DEV Community: Cortex AI

How an AI agent for QuickBooks works

Cortex AI — Fri, 16 Oct 2026 17:46:41 +0000

An AI agent for QuickBooks connects to QuickBooks Online through OAuth 2.0, with no API keys, and reads or drafts work such as unbilled jobs and overdue invoices. Tokens expire (one hour for access, 100 days for refresh), webhooks carry only record IDs, and updates replace the whole record. Your office manager or bookkeeper should approve anything that changes the ledger or reaches a customer.

An AI agent for QuickBooks is a program that reads your books through Intuit's API, notices what needs a human (an invoice 30 days past due, a finished job nobody billed, a customer who paid but still shows open) and drafts the next step for your office manager to approve. For a service company the books are where field work turns into money, so the useful agents sit between QuickBooks and the software your crew uses all day. The one rule that matters most: a person approves anything that changes the ledger or goes to a customer.

This guide covers what QuickBooks already includes, how an outside agent connects, the technical limits that trip up builds, and what to keep with your bookkeeper.

What QuickBooks already does

Check this first so you don't pay twice. Intuit has added AI agents to QuickBooks Online. According to a vendor's rundown of the agents, there is an Accounting Agent that categorizes transactions and flags anomalies, and a Payments Agent that predicts late payments and sends invoice reminders. Which agents you get depends on your plan. The same article says the agents work inside the ledger and do not update your field service software or message your team, and that bookkeepers still advise "trust, but verify."

That article is written by a company that sells an AI assistant, and the plan details come from Intuit's marketing, so check them in your own account. The takeaway is simple: reminders and categorizing are covered. The gaps are across systems.

Where the built-in tools stop

QuickBooks sees invoices. It does not see that a job finished on Tuesday and nobody created the invoice, because that fact lives in ServiceTitan, Jobber or Housecall Pro. For a trades company the expensive gaps are these:

Unbilled work. Completed jobs with no invoice. See unbilled work.
Slow collections. A reminder that goes out on a schedule and ignores the customer who just called about the balance. See invoice reminders.
Two sources of truth. The field system says paid, QuickBooks says open.
The owner's morning view. Cash in, money owed, jobs done but not billed, in one place.

How an agent connects to QuickBooks

QuickBooks Online uses OAuth 2.0 only, according to one developer guide. There are no API keys. The company owner (or an admin) approves the connection once, and the agent then holds tokens.

Three details matter in practice:

Tokens expire. Intuit's own sample code says the access token is valid for one hour and the refresh token for 100 days before you must re-authorize. The developer guide adds that the refresh token changes each time you renew, so a build has to store the newest one. A connection that nobody watches simply dies, and the agent goes quiet.
Limits. One guide puts the cap at about 500 requests per minute per company, with batches of up to 30 operations, and says going over returns an HTTP 429. Those numbers are approximate, so read Intuit's current limits page before you size a build.
Webhooks are thin. The developer guide says notifications cover 24 entity types (including invoices, payments, estimates and customers) and that "webhook payloads only contain entity IDs and operation types," so the agent has to fetch the record afterward. The limits guide recommends pairing webhooks with periodic checks because notifications can be delayed or duplicated.

Connecting ChatGPT to QuickBooks

We found no direct ChatGPT connector from Intuit in our research. An assistant would need the OAuth connection above and a place where your bookkeeper approves what it drafts. Without that second part, a chatbot with access to your books is a risk, not a helper.

What an agent does around QuickBooks

These are illustrations of the pattern, not a client story.

Agent	Starts when	Reads	Drafts	Person approves
Unbilled work finder	A job is complete in the field system with no matching invoice	Job records, QuickBooks invoices	A list of jobs to bill, with the likely invoice	Office manager
Collections follow-up	An invoice passes a set number of days overdue	Invoice, balance, last contact	A personal reminder, or a call note	Office manager
Payment match	A payment arrives	Payment, open invoices	A suggested match	Bookkeeper
Owner cash report	Every morning	Invoices, payments, aging	One summary	None, it only reads

Setup steps and what you need

A QuickBooks admin who can approve the app connection.
A read-only start. Let the agent read for two weeks and compare what it says with your own numbers before it drafts anything.
Access to the field system and phone provider, so the agent can see what QuickBooks can't.
One person per agent approving drafts, usually your office manager or bookkeeper.

Updates in QuickBooks replace the whole record, according to the developer guide, which says partial updates are not supported and that each change must carry the record's current version number. An agent that edits an invoice from stale data can wipe out fields. This is one reason we start with agents that read and draft, not agents that write to the ledger.

Limits and risks

Money moves stay human. Writing off a balance, issuing a credit or voiding an invoice is the owner's or bookkeeper's call.
Accuracy is a real problem. In the Federal Reserve's survey of small firms already using AI, 46% named accuracy as a top challenge. That is a reason to review drafts, not to avoid them.
Double reminders. If QuickBooks sends its own payment reminders and an agent also writes to the customer, pick one for each stage.
Upkeep. Tokens expire, limits change and fields get renamed. Someone has to own that, which is the reason for managed AI operations.

Where Cortex fits

Cortex builds AI agents that work inside the software you already run, with a person approving anything that changes your books or reaches a customer. We are not an Intuit partner, and we'll say so when QuickBooks' own tools are enough. See what we connect on integrations, and read how the same pattern works for Jobber. The approval model is explained under human-in-the-loop AI.

The free AI Operations Audit looks at your field software, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does QuickBooks have AI built in?

Yes. A vendor's rundown describes an Accounting Agent that categorizes transactions and a Payments Agent that predicts late payments and sends reminders, with access depending on plan. Check your own account for what your plan includes.

Does QuickBooks Online have an API?

Yes. It uses OAuth 2.0 only, with no API keys, according to a developer guide. The owner or an admin approves the connection once, and the app then holds tokens that must be refreshed.

Can I connect ChatGPT to QuickBooks?

We found no direct connector from Intuit. An assistant would need the OAuth connection and a place where your bookkeeper approves what it drafts, because a chatbot with write access to your books is a risk.

Does QuickBooks support webhooks?

Yes, for 24 entity types including invoices, payments, estimates and customers. Payloads carry only record IDs and the operation, so the agent fetches the record afterward. Notifications can be delayed or duplicated.

How do I keep an AI agent from changing the wrong thing in my books?

Start read-only for two weeks, compare its output to your own numbers, then let it draft while a person approves. QuickBooks updates replace the whole record, so edits from stale data can erase fields.

Sources

QuickBooks Online API: Complete Developer Guide for 2026, Satva Solutions
Top 5 QuickBooks API Rate Limits (2026), Satva Solutions
OAuth2 Tokens help, Intuit Developer (GitHub sample)
QuickBooks AI: The Agents, the Tiers, and What They Miss (2026), Carly
2026 Report on Employer Firms, Federal Reserve Banks (Small Business Credit Survey)

Originally published at buildcortex.dev.

