AI Solopreneur TrendDiscover the World's AI Solopreneurs

Nov 6, 2026

What 31 Solo Professionals Hand to AI, and What They Keep for Themselves

Across 20 fields, from physiotherapy to jewellery making, many people who work alone let AI take the paperwork and the first draft, keep client data out of it, and hold on to their own voice, their prices and their judgment.

Over the past few months we have written 20 case studies about people who run their work alone and use AI in some part of it. They include a dietitian in Canberra, a handyman near Tucson, a goldsmith in the Netherlands, a translator in Austria, a plaintiff's lawyer in California and a history teacher in Spain. For each one we checked who they are, whether they really work alone, and what they actually do with AI, using their own writing, trade press and other sources we could open and read.

Their jobs have almost nothing in common. Their use of AI has a lot in common. This article brings together the patterns that came up again and again, and the places where people went in opposite directions.

Every person below has a full case study on this site, linked in the text and listed at the end.

The 31 people at a glance

  • 20 articles, 31 people. Nine articles profile one person; eleven compare two people in the same field.
  • Where they are. 18 are in the United States, 4 in Australia, 2 in Spain, and one each in Austria, France, the Netherlands, Singapore, South Africa and the United Kingdom. One, the writer Kieran Drew, left the UK in 2025 and works while travelling.
  • What they do. Health care (a dietitian, a physiotherapist, a psychologist, a counsellor and a nurse), law, accounting, real estate, farming, wedding photography, interior design, web design, writing and editing, translation, event planning, jewellery making, recruitment training, teaching, handyman work, marketing consulting and digital operations.
  • Which AI tools they name. ChatGPT and Claude come up most often, with seven people each. Heidi, an AI medical scribe, is used by three clinicians, and Aftershoot, which sorts wedding photos, by two photographers. Most other tools are named by only one person.
  • AI built into the software they already use. Several people use AI that arrives as a feature of their trade software rather than as a separate tool: a counsellor's clinical record system, an accountant's practice software, and a landlord's property management platform.
  • AI as the product. For 29 people, AI helps them run their work. For two, a nurse and a teacher, AI is at the centre of the product they built.

1. AI pays off in the paperwork

The clearest pattern is where AI saves time. For many of these people it is not the work clients pay for, but the writing around it: notes, letters, reports and summaries.

  • Clinicians. Dietitian Justin Aniello says his daily admin fell from up to two hours to about 15 to 20 minutes, and that he went from seeing four or five clients a day to 10 to 12. Physiotherapist Gil Jelin stopped doing documentation after hours two weeks after she started using an AI scribe. Psychologist Siew Soon, in the same article, saves up to 45 minutes on each new intake note. Counsellor Tom Schlosser saves about an hour a week on progress notes, which he can use to see another client.
  • Law and accounting. Plaintiff's lawyer Lorraine Gingery cut the time she spends on a demand letter from more than four hours to under 30 minutes. For CPA Sam Leon, building tax workpapers from financial statements went from three to five hours of a person's time to about five minutes of AI, according to an account by Blake Oliver. That is one step of the work: The Accounting Podcast titled its episode with him as using Claude to prepare tax returns five times faster overall.
  • Consulting and recruitment. PPC consultant Lionel Fenestraz timed his work task by task: a monthly report per account went from four hours to one and a half, and he estimates a saving of about 14 hours a week in total. Recruitment trainer Vanessa Raath has AI turn CVs, interview notes and LinkedIn profiles into candidate summaries in under five minutes.
  • Events. Event planner Fiona Nieman uses AI to compare rooming lists against registrations, build food and beverage budgets, and create a run of show for each event.

The exception is wedding photography, where the work AI took on was culling, the job of sorting thousands of photos. John Branch IV lets AI cull overnight and says checking its selection now takes him about five minutes per wedding. Amber Marie, in the same article, started with culling and later moved on to editing and light retouching.

A caution: several of these figures come from case studies published by the companies whose software these people use (Heidi, Anytime AI, CheckpointEHR, Aftershoot). The numbers are the users' own, but the sources are selling something.

Where the saved time goes

Several people put the time back into taking on more clients rather than working less. Justin Aniello more than doubled his daily caseload. Lionel Fenestraz took on three new clients in six months without hiring. Accountant Emma Fabbro says her firm roughly doubled both its revenue and its client capacity over three to four years, without building a staff team and while relying heavily on her accounting software.

