DEV Community

ibo
ibo

Posted on Originally published at Medium

The Boring Businesses Won

Searches for it fell 71% this year. Here’s what people are searching for instead.

Every list of dying businesses says the same thing. AI is coming for the boring work. Bookkeepers, translators, copywriters, support reps. Learn to prompt or get replaced.

I run a database that pulls business ideas from Reddit complaints and App Store reviews, then checks real search-volume data behind each one. 1,416 scored threads. 192 published ideas. 941 companies with revenue verified straight from Stripe.

When I sorted those 192 ideas by year-over-year search growth, the bottom of the list was not what I expected.

At a glance

  • Searches for “ai writing tool” fell 71% year over year. “ai agent” fell 46%. “ai detector” fell 19%.
  • Meanwhile “fleet management software” rose 50%, “route planner” rose 50%, and “invoice reminder software for contractors” rose 45%.
  • Of 941 Stripe-verified companies, the 37 in Services average $22,457 MRR. The 129 mobile apps average $4,387.
  • The dying category is not boring work. It is the tool layer built on top of a model anyone can call.
  • Demand did not disappear. It moved to the industries nobody wants to write a Medium post about.

Where these numbers come from

Search volume and year-over-year growth come from DataForSEO, the same keyword source most SEO tools resell. Revenue comes from TrustMRR, which reads a company’s actual Stripe account rather than asking the founder what they make.

That second part matters for this piece. Most “here is what’s growing” articles quote founders. Founders round up. Stripe does not.

The categories that are shrinking

  • Keyword: cloud storage

Monthly searches: 60,500

Year over year: -99%

  • Keyword: tax preparation software

Monthly searches: 6,600

Year over year: -75%

  • Keyword: ai writing tool

Monthly searches: 8,100

Year over year: -71%

  • Keyword: church management software

Monthly searches: 4,400

Year over year: -57%

  • Keyword: ai agent

Monthly searches: 18,100

Year over year: -46%

  • Keyword: 3d printing software

Monthly searches: 9,900

Year over year: -46%

  • Keyword: crm for small business

Monthly searches: 4,400

Year over year: -34%

  • Keyword: property management software

Monthly searches: 12,100

Year over year: -34%

  • Keyword: workout planner

Monthly searches: 60,500

Year over year: -33%

  • Keyword: bookkeeping software

Monthly searches: 8,100

Year over year: -33%

  • Keyword: employee monitoring software

Monthly searches: 2,900

Year over year: -33%

  • Keyword: ai detector

Monthly searches: 4,090,000

Year over year: -19%Three of the twelve steepest declines are AI products.

That is worth sitting with. “AI agent” was the most repeated phrase in indie hacker circles for two years running, and search interest is down almost by half. “AI writing tool” lost 71% of its search volume. “AI detector” still gets four million searches a month, which is enormous, but the direction is down.

I want to be careful about what this does and does not prove. Falling search volume for a keyword is not the same as falling revenue in a market. It measures how many people are actively looking for a new one. When a category matures, people stop searching and start renewing. Some of this decline is saturation rather than death.

But saturation is exactly the problem if you are about to start something. A market where nobody is searching for a new provider is a market where you have to take customers from an incumbent instead of catching someone with a problem and a credit card.

Look at “cloud storage” losing 99% of its volume. Nobody stopped storing files. They stopped shopping. Dropbox and Google won and the search behavior evaporated. That is what the end of an opportunity looks like in the data.