How an AI agent for Shopmonkey works

Cortex AI — Fri, 16 Oct 2026 15:46:35 +0000

An AI agent for Shopmonkey connects through an API key an admin creates under Settings, Integration, API Keys, and through webhooks that report changes to orders, messages, inspections and other records. Estimates, repair orders and invoices are all orders told apart by status, and webhooks report record changes, not business events. A service advisor should approve every customer message.

An AI agent for Shopmonkey is a small program that watches your shop's records (an estimate nobody answered, a repair order that finished while the car sat in the lot, a customer who declined a brake job in the spring) and drafts the next step for a service advisor to approve. Shopmonkey gives it two ways in: an API key an admin creates, and webhooks that tell the agent when something changed. What the agent should not do is send anything to a customer without a person reading it first.

This guide covers how the connection works, what the data looks like, where builds go wrong, and what stays with your advisors.

What Shopmonkey gives an agent to work with

Shopmonkey describes its API and webhooks page in plain terms: an API is "a door through which information and capabilities can be shared between two softwares," and a webhook is a kind of API driven by events instead of requests. The page says Shopmonkey can send information to other systems so they update on their own. It does not say which plans include access or how to get a key, so ask Shopmonkey or your account rep before you plan around it.

The developer site, shopmonkey.dev, links to the API overview, a webhooks page and an enterprise data streaming option. Third parties fill in the rest, and the quality varies. One of them, Carly, is a company that competes in this space, so read its claims as a starting point and confirm them in your own account.

How an agent connects

An API key from an admin. Carly's guide says an admin creates the key under Settings, Integration, API Keys, and that a key carries the same permissions as the person who made it. Two practical consequences follow. If that admin leaves or has their role narrowed, the integration starts failing with a 403. And the key can't be viewed again after you create it, so a lost key means a new one and a broken connection until it is swapped. Create keys from a dedicated account with only the access the agent needs.

A Supergood review adds that there is no sandbox or developer tier, so a builder tests inside a real shop's account, and that there is no fine-grained scope for keys. That review comes from a company that sells API tooling, so treat it as one view. The practical point holds either way: test with read-only work first.

Webhooks. Carly lists eleven kinds of records that send events when they are created, changed or deleted: appointments, customers, inspections, inventory, messages, orders, payments, purchase orders, users, vehicles and vendors. These are changes to records, not business events. One edit in the shop can fire several webhooks. An agent has to work out for itself that "this estimate has gone quiet" is a fact, because no event says so.

Two details matter for reliability. Delivery times out after about 55 seconds, so the receiving end should acknowledge fast and do its work afterward. And after repeated failures a webhook is switched off and missed events are not replayed on their own, according to the same guide. Someone has to notice and replay them.

One thing that surprises people: estimates are orders

In Shopmonkey's data, per Carly's guide, an estimate, a repair order and an invoice are all the same kind of record, an order, told apart by a status field. There is no separate estimate endpoint. An agent that follows up on quiet estimates therefore watches orders in the Estimate status and ignores the rest. A build that skips this will either miss estimates or text customers about finished jobs.

What an agent does around Shopmonkey

These are examples of the pattern, not a client story.

Agent	Starts when	Reads	Drafts	Person approves
Estimate follow-up	An order in Estimate status has had no reply after a set number of days	Order lines, vehicle, customer messages	A personal text, ranked by dollar value	Service advisor
Declined work revisit	A customer is due back with declined inspection items	Inspection results, last visit, vehicle	A message naming the specific item	Service advisor
Ready for pickup	An order moves to complete and the car is still on site	Order status, customer contact	A pickup message	Front desk
Owner morning report	Every morning	Orders, payments, appointments	One summary	None, it only reads

Setup steps and what you need

A Shopmonkey admin willing to create a limited key and add a webhook.
Two or three jobs, in order. Quiet estimates and declined work usually come first.
Access to the other systems in play: your phone provider, QuickBooks, review accounts. Supergood reports that Shopmonkey's built-in catalog is short (QuickBooks Online, CARFAX, PartsTech, Nexpart and RepairLink), so most other connections need custom work.
One advisor per agent approving drafts for the first few weeks.

Start with a person approving every message and loosen it only after you have read enough drafts to trust them. We explain the model in human-in-the-loop AI.

Limits and risks

Safety work stays human. An agent never tells a customer a car is safe to drive or talks anyone out of a brake or steering repair.
Prices and warranty. Discounts, goodwill repairs and warranty calls belong to the advisor.
Wrong customer. Match on phone number and vehicle, not just a name.
Double messages. If Shopmonkey's own messaging and an agent both text about the same car, pick one for each stage.
Upkeep. Keys, permissions and events change. Someone has to own that, which is the reason for managed AI operations.

Where Cortex fits

Cortex builds AI agents that work inside the software you already run, with a person approving anything that goes to a customer. We are not a Shopmonkey partner or a certified app, and we will say so when Shopmonkey's own tools are enough. See what we connect on integrations and how this looks for auto repair. For recovering work that fell off the board, read about declined work. If your shop runs a different system, see AI agents for Tekmetric or AutoLeap.

The free AI Operations Audit looks at your shop software, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does Shopmonkey have an API?

Yes. Shopmonkey's API and webhooks page describes both, and its developer site is shopmonkey.dev. The page does not say which plans include access, so confirm that with Shopmonkey before you plan a build.

Does Shopmonkey support webhooks?

Yes. A third-party guide lists eleven record types that send events, including orders, messages, inspections and payments. These are record changes, so an agent has to work out for itself that an estimate has gone quiet.

Can I connect ChatGPT to Shopmonkey?

We found no native ChatGPT connector. An assistant would need an admin-created API key or a webhook feed, plus a place where a service advisor approves drafts before anything reaches a customer.

Is there a sandbox to test an AI integration?

One third-party review reports there is no sandbox or developer tier, so testing happens in a real shop account. Start with read-only work and give the agent a key with limited access.

What happens if the admin who made the API key leaves?

A third-party guide says a key carries the same permissions as its creator, and the integration fails with a 403 if that person is offboarded or their role is narrowed. Create keys from a dedicated account.

Sources

API & Webhooks Integration, Shopmonkey
Shopmonkey API documentation, Shopmonkey
Shopmonkey API guide, Carly
Shopmonkey API report card, Supergood

Originally published at buildcortex.dev.

How an AI agent for Tekmetric works

Cortex AI — Thu, 15 Oct 2026 23:49:04 +0000

An AI agent for Tekmetric connects one of two ways: an app approved by Tekmetric that the shop admin authorizes under Shop Settings, Integrations, or a webhook an admin adds themselves. API credentials are not self-serve (one review reports a wait of 2 to 3 weeks). Tekmetric already covers texting, reminders and campaigns; an agent adds follow-up judgment on estimates and declined work, with a service advisor approving every message.

An AI agent for Tekmetric is a small program that watches what happens in your shop (an estimate nobody answered, a repair order that finished while the car sat on the lot, declined work for a customer who is due back) and drafts the next step for a service advisor to approve. In Tekmetric it can reach your data in two ways: through an app Tekmetric has approved, which your shop admin authorizes, or through a webhook an admin adds themselves. What it should not do is send anything to a customer without a person reading it first.