2. AI drafts, the person decides, and checks

Many of those who let AI produce something say they review it before it goes anywhere. Emma Fabbro sums up her setup as the AI suggesting the draft and her giving the human approval. Lorraine Gingery reviews and refines every demand letter and makes the strategic calls herself. Lawyer Sergei Tokmakov reviews every AI output before it leaves his firm. Lionel Fenestraz reads every important email, monthly report and campaign brief from start to finish before sending it. Handyman Ray Duke lets AI draft a difficult email, then edits it himself so that he is the last one to touch it.

The checking is not a formality. These people describe specific mistakes:

  • Digital operations partner Sophie Kazandjian was told by Claude that a client's email open rates had gone up by 23% when they had actually fallen.
  • Lionel Fenestraz found that when asked for a current industry statistic, Claude may invent a plausible figure, and that it can get cells wrong across a large spreadsheet, so he does exact calculations in code instead.
  • Interior designer Jacob A Adams found that feeding his 3D models straight into AI made the result too loose for work a builder relies on, so he keeps his own rendering software in the process to control the geometry.
  • Web designer Pravin Kumar finds Claude's design feedback useful but not always correct, and translator Kristal Fellinger stresses that AI's explanations of idioms must be checked.

Our reading, rather than something these people said in so many words, is that their experience is what makes this possible. Someone who has built tax workpapers by hand for years can tell quickly whether an AI-built one is right. Someone new to the work could not.

The reverse pattern: the person writes, AI polishes

Four people work the other way round, at least for their writing. They write first, and AI reshapes what they wrote. Tom Schlosser writes his own session note and uses AI to turn it into a polished progress note. Goldsmith Sarah Kobak drafts every text herself before handing it to AI to rephrase. Kristal Fellinger translates the text herself and, when a sentence sounds clunky, asks AI for five possible translations to combine with her own. Kieran Drew writes his newsletter himself and uses AI to turn it into video scripts and to find what is worth extracting for social media, rather than to write new posts.

3. Client data stays out

Many of those who work with sensitive information have a rule about what goes into an AI tool, and the rules are concrete:

  • Event planner Kristen Collins never uploads a whole contract, attendees' registration details or other proprietary information.
  • Sergei Tokmakov keeps confidential client work on paid plans that do not train on what he enters, and redacts details that identify clients.
  • Vanessa Raath removes sensitive personal information before uploading candidate documents. Kristal Fellinger removes identifying information when needed.
  • Siew Soon's setup keeps identifiable client information separate from session transcripts. Sophie Kazandjian keeps identifiable client information out of the AI tools she uses; her About page says it stays on a local model on her own computer.
  • Editor Olivia Bedford turned off Microsoft Word's "connected experiences" to keep clients' manuscripts away from AI, and published her AI policy on her website.
  • Lionel Fenestraz advises anonymizing personal data before sharing it with a model and telling clients in writing how AI is used.

4. What they keep for themselves

When people describe the line they draw, it falls in a few recurring places.

  • Their own voice. Kieran Drew does not let AI write any part of his newsletter: not the first draft, not the final draft and not the structure. Web designer Eric Karkovack says all of his writing is his own. Ray Duke wants his emails to still sound like him.
  • Their prices. Ray Duke tells other tradespeople that AI cannot tell them what to charge.
  • Their client relationships. Pravin Kumar does not use AI for client emails or project updates, because his relationships depend on genuine communication. Vanessa Raath puts it as AI handling the tasks and humans handling the trust. Fiona Nieman believes no amount of AI can replace the personal touch.
  • Their hands. Jeweller Leto Lama uses AI around her business, but every piece is still made by hand by her in London. Sarah Kobak, in the same article, also makes her work by hand.
  • Their professional judgment. Sam Leon applies his own judgment to every return. Nurse Rachell Dumas built an AI app for patient advocacy whose site states that it does not provide medical diagnoses or advice.
  • Their own understanding. Ray Duke admits that when AI talked him through fixing a video doorbell, he still did not know what he had actually done, and he warns tradespeople to keep learning rather than hand their thinking over to AI.

5. How much they use AI varies widely

There is no single right amount. Among the people whose articles describe how much they use AI:

  • No generative AI. Olivia Bedford uses no generative AI at all, but relies on other kinds of AI tools, such as dictation and text-to-speech.
  • Sparingly. Kristen Collins uses AI sparingly, as a supplement to her own expertise.
  • Occasionally. Eric Karkovack does not use AI every day and turns to it when he is stuck on code. Kristal Fellinger checks in with AI occasionally when she needs ideas.
  • Widely. Sophie Kazandjian, Leto Lama, Fiona Nieman, Sergei Tokmakov and real estate agent Reginald Reglus use AI across many parts of their work.
  • In almost everything. Pravin Kumar has built AI into almost every part of his working day. Kieran Drew uses it for almost everything in his business except the writing itself.