The categories that are growing

  • Keyword: family password manager

Monthly searches: 1,300

Year over year: +91%

  • Keyword: legal citation verification tool

Monthly searches: 7,200

Year over year: +89%

  • Keyword: adhd planner

Monthly searches: 6,600

Year over year: +86%

  • Keyword: competitor analysis

Monthly searches: 9,900

Year over year: +83%

  • Keyword: book summary app

Monthly searches: 156,000

Year over year: +67%

  • Keyword: subcontractor insurance tracking software

Monthly searches: 3,900

Year over year: +58%

  • Keyword: second brain app

Monthly searches: 2,500

Year over year: +56%

  • Keyword: customer reorder reminder software

Monthly searches: 4,800

Year over year: +52%

  • Keyword: inventory management software

Monthly searches: 12,100

Year over year: +52%

  • Keyword: fleet management software

Monthly searches: 6,600

Year over year: +50%

  • Keyword: route planner

Monthly searches: 12,100

Year over year: +50%

  • Keyword: invoice reminder software for contractors

Monthly searches: 7,200

Year over year: +45%Read that list again and notice how unglamorous it is.

Subcontractor insurance tracking. Customer reorder reminders. Invoice chasing for contractors. Fleet management. Route planning. These are not businesses anyone starts because they are exciting. They are businesses someone starts because a specific person in a specific trade is losing money every week to a spreadsheet.

There is a second cluster in there worth noticing: family password manager, ADHD planner, second brain, legal citation verification. Every one of those is a consumer or professional tool where the buyer needs to trust the output. A legal citation verifier exists precisely because lawyers got burned filing AI-hallucinated cases. Growth in that keyword is a direct consequence of AI failing in public.

So the honest read is not “AI is dying.” It is that the market moved one layer down. The generic AI tool got commoditized. The thing that checks AI’s work, or the thing that solves a problem so specific no general model has the context for it, is where demand went.

What the revenue data says

Search volume tells you what people want. It does not tell you what they pay for. So I pulled the Stripe-verified numbers on all 941 companies in the database and grouped them.

  • Category: Services

Companies: 37

Average MRR: $22,457

  • Category: Software

Companies: 417

Average MRR: $18,042

  • Category: Consulting

Companies: 23

Average MRR: $13,686

  • Category: Education

Companies: 16

Average MRR: $12,807

  • Category: Media

Companies: 18

Average MRR: $5,103

  • Category: Mobile App

Companies: 129

Average MRR: $4,387Services companies out-earn mobile apps by 5x on average.

This is the least fashionable finding in the whole dataset and probably the most useful. There are 129 mobile apps in there and 37 services businesses, so builders are choosing apps roughly four to one. The services businesses are making five times more each.

The usual caveat applies. Averages hide distribution, and a few large accounts pull the Services number up. But the gap is wide enough and the direction consistent enough that it is hard to explain away.

What I would actually do with this

Stop building the tool layer on top of a model. If your product is a thin wrapper around an API call that a competitor can rebuild in a weekend, the search data says that market is already past its peak. “AI writing tool” down 71% is your warning.

Go where the buyer has a spreadsheet and a deadline. Subcontractor insurance tracking grew 58% because somewhere there is a construction office manager chasing certificates of insurance before a job can start. That person has a budget and no patience.

Sell to trades and logistics. Fleet management, route planning, invoice chasing for contractors. Three separate keywords, all up 45% or more, all in industries where the incumbent software is fifteen years old and looks it.

Build the thing that verifies. Legal citation verification up 89% is a template. Every field where AI produces plausible output that has to be correct now needs a checking layer.

Consider not building software at all. The revenue data says services out-earn apps five to one. That is not the answer anyone wants, which is part of why it is still available.

The part I got wrong

I started this expecting to write the standard piece. AI eats the boring jobs, here are the safe harbors, learn to prompt.

The data would not cooperate. The steepest declines in my set are AI products, and the strongest growth is in trade and logistics software so dull that nobody has bothered to compete properly. The businesses quietly dying in 2026 are not the ones everybody is worried about. They are the ones everybody is building.

Every keyword, growth figure, and revenue number above comes from the Business Ideas DB dataset. The ideas are sourced from real Reddit threads and App Store reviews, scored, and shipped with the competitors already earning money in each niche. You can browse the verified-revenue companies without paying anything.

If you want the version of this with the source thread attached to each idea, that is what the database is.

Top comments (0)