This guide covers what Tekmetric already does, where the gaps are, how an agent connects, and what to keep with your advisors.

What Tekmetric already does

Check this first so you don't pay for something you already have.

Tekmetric's integrations page says it has 70+ integrations and lists two-way texting, text-to-pay, digital vehicle inspections, automated reminders, scheduled campaigns and Google review requests. Partners such as ShopGenie and AutoBucks add things like service reminders and a text-based rewards program. We did not find AI features described on that page.

So the reminders and campaigns are already there. They run on rules: this message, at this time, to this list.

Where the built-in tools stop

A reminder fires on a date. It doesn't know that the customer who declined a suspension job last month asked about financing, or that two estimates are sitting unanswered on the same day your advisor is out.

The gaps we hear about most:

Estimates that go quiet. Sent, viewed, never answered, then forgotten.
Declined work. Inspection findings the customer put off, never revisited at the right moment.
Finished cars that wait. A repair order marked complete and a customer who hasn't been told.
Money and parts elsewhere. Balances in QuickBooks, parts in a supplier portal, leads in your phone system.
The owner's morning view. Cars in, estimates open, cars ready, balances owed.

What an agent does around Tekmetric

These are illustrations of the pattern, not a client story.

Agent	Starts when	Reads	Drafts	Person approves
Estimate follow-up	An estimate has no response after a set number of days	Estimate lines, vehicle, notes, history	A personal text or a call reminder, ranked by dollar value	Service advisor
Declined work revisit	A customer is due back and has open declined items	Declined lines, mileage, last visit	A message naming the specific item	Service advisor
Ready for pickup	A repair order is complete and the car is still there	Repair order status, contact	A pickup message	Front desk
Owner morning report	Every morning	Orders, estimates, balances	One summary	None, it reads only

How an agent connects to Tekmetric

Two routes exist, and they are quite different.

An approved app, authorized by your admin. Tekmetric points developers to an application page. One third-party review reports that credentials are not self-serve: they need an active paid subscription and manual approval from Tekmetric that typically takes 2 to 3 weeks, and they use OAuth 2.0 client credentials. It also reports that each query returns at most 25 records, so reading a whole shop's history takes many requests. That review is published by a company that sells managed API layers, so confirm these details with Tekmetric. Once a connected app exists, the shop admin grants it access: Shop Settings, then the Integrations tab, then Manage, then Grant Access. Revoking it works from the same screen.
A webhook you add yourself. A partner's setup guide shows that an admin can add a webhook under Settings, then Integrations, and that webhooks deliver appointment and repair order updates in real time, instead of waiting for an hourly sync. We found no public list of events, so ask Tekmetric which ones exist before building on them.

Webhooks tell an agent something changed, which suits follow-up. A good agent then reads the record, checks it hasn't already acted on that event, and drafts. Events can repeat, and a careless build messages the customer twice.

Connecting ChatGPT to Tekmetric

We found no native ChatGPT or Claude connector. An assistant would need approved API access or a webhook feeding it, plus a place for a person to approve drafts. That last part is what turns a chatbot into something safe to put near customers.

Setup steps and what you need

A Tekmetric admin who can grant access and add integrations.
Two or three jobs in order. Quiet estimates and declined work are usually first.
Access to the other systems: phone provider, QuickBooks, review accounts.
One advisor per agent approving drafts for the first few weeks.

Start with a person approving every draft and loosen it only after you've read enough to trust it. We explain the model in human-in-the-loop AI.

Limits and risks

Safety work stays human. An agent never tells a customer a car is safe to drive or talks anyone out of a brake or steering repair.
Prices and warranty. Discounts, goodwill repairs and warranty calls are the advisor's.
Double messages. If Tekmetric's automated reminders and an agent both text about the same car, pick one per stage.
Wrong customer. Match on phone number and vehicle, not just a name.
Upkeep. Credentials and events change. Someone has to own that, which is the reason for managed AI operations.

Where Cortex fits

Cortex builds AI agents that work inside the software you already run, with a person approving anything that goes to a customer. We are not a Tekmetric partner or certified app, and we'll say so when Tekmetric's own tools are enough. See what we connect on integrations and how this looks for auto repair. For recovering work that fell off the board, see declined work. If your shop runs another system, read AI agents inside AutoLeap.

The free AI Operations Audit looks at your shop software, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does Tekmetric have AI built in?

Tekmetric's integrations page lists two-way texting, automated reminders, scheduled campaigns, Google review requests and digital vehicle inspections. We did not find AI features described on that page.

Does Tekmetric have an API?

Yes, but it is not self-serve. Tekmetric sends developers to an application page, and one third-party review reports it needs an active paid subscription plus manual approval that typically takes 2 to 3 weeks.

Can I connect ChatGPT to Tekmetric?

We foun

How an AI agent for AccuLynx works

Cortex AI — Thu, 15 Oct 2026 19:46:49 +0000

An AI agent for AccuLynx connects through the AccuLynx API using a key that an administrator creates, limited to 10 requests per second per key, with webhook subscriptions for change events. AccuLynx already scores leads with its own AI (Lead Rank); an agent adds follow-up on stalled estimates and jobs, and reporting across your phones and books, with a person approving what goes to homeowners.

An AI agent for AccuLynx is a small program that watches your account for events (a lead that nobody called back, an estimate that went quiet, a job with no invoice) and drafts the next step for someone in your office to approve. It connects through the AccuLynx API with a key an administrator creates. AccuLynx already uses AI to rank leads. What it doesn't do on its own is chase the estimate a homeowner went silent on, or compare your jobs with what you've actually collected.

This guide covers what AccuLynx already does, where the gaps are, exactly how an agent connects, and what should stay with a person.

What AccuLynx already does

Start here, so you don't pay for something you already own.

Lead Intelligence. AccuLynx announced it in 2023: it uses AI to predict homeowner buying behavior, scoring leads against hundreds of third-party data attributes and showing a Lead Rank on leads and prospects. It is included in all subscriptions. The release gives no accuracy or win-rate figures, so treat it as a sorting aid, not a promise.

Texting, documents, QuickBooks and Zapier. AccuLynx's product roundup lists text messaging, document automation, an Invoice Worksheet that records invoices to QuickBooks, and a Zapier integration that can send leads in from other apps. AccuLynx notes there are limits on which software and data can move through Zapier.

If a ranked lead list and templated texts solve your problem, use those first.

Where the built-in tools stop

A rank tells you who to call. It doesn't make the call, and it doesn't remember that the homeowner on Maple Street asked about a metal upgrade and went quiet three weeks ago.

The gaps we see most in roofing offices:

Estimates that go silent. After the inspection and the proposal, follow-up depends on whoever remembers.
Slow lead response. A web form lands after hours and gets a reply the next afternoon.
Job stalls. A job sits between steps (approved, materials, scheduled) and nobody owns the gap.
Money in two places. A completed job in AccuLynx and an open balance in QuickBooks are two screens.
The owner's view. Leads by source, estimates out, jobs in production and collections in one message.

