We asked 89 founders and operators a plain question: which AI has actually become part of how you run your company, what do you use each one for, and which did you try and then drop, and why.
The answer was close to unanimous, and it cuts against the marketing. AI earns a permanent spot when it removes routine load and sits close to your own data and workflow. It gets cut, fast, when it tries to replace judgment or puts distance between you and the customer.
What the operators said
The kept stack is unglamorous: first-draft engines for internal text, orchestration that routes leads and removes human lag, meeting capture that turns talk into tasks, research that replaces fifteen open browser tabs.
The dropped pile is remarkably consistent: generic copy generators that cost more to edit than to write, autonomous customer-facing bots that erode trust on contact, and bulk content engines that AI search quietly ignores.
The filter most operators use: a hard utility bar (one cut a dozen experiments to three with a "5 hours a week or it is gone" rule) and problem-first buying (another audited 12 companies that each wasted six figures on software they barely touched).
The signal worth watching: AI search is breaking the dashboards, sending customers to a booking with zero website visit, so operators are starting to track revenue, not clicks.
Methodology: 89 operator submissions, 75 usable after dedup. The pool skews toward marketing, agency, and digital-services operators, and that skew is disclosed in the report.
PRAPI runs these research briefs in the open: a real question, real operator submissions, named citations. More at prapi.dev/research.