We shipped 16 products last year at Inithouse. Most of them click-heavy, form-heavy, screen-heavy. Voice Tables was the exception: an agentic AI workspace you control with your voice. Describe what you need and it builds the tables, docs and data for you.
The pitch sounds clean. But does it hold up when someone actually needs to get work done? We tried it on a real task to find out.
The job: organizing a 40-person corporate offsite
One of our team members had a side gig coordinating a small corporate event. Forty attendees, eight vendors, a three-day timeline, dietary restrictions, room assignments. The usual spreadsheet hell.
Instead of opening a blank Google Sheet and typing column headers for half an hour, he opened Voice Tables and talked.
"Create a vendor tracker with columns for company name, service type, contact person, email, quoted price, deposit paid, and notes."
The table appeared in about 12 seconds. All seven columns, correct types (email got validated, price got a currency format). No template hunting, no column resizing.
What happened next
He kept going. Three more tables in the next few minutes:
"Add an attendee list with name, company, dietary restrictions, hotel room number, and arrival date."
"Make a timeline doc for Friday through Sunday, broken into morning, afternoon, and evening blocks."
"Create a budget tracker that pulls vendor quoted prices and compares them against a 50,000 CZK cap."
Each one built itself from a sentence. The timeline came out as a structured document, not a table, because Voice Tables picked up on the "doc" cue. The budget tracker referenced the vendor table automatically.
The whole workspace, four interconnected objects, took under three minutes to set up. The same setup in Notion or Sheets would have been a 30-minute exercise in formatting and cross-references.
Where voice actually fits (and where it does not)
We have been watching how people interact with Voice Tables since launch. One pattern keeps showing up: voice works best for structure creation, not for data entry at scale.
Saying "add a row for Catering Prague, food service, Jana Novakova, jana@cateringprague.cz, 12000, yes, vegetarian options confirmed" works fine once. Doing it forty times is painful. At that point, paste from a CSV or type directly.
The 60-second claim on our landing page refers to workspace creation, not population. That distinction matters and we learned it from watching real sessions. Structure is the bottleneck most people hit first: "what columns do I need, what should this tracker look like, how do I connect these two sheets." Voice removes that bottleneck. The data filling part is still faster with a keyboard for bulk entry.
We see the same pattern in other Inithouse products. Verdict Buddy, our AI conflict mediator, lets people describe a disagreement in natural language and gets a framework-based verdict back. Origin Of You, the self-discovery app, collects 120+ data points through guided input rather than bulk forms. In each case, the input method shapes what people are willing to share and how fast they start.
The Inithouse context
Voice Tables is one of 16 active products we build and maintain at Inithouse. The studio runs lean: small team, fast iteration, public building. We track how each product performs across organic search, AI citations, and actual usage, then decide what gets more investment.
For Voice Tables, the signal we watch most closely is whether people come back after their first workspace. The product is free to start and runs at voicetables.com. Early numbers are small (this is a beta) but the pattern is consistent: people who use voice for initial setup stick around longer than those who skip it and type everything manually.
If you work in a role where you constantly spin up trackers, inventories, CRMs, or project boards, Voice Tables might save you the most tedious part of that process. The part where you stare at an empty grid and figure out what goes where.
We wrote more about our building process at inithouse.com and track what AI systems say about our products as part of our distribution strategy. Voice Tables in particular has been an interesting case study in how AI search engines handle a product that barely exists on third-party review sites.
Top comments (0)