What an agent does inside AccuLynx

These are illustrations of the pattern, not a client story.

Agent	Starts when	Reads	Drafts	Person approves
New lead reply	A lead is created from a form or call	Lead record, source, address	A first reply and a task for the right rep	Sales rep
Estimate follow-up	An estimate has no response after a set number of days	Estimate, notes, past messages	A personal text or call reminder, ranked by job size	Sales rep
Stalled job check	A job sits in one milestone longer than normal	Job milestone, dates, notes	A flag and a question for the project manager	Project manager
Completed job, open balance	A job is complete and QuickBooks shows an open balance	Job, invoice	A payment reminder	Office manager

A fifth agent sends the owner one morning summary. It reads and never writes.

How an agent connects to AccuLynx

An API key from an administrator. You need an active account and a created, named API key, and the docs say a technical understanding of APIs is required. A third-party help guide reports the keys live in a paid add-on called AppConnections, that only a Location or Company Administrator can create them, that each location needs its own key, and that the full key is shown once. Confirm the add-on and its cost with AccuLynx.
Rate limits. AccuLynx's docs set 30 requests per second per IP and 10 requests per second per API key. Exceed them and you get an HTTP 429 and a ban lasting from about 30 seconds to a few minutes. A careful agent reads what changed instead of re-pulling every job.
Webhooks. You can create a subscription by sending a request with a JSON body and choosing topics, using your key as a bearer token. The page we read doesn't describe retries or signature checks, so ask AccuLynx or the builder how duplicates and failed deliveries are handled.
Higher limits. The docs point to AccuLynx's Partner Program for anything above the standard limits.

Connecting ChatGPT to AccuLynx

We found no native ChatGPT or Claude connector in AccuLynx's documentation. The working paths are Zapier for one fixed step, or a custom app on the API when you need memory, judgment across systems and an approval queue.

Setup steps and what you need

An AccuLynx administrator to create the key, with the add-on enabled.
Two or three jobs to hand off, in order. New lead reply and quiet estimates usually come first.
Access to the other systems: phones, QuickBooks, ad accounts for the morning report.
One person per agent who approves drafts for the first few weeks.

Start with a person approving every draft and loosen it only after you've read enough to trust it. We explain the model in human-in-the-loop AI.

Limits and risks

Prices and promises. Scope, price and discounts stay with a person. So do insurance conversations.
Duplicate messages. If AccuLynx texting and an agent both reach a homeowner about the same estimate, pick one per stage.
Key handling. The key is shown once and opens the account. Store it safely and know who holds it. If it leaks, create a new one.
Maintenance. Keys, limits and endpoints change. Someone has to own that, which is the reason for managed AI operations.

Where Cortex fits

Cortex builds AI agents that work inside the software you already run, with a person approving anything that leaves the building. We are not an AccuLynx partner or certified app, and we'll tell you when AccuLynx's own tools are enough. See what we connect on integrations and how this looks for roofing companies. For jobs that stall, see stalled jobs. For the bigger picture, read the AI operations layer.

The free AI Operations Audit looks at your AccuLynx account, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does AccuLynx have AI built in?

Yes. AccuLynx announced Lead Intelligence in 2023, which scores leads with a Lead Rank using third-party data and is included in all subscriptions. It ranks leads. It does not follow up on a stalled estimate for you.

Does AccuLynx have an API?

Yes. You need an active account and an API key, and the docs say a technical understanding of APIs is required. A third-party help center reports the keys sit in a paid add-on called AppConnections.

Who can create an AccuLynx API key?

According to a third-party help guide, only a Location or Company Administrator. Each location needs its own key, and the full key is shown only once, so it has to be stored when it is created.

Can ChatGPT connect to AccuLynx?

We found no native ChatGPT connector in AccuLynx's documentation. Options are Zapier, which AccuLynx says has limits on what data it can move, or a custom app using the API.

What are the AccuLynx API rate limits?

AccuLynx's docs set a limit of 30 requests per second per IP and 10 requests per second per API key. Go over and you get an HTTP 429 and a short ban, so a good agent reads only what changed.

Sources

Getting Started, AccuLynx API docs
Rate Limits, AccuLynx API docs
Webhooks End User Reference, AccuLynx API docs
How to create an API key in AccuLynx, Alivo
AccuLynx Launches New AI-Powered Lead Intelligence Feature, PRWeb
Product Update Roundup, AccuLynx

Originally published at buildcortex.dev.

How an AI agent for AutoLeap works

Cortex AI — Thu, 15 Oct 2026 17:48:39 +0000

An AI agent for AutoLeap works through AutoLeap's partner API, which one third-party review says is issued by hand to approved partners, uses a static key pair and has no documented webhooks, so an agent has to poll for changes. AutoLeap already sells its own AI receptionist (AutoLeap AIR); an outside agent adds follow-up judgment on estimates and declined work, with a service advisor approving every message.

An AI agent for AutoLeap is a small program that watches the data in your shop management system (estimates that nobody answered, repair orders that sat, customers due back) and drafts the next step for a service advisor to approve. The catch with AutoLeap is access. According to one third-party review, its partner API is issued by hand, and it has no documented webhooks, so an outside agent has to check for changes on a schedule rather than being told the moment they happen.

This guide covers what AutoLeap already offers, where the gaps are, how an agent can connect, and what a person should keep.

What AutoLeap already does

Check this first, because some shops pay for an "AI agent" that repeats something they already own.

AutoLeap sells AutoLeap AIR, an AI receptionist that answers the phone so a call during a brake job doesn't go to voicemail. Its feature list also includes reminders, customer communication and a QuickBooks connection for accounting. We did not find a public description of anything that decides which open estimate deserves a call today, so that is the gap an outside agent can fill.

If your problem is unanswered phones, look at AutoLeap AIR before you hire anyone. It is built into the system your advisors already use.

Where the built-in tools stop

A receptionist and reminders are rules and scripts. They don't read an estimate and notice that a big brake and rotor job was declined by a customer who said "after payday."

The gaps we hear about most in shops:

Estimates that go quiet. The advisor sent it, the customer said they'd think about it, and nobody looks again.
Declined work. Recommendations from the inspection that were never revisited, even when the customer's next oil change is due in three weeks.
Money and parts in other places. Paid status in QuickBooks, parts orders in a supplier portal, ad leads in another tab.
The owner's morning view. One page: cars in, estimates waiting, cars ready for pickup, balances open.

What an agent does around AutoLeap

These are illustrations of the pattern, not a client story.

Agent	Starts when	Reads	Drafts	Person approves
Estimate follow-up	An estimate has sat unanswered for a set number of days	Estimate lines, vehicle, customer notes, past visits	A personal text or a call reminder, ranked by dollar value	Service advisor
Declined work revisit	A customer is due for service and has open declined items	Past declined lines, mileage, last visit	A reminder that names the specific item	Service advisor
Ready-for-pickup nudge	A repair order is marked complete and the car is still on the lot	Repair order status, customer contact	A pickup message	Front desk
Owner morning report	Every morning	Orders, estimates, balances	One summary message	None, it only reads

How an agent can connect

This part comes with a warning. We could not open AutoLeap's own developer site, so everything below comes from a third-party review by Supergood, a company that sells managed API layers and so has a stake in making APIs look hard. Treat it as a starting point and confirm each item with AutoLeap.

Access is requested, not self-serve. The review says developers ask AutoLeap for a Partner ID and Auth Key, that it requires a paid subscription and a partnership, and that there is no sandbox or third-party developer tier.
A static key pair. It lists no OAuth and no per-shop consent screen. In practice that means whoever holds the key can reach whatever it unlocks, so it must be stored carefully and a person should know it exists.
A REST API with a published spec. The review describes a versioned v2 REST API with an OpenAPI spec covering customers, vehicles, items, inventory, repair orders, payments, technician timesheets and taxes.
No webhooks documented. An agent would poll, which means a delay of minutes instead of seconds. That is fine for follow-up, which works in days, and not fine for anything that needs an instant reaction.
Limits are unpublished. The review says rate limits are not documented, so a careful build reads only what changed.

For contrast, Tekmetric lists 70+ integrations and lets a shop admin add a webhook under Settings, then Integrations for a connected app, so the two systems differ here. If AutoLeap will not grant access for what you want, there are two fallbacks: scheduled exports for read-only reporting, or keeping the agent outside AutoLeap, where it watches your phones, texts and QuickBooks and hands the advisor a draft to enter.

Connecting ChatGPT to AutoLeap

We found no native ChatGPT or Claude connector. The same review reports no MCP server as of August 2026. A ChatGPT-style assistant would sit on top of whatever the partner API allows, with the same limits.

Setup steps and what you need

An AutoLeap admin who can ask AutoLeap about integration access for your shop.
Two or three jobs you want handled, in order. Estimates that went quiet are usually first.
Access to the other systems involved: your phone provider, QuickBooks, review and ad accounts for the morning report.
One advisor per agent who approves drafts for the first few weeks.

Start with a person approving every draft. Loosen that only after you've read enough of them to trust the agent. We explain the model in human-in-the-loop AI.

Limits and risks

Safety work stays human. An agent never tells a customer a car is safe to drive, and never talks someone out of a brake or steering repair.
Prices and promises. Discounts, warranty decisions and anything about price are the advisor's.
Double messages. If AutoLeap reminders and an agent both text the same customer about the same car, pick one per stage.
Credential drift. Keys and versions change. Someone has to own that, which is the reason for managed AI operations.

Where Cortex fits

Cortex builds AI agents that work inside the software you already run, with a person approving anything that goes to a customer. We are not an AutoLeap partner or a certified app, and we'll say so when AutoLeap's own tools are enough. See what we connect on integrations and how this looks for auto repair. For recovering work that quietly fell off the board, see declined work. Related reading: AI agents inside Jobber shows the same pattern in a system with an open API.

The free AI Operations Audit looks at your shop software, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does AutoLeap have AI built in?

AutoLeap sells an AI receptionist called AutoLeap AIR for answering calls, and lists appointment reminders and customer communication among its features. We found no description of AI that decides which estimate to chase.

Can I connect ChatGPT to AutoLeap?

Not directly that we could find. Per one third-party review, AutoLeap has no public developer tier and no MCP server (checked August 2026). A connection would run through the partner API, which AutoLeap grants to approved partners.

Does AutoLeap have an API?

Yes, but a third-party review says access is not self-serve. You ask AutoLeap for a Partner ID and Auth Key, and it requires a paid subscription and a partnership. Ask AutoLeap directly what it will give a shop-specific integration.

Does AutoLeap send webhooks?

None are documented in the public partner API reference, according to that same review. An agent would check for changes on a schedule instead of being told the moment something happens.

Is it safe to let an AI agent write in my shop management system?

Start with read-only access and drafts that a service advisor approves. Pricing, discounts, warranty promises and anything about brakes, steering or tires stay with a person.

Sources

Technical Advice (AutoLeap blog), AutoLeap
AutoLeap API report card, Supergood
Integrations, Tekmetric
Setting up Webhooks in Tekmetric, Shopgenie

Originally published at buildcortex.dev.

AI agents inside PestPac

Cortex AI — Wed, 14 Oct 2026 19:55:02 +0000

Originally published at buildcortex.dev.

An AI agent for PestPac connects through WorkWave's PestPac API (available to customers and approved developers, with a setup fee) or scheduled CSV reports, watches for cancellations, unscheduled service, idle leads and past-due invoices, and drafts the next step for a person to approve. PestPac already automates reminders and conversation summaries; the agent adds judgment and work that crosses into phones, QuickBooks and the CRM.

An AI agent for PestPac is a program connected to your PestPac database through WorkWave's API that watches for things your office would catch if it had time: a recurring customer who cancelled, a service order that never got scheduled, a lead nobody called, an invoice going past due. It drafts the next step, and a person on your team approves it. PestPac already sends reminders and summarizes conversations on its own. The agent covers the judgment calls and the work that crosses into your phones, QuickBooks and CRM.

This guide covers what PestPac does natively, where agents fit, what it takes to connect, and what should never be automated.

Why pest control is a good fit for agents

Pest control runs on recurring service. NPMA's 2025 industry figures put recurring revenue at 85.4% of residential service revenue, across an industry that brought in $13.416 billion. The same release counted 16,565 firms, with 81.4% running one or two locations.

That combination matters. A small office handling thousands of recurring accounts loses money quietly: a skipped quarterly, a cancellation nobody tried to save, a renewal that lapsed. None of it shows up as a crisis. It shows up as a slow year.

Staffing makes it harder. The Bureau of Labor Statistics puts pest control worker median pay at $45,250 and expects about 13,700 openings a year through 2035. The technicians are busy, and so is whoever answers the phone.

What PestPac already automates

Check these before you pay anyone for an agent.

Customer Notifications. PestPac can send call, text or email service reminders and on-the-way messages, with up to five touchpoints per reminder rule, and log customer confirmations.
Communication Center. Email, two-way texting and web chat in one place, with AI-generated summaries of each conversation.
Customer portal and Ask to Pay. Customers can update their own accounts and pay from a text link.
Scheduled reports. Report Writer can run reports on a schedule and deliver them automatically, and every report exports to CSV.
WAIve. WorkWave announced its AI platform in January 2026, including WAIve Teammates, specialized agents rolling out through 2026 starting in the second quarter, with the customer keeping the final say. Ask your WorkWave rep what is live on your account today.

If reminders and on-the-way texts are your whole problem, turn on Customer Notifications and stop there.

Where an agent adds something

PestPac's automations follow the rules you set. They don't decide which cancelled account is worth a call from the owner, and they don't see your phone system, ad accounts or QuickBooks unless an integration brings that data in.

These are the agents we'd consider in a PestPac shop. They are illustrations of the pattern, not client results.

Agent	Starts when	Reads	Drafts	Person approves
Cancellation save	A recurring service setup is cancelled or suspended	Location history, service notes, last invoices, reason code	A save call script or a text, ranked by account value	Office manager or owner
Unscheduled work	A service order or follow-up has no date after a set number of days	Service orders, routes, technician notes	A scheduling task with the customer's preferred times	Scheduler
Lead follow-up	A lead sits with no activity	Lead, source, any call recording notes	A first reply and a follow-up plan	Sales rep
Past due	An invoice passes your terms	Invoices, payments, account notes	A reminder in your voice, or a flag for a call	Billing

One more is worth it for most owners: a morning summary of new leads, cancellations, unscheduled work and open balances, pulled from PestPac and QuickBooks. Our customer reactivation page walks through the save workflow in more detail.

How an agent connects to PestPac

PestPac is less open than some field service platforms, so plan for this step.

API access. WorkWave's developer portal says you must be a PestPac customer or an approved developer working for one, that there is a one-time setup fee for the API platform, and that further costs depend on data access and calls per day. A 30-day trial is available on request through WorkWave's API sales team.
Plan tier. PestPac's pricing page lists "API Integrations" in the Enterprise feature list. Confirm your tier before anyone scopes a build.
Credentials. WorkWave issues them. Third-party integrations such as ClickPoint's document the set they need: API key, tenant ID, client ID and secret, a username and the PestPac host URL. Create a dedicated PestPac user for the agent so its actions are logged under its own name.
What it can reach. WorkWave describes the PestPac API as covering servicing, scheduling, billing and payments, with webhook notifications when your data changes. In practice an agent reads locations, service orders, leads and invoices, and writes tasks and notes.
When the API isn't an option. Report Writer exports to CSV on a schedule. An agent can work from those files for reporting and flagging, with no writes back. It's slower, but it's honest and cheap.

Connecting ChatGPT to PestPac

There's no ChatGPT connector in WorkWave's materials that we found. The working paths are the PestPac API (through a custom app or an integration platform) or scheduled CSV reports passed to a model. Either way, don't paste customer lists into a personal ChatGPT account. Use a business account with data controls, or run the model privately.

Setup steps and requirements

Confirm API access and cost with WorkWave, in writing.
Pick one job to start. For most pest companies it's cancellation saves or past-due invoices.
Map your codes. Cancellation reasons, service types and lead sources must be used consistently, or the agent can't tell a move-out from an angry customer.
Name the approver for each agent.
Run every draft past that person for the first few weeks.

Limits and risks

Regulatory records stay human. PestPac holds material usage and state reporting. An agent should never write to chemical application records or regulatory reports.
Pricing stays human. No discounts or price changes from an agent, even to save a cancellation.
Double messages. If PestPac reminders and an agent both text the same customer, it looks careless. Assign each message type to one system.
Cost surprises. API pricing depends on calls per day, so an agent that polls everything constantly costs more. Good builds react to changes and read only what moved.
Upkeep. Credentials rotate, APIs change, and models are retired. Someone has to own that, which is why we offer managed AI operations.

Where Cortex fits

Cortex builds AI agents for service businesses, including pest control companies, inside the software they already run. We are not a WorkWave partner, and we'll tell you if PestPac's own tools already cover what you need. See how we approach integrations and why the cross-system view matters in the AI operations layer.

The free AI Operations Audit is a 30-minute look at your PestPac setup, phones and books. You leave knowing which job to hand an agent first and what API access you'd need.

Questions people ask

Does PestPac have an API?

Yes. WorkWave offers a PestPac API to customers and approved developers working for them, with a one-time setup fee and further costs based on data access and daily calls. PestPac's pricing page lists API integrations in its Enterprise tier.

Does PestPac have AI features?

PestPac's Communication Center adds AI-generated summaries of customer conversations, and WorkWave announced its WAIve AI platform in January 2026 with specialized agents rolling out through the year. Ask WorkWave what is live on your account.

How much is PestPac software?

PestPac's pricing is quote-based across Small Business, Professional and Enterprise tiers; WorkWave doesn't publish a flat rate card. API access carries its own setup fee and usage-based costs.

Can I connect ChatGPT to PestPac?

Not through a native connector that we found. The options are a custom app on the PestPac API or scheduled CSV exports passed to a model, using a business account with data controls rather than a personal ChatGPT login.

What should an AI agent never do in PestPac?

It should not write chemical application or regulatory records, change prices or offer discounts, or message customers who already get PestPac's own reminders for the same event. Those stay with people or with one system.

Sources

US Pest Control Industry Sustains Steady Growth with 6% Increase in 2025, National Pest Management Association
Occupational Outlook Handbook: Pest Control Workers, U.S. Bureau of Labor Statistics
Customer Communication, PestPac by WorkWave
Pest Control Reporting Software, PestPac by WorkWave
WorkWave Transforms Field Service Operations with WAIve, WorkWave / TEAM Software
WorkWave Developer Portal, WorkWave
Pricing, PestPac by WorkWave
PestPac by WorkWave integration, ClickPoint SalesExec Help
PestPac API, PestPac by WorkWave

AI agents inside Jobber

Cortex AI — Wed, 14 Oct 2026 15:47:09 +0000

Originally published at buildcortex.dev.

An AI agent for Jobber connects through Jobber's official GraphQL API (authorized by a company admin through OAuth) and webhooks, watches for events like quiet quotes, missed calls and unpaid closed jobs, and drafts the next step for a person to approve. Jobber already handles two fixed quote and invoice follow-ups and offers its own AI Receptionist; an agent adds judgment, connections to phones and QuickBooks, and cross-system reporting.

An AI agent for Jobber is a small program, connected through Jobber's official API, that watches your account for events (a new request, a quote that went quiet, a job closed with no invoice) and drafts the next step for someone on your team to approve. Jobber already automates some of this on its own. An agent earns its keep where Jobber's built-in rules stop: deciding which quote deserves a call, connecting Jobber to your phones and QuickBooks, and reporting across all of them.

This guide covers what Jobber already does, where the gaps are, exactly how an agent connects, and what should stay with a person.

What Jobber already automates

Start here, because some owners pay for an "AI agent" that duplicates a setting they already have.

Quote and invoice follow-ups. On the Connect plan and up, Jobber can send two automatic quote follow-ups for quotes still awaiting a response, up to 90 days after sending, to clients who have follow-ups turned on. Invoice follow-ups work the same way for past-due invoices. The same Automations page can auto-draft quotes from new requests and notify you about high-value quotes.

AI Receptionist. Jobber's own receptionist answers calls and texts, books visits and takes requests, texts back callers who hang up, and escalates urgent calls. Jobber's pricing page lists it as a $29 a month add-on, or included in the Plus plan. Connect starts at $99 a month billed annually.

Reports and Zapier. Plus-plan admins can describe a custom report and have Jobber AI build it. Jobber also connects to Zapier, which offers triggers such as Quote Sent, Quote Approved and Job Closed, and actions such as Create Client and Create Job.

If two fixed follow-ups and an answering service cover your problem, turn those on first. They're cheaper than anything we'd build.

Where the built-in tools stop

Jobber's automations are rules: same message, same timing, every quote. That's fine for a gutter cleaning. It's not how a good salesperson treats a full patio and retaining wall quote from a client who asked three questions about drainage.

The gaps we see most often in Jobber accounts:

No judgment on follow-up. Two automated nudges, then nothing. Nobody decides which quiet quotes deserve a personal call this week.
Phones live outside Jobber. If you keep your own phone system instead of the AI Receptionist, missed calls never become requests unless someone types them in.
Money lives in QuickBooks. A job closed in Jobber and an unpaid balance in QuickBooks are two screens. Nobody compares them daily.
One view for the owner. Jobber reports on Jobber. Ad spend, call volume and collections are somewhere else.

Small companies feel this most because the same two people answer phones, schedule and bill. The Census Bureau reported that fewer than 20% of firms with four or fewer employees use AI, and a Federal Reserve analysis put AI adoption at about 18% of firms as of year-end 2025. Most Jobber users are not behind their competitors yet.

What an AI agent does inside Jobber

Here is what three common agents watch, read, draft and write back. These are illustrations of the pattern, not a client story.

Agent	Starts when	Reads from Jobber	Drafts	Person approves	Writes back
Quote follow-up	A quote has sat in "awaiting response" for a set number of days	Quote lines, client notes, past jobs, any replies	A personal follow-up or a call reminder, ranked by value	Salesperson sends or edits	A note on the quote
Missed call to request	Your phone system reports an unanswered call	Client lookup by phone number, open requests	A callback text and a draft request	CSR sends, confirms request	New request or client note
Closed job, no payment	A job is closed and the balance is still open in Jobber or QuickBooks	Job, invoice, visit notes	Invoice reminder or a flag for the office	Office manager	Note on the invoice

A fourth agent sends the owner one morning message: new requests, quotes sent and won, jobs completed, money still open. That one reads but never writes.

How an agent connects to Jobber

There is no hidden back door. A legitimate agent uses the same public API as every app in Jobber's App Marketplace.

Authorization. Jobber's API uses OAuth 2.0, not API keys, and a company admin authorizes the app for that account. The app asks for scopes (what it can read or write). Ask for the scope list and refuse anything the job doesn't need.
The API. It's a GraphQL API with a single endpoint. Every request must name an API version; Jobber supports old versions for at least 12 months, so someone has to update the agent when versions retire.
Webhooks. The agent subscribes to events such as client created or invoice updated. Jobber's webhook rules matter in practice: the app must answer within 1 second, the same event can arrive more than once, and the message carries only an ID, so the agent has to fetch the record afterward. A well-built agent checks the signature on every message and ignores duplicates.
Rate limits. Jobber allows 2,500 requests per 5 minutes per app and account, plus a cost budget per query. An agent that re-reads your whole client list every hour will get throttled. Good ones read what changed.
Private vs published apps. A custom app built for your company can stay unpublished. Jobber's docs say a draft app is blocked if it connects to more than 5 paying Jobber accounts, so an agency using one unpublished app across many clients is a warning sign.

Connecting ChatGPT to Jobber

We didn't find a native ChatGPT or Claude connector in Jobber's documentation. The working paths are Zapier (fixed triggers into a ChatGPT step) or a custom app on Jobber's API. Zapier is fine for one step, like summarizing new requests into Slack. It gets fragile once you need memory, judgment across systems or an approval queue.

Setup steps and what you need

A Jobber admin login to authorize the app.
A list of the two or three jobs you want handled, in order. Usually quiet quotes first.
Access to the other systems involved: phone provider, QuickBooks, ad accounts if you want them in the morning report.
One person per agent who approves drafts for the first few weeks.

Expect the first agent to run with a person approving every draft. Loosen it only after you've read enough drafts to trust it. We explain that model in human-in-the-loop AI.

Limits and risks

Duplicates. Because Jobber can send the same event twice, a careless agent texts a customer twice. Ask how they prevent it.
Two systems texting. If Jobber's own follow-ups are on and an agent also follows up, the customer hears from you four times. Pick one per quote stage.
Token and version drift. Access tokens expire and API versions retire. Someone has to own that, which is the reason for managed AI operations.
Mistakes. Jobber's own help pages note that Jobber AI can make mistakes. So can ours. Pricing, discounts and promises to customers stay with a person.

Where Cortex fits

Cortex builds AI agents that work inside Jobber and the other software you run, with a person approving anything that leaves the building. We are not a Jobber partner or a certified app, and we'll tell you when Jobber's own automations are enough. See which systems we connect on integrations and how this works for home service companies. For the bigger picture of agents that read across your whole stack, read the AI operations layer.

The free AI Operations Audit looks at your Jobber account, phones and books in 30 minutes and tells you which job to hand an agent first.

Questions people ask

Does Jobber have AI built in?

Yes. Jobber offers an AI Receptionist add-on, AI-built custom reports on the Plus plan, auto-drafted quotes from requests and automatic quote and invoice follow-ups on Connect and up. Those cover fixed rules inside Jobber; they don't look across your other systems.

How do I connect AI to Jobber?

Either through Zapier, which offers Jobber triggers like Quote Sent and Job Closed, or through an app built on Jobber's GraphQL API that a company admin authorizes. Zapier suits one fixed step; an API app suits agents that need history, judgment and approvals.

Can ChatGPT access my Jobber account?

We found no native ChatGPT connector in Jobber's documentation. ChatGPT can be one step in a Zapier flow, or a custom app can pull Jobber data and pass it to a model with only the access the job needs.

Is it safe to give an AI agent access to Jobber?

It can be when the app asks only for the scopes its job needs, a person approves anything sent to customers, and someone monitors it. Jobber lets an admin authorize or disconnect an app, so access stays under your control.

Will an AI agent conflict with Jobber's own follow-ups?

It can if both are on for the same quotes, and customers then hear from you too often. Decide which handles each stage, for example Jobber's automatic nudges first and the agent's personal follow-up for high-value quotes after that.

Sources

Automations, Jobber Help Center
AI Receptionist, Jobber
Pricing, Jobber
Jobber integrations, Zapier
Getting started, Jobber Developer Center
Setting up webhooks, Jobber Developer Center
API rate limits, Jobber Developer Center
Custom integrations, Jobber Developer Center
Large Firms With at Least 20 Employees Biggest AI Users, U.S. Census Bureau
Monitoring AI Adoption in the U.S. Economy, Federal Reserve Board

The AI operations layer: one brain across your business software

Cortex AI — Wed, 14 Oct 2026 00:02:14 +0000

Originally published at buildcortex.dev.

An AI operations layer is a set of AI agents that sits across a company's field service software, CRM, phones and accounting, and watches for work that falls between them, such as missed calls, unanswered estimates and unbilled jobs. It doesn't replace those tools; it reads their events, flags what didn't happen, and drafts the next step for a person to approve.

An AI operations layer is a set of AI agents that sits across all the software a business runs (field service, CRM, phones, accounting) and watches for work that falls between them. It doesn't replace any of those tools. It reads what each one reports, notices when something that should have happened didn't, and drafts the next step for a person to approve.

For a home-service company, "something that didn't happen" is concrete. A call came in and nobody booked it. An estimate went out and nobody followed up. A job was marked complete and never invoiced. Each of those lives in a different system, which is why no single tool catches them.

AI operations layer vs AIOps

The terms get mixed up in search results, so it is worth separating them.

AIOps is AI for IT operations: monitoring servers, spotting outages, finding the root cause of a failed deploy. It is a tool for IT teams.

An AI operations layer, as we use the term, is about business operations: leads, jobs, invoices, follow-up. It is for owners, office managers and dispatchers. The closest everyday comparison is a very attentive office manager who can see every screen at once.

Why work falls between systems

Most service companies run five to ten pieces of software. A typical stack: ServiceTitan, Housecall Pro or Jobber for jobs and dispatch. A CRM or GoHighLevel for leads and texting. A phone system. QuickBooks for the books. Google Business Profile for reviews. Ad accounts.

Each tool is good at its own job. None of them is responsible for the handoff to the next one.

Larger companies have the same problem at a bigger scale. Okta's 2025 Businesses at Work report found the average company now uses 101 apps. In a MuleSoft survey of IT leaders at large companies, only 29% of their applications were integrated, and 80% said integration problems were slowing them down. If enterprises with IT departments can't connect their tools, a 30-person plumbing company with one office manager has no chance of doing it by hand.

The phone is where it shows first. Invoca analyzed more than 70 million calls and found that only 52% of callers to home services businesses speak with a person. The rest hit voicemail, hold queues or menus. The call log shows the miss. The CRM never hears about it. Nobody calls back.

Why a chatbot in each tool doesn't fix it

Every software vendor is adding AI. That helps inside each product. A field service AI can summarize a job. A CRM AI can draft a text. But each one only sees its own data.

The missed call is in the phone system. The customer's history is in the FSM. The open estimate is in the CRM. The unpaid invoice is in QuickBooks. A chatbot inside any one of them can't see the gap, because the gap is the space between them.

An operations layer reads from all of them. That is the whole idea.

How an AI operations layer works in practice

Under the hood, it works on events and checks.

1. Listen for events. Most modern business software can send a notification the moment something changes. Housecall Pro offers webhooks for more than 50 events, including jobs created and completed, estimates and invoices, though only on its MAX plan. GoHighLevel sends real-time notifications for 50+ events, including new contacts and inbound messages. Twilio, which runs many business phone systems, sends a status callback when each call ends. ServiceTitan's online booking can send an event when a customer leaves contact details but abandons the booking.

2. Compare across systems. The agent holds simple rules a person wrote down. A call ended with no answer and no job was created for that number within an hour. An estimate was sent five days ago and the CRM shows no reply. A job was completed yesterday and has no invoice.

3. Draft, don't act. When a rule trips, the agent drafts the fix: the callback text, the follow-up email, the invoice reminder for the office. A person reviews and sends. What the agent may do alone is a permission you set, one job at a time. We describe that model in human-in-the-loop AI.

4. Report. Once a day, the owner gets one message: what was caught, what was handled, and what is still open.

Newer standards make the plumbing easier. The Model Context Protocol is an open standard for connecting AI models to business tools and data. It also says the host application must get the user's explicit consent before a tool runs, which fits the draft-then-approve approach.

Examples from a home-service week

This is an illustration of how the layer behaves, not a client story.

Monday, 8:10 AM. Three calls rang out over the weekend. The layer matches them against the FSM, finds two with no booking, and puts two callback drafts in front of the CSR.
Tuesday. Eleven large replacement estimates have had no reply for a week. The comfort advisor gets a list with a drafted follow-up for each, written in the company's voice.
Wednesday. A tech closed out a water heater install, but no invoice was created. The office manager gets a flag before the end of the day.
Every morning. The owner gets yesterday's booked jobs, revenue and open items in one text, pulled from the FSM, QuickBooks and the ad accounts.

None of those steps is clever. The value is that they happen every day without someone remembering to check.

Risks and how to control them

Connecting AI to every system raises fair questions. Who can see what? What if it sends something wrong?

Three rules we follow:

Least access. Each agent gets only the records its job needs. The estimate agent doesn't need payroll.
Approval on anything with money or a promise. Pricing, discounts and upset customers stay with people.
A written risk process. NIST's AI Risk Management Framework organizes this into four functions: govern, map, measure, manage. A small business doesn't need a thick policy, but it should know who owns each agent and how failures are caught.

Where the layer runs is a choice too. Some companies keep it in the cloud. Others want it on a private server. We cover that trade-off in private AI for service businesses.

How to start

Small firms are behind on AI use. The Census Bureau found AI use didn't change significantly among firms with fewer than 20 employees from late 2025 into 2026, while larger firms kept adding it. The good news is that an operations layer doesn't have to start big.

Start with one handoff that hurts. Usually it's missed calls or quiet estimates. Connect the two systems involved, write the rule, and run it with a person approving every draft for a few weeks. Add the next handoff once the first one is trusted.

That is how Cortex builds it: AI agents with defined jobs, working inside the software you already pay for. See which systems we evaluate on integrations and why we work inside existing software instead of replacing it. The free AI Operations Audit maps where work drops in your stack and which handoff to fix first.

Questions people ask

What is an AI operations layer?

It is a set of AI agents that reads from all the software a business runs and catches work that falls between systems. It flags what was missed and drafts the next step. The underlying tools stay the same.

Is an AI operations layer the same as AIOps?

No. AIOps is AI for IT teams: monitoring servers and finding the cause of outages. An AI operations layer works on business operations such as leads, jobs, invoices and follow-up.

What is an intelligence layer?

It is a common name for software that sits above other systems and makes sense of their data together. An AI operations layer is one kind: it adds agents that act on what they see, with a person approving anything important.

Do I need to replace my field service software to add an AI layer?

No. The layer connects to the software you already use through its API and webhooks. What it can reach depends on the vendor and sometimes on your plan, which is why an audit of your stack comes first.

Is it safe to connect AI to all my business systems?

It can be, if each agent gets only the access its job needs, a person approves anything involving money or a promise, and someone owns monitoring. Those rules matter more than which model is used.

Sources

Businesses at Work 2025, Okta
2023 Connectivity Benchmark Report, Salesforce / MuleSoft
Home Services Lead Conversion Benchmarks Report 2026, Invoca
How to enable webhooks, Housecall Pro Help Center
HighLevel API documentation, HighLevel
Voice webhooks, Twilio
Set up Scheduling Pro webhooks, ServiceTitan
Model Context Protocol Specification (2025-06-18), Model Context Protocol
AI Risk Management Framework, National Institute of Standards and Technology
Large Firms With at Least 20 Employees Biggest AI Users, U.S. Census Bureau