Three of our articles deliberately pair two people in the same job who sit at opposite ends: the event planners, the web designers and the writers. In each pair, both people run a working business.

6. A colleague when you work alone

Working alone means no one to ask, and several people use AI to fill that gap rather than to produce anything.

  • Eric Karkovack compares having AI available to having a colleague on hand around the clock.
  • Farmer Brian Krukewitt treated AI as a collaborator when building his app, asking it after each prompt what questions and suggestions it had.
  • Kristal Fellinger uses AI as a sounding board. Kristen Collins uses it to sense-check her thinking on a contract clause. Kieran Drew relies on it heavily to help him make decisions.
  • Pravin Kumar shares screenshots of work in progress with Claude, which often points out accessibility, hierarchy and spacing problems he has missed. Vanessa Raath has AI ask her questions before it writes a prompt for her.
  • Tom Schlosser notes that in a solo practice no one reviews his notes, so AI takes some of the pressure off making sure he has included everything.

7. When AI changes the shape of the business

For some people, AI did more than speed up existing work.

Building software without writing code. Brian Krukewitt, a farmer in Illinois, built a maintenance app with Rork, a chat-based app builder, after a developer quoted him $5,000 to $10,000. Genevieve Priebe built Reroot, a local food marketplace, without writing a line of code herself, and chose not to look for a technical co-founder. Rachell Dumas built her patient advocacy app HEARD with Lovable. Teacher Julio Candela built Habla con la Historia, where students chat with historical figures; it had 83,820 users from 25 countries between January 1 and June 10, 2024, while he kept teaching.

Selling what they learned. Sergei Tokmakov sells an AI audit and policy package on his site. Reginald Reglus sells a bundle of the prompts and workflows he says he uses every day. Vanessa Raath says she has trained more than 8,000 recruitment professionals in over 100 countries. Ray Duke teaches other tradespeople through his YouTube channel and newsletter.

8. Doubts and costs

These are not uncritical users. Several of them name what AI costs them or others.

  • The environment. Kristen Collins has deep concerns about AI's use of natural resources and says part of her feels guilty when she uses it. Olivia Bedford weighs each tool's environmental cost and keeps only those she thinks justify it.
  • Other people's work. Olivia Bedford does not use AI for work she would otherwise pay a person to do. Eric Karkovack feels guilty about contributing to falling traffic to Stack Overflow and similar communities.
  • Which company to trust. Sophie Kazandjian cancelled her paid ChatGPT subscription after OpenAI's deal with the Pentagon and dropped Perplexity for ethical reasons. She calls her use of Claude, from a US company, a trade-off she accepts.
  • New bottlenecks. Sam Leon says reviewing AI's output, not producing it, is now the slowest step. Sergei Tokmakov has published the limits of his own AI intake assistant, including that only 14 percent of visitors leave an email address and that some problems appeared only with live traffic.
  • Losing skills. Ray Duke tells the doorbell story, above, alongside his warning that tradespeople should keep learning rather than hand their thinking over to AI.

If you want to try this

These are our suggestions, drawn from the patterns above.

  • Start with the writing around your work. Notes, reports, letters and summaries are where most of these people saved the most time.
  • Decide your data rule before you start. Write down what never goes into an AI tool, and tell clients how you use AI.
  • Build in a check you actually do. Read the output end to end, compare numbers against the source, and do exact calculations outside the chat.
  • Choose what stays yours. For many of these people it was their voice, their prices, their client relationships or the craft itself.
  • Use as much or as little as suits you. People at every level, from no generative AI to AI in almost everything, are running working one-person businesses.

How we put this together

Each of the 31 people meets four criteria we check for every case study on this site: they run their work alone, AI plays a real part in it, we can confirm who they are and the facts about their work, and their business is active. For 22 of them, the person or a publication says explicitly that they work alone; for the other 9, we judged this from their own website or other materials, and the individual articles explain the basis.

Most of what we report comes from the people's own writing, trade publications, and case studies published by AI companies. The last group is promotional, and figures such as time saved are the users' own estimates. Where an observation in this article is ours rather than something the people themselves said, we say so.

Sources

All 20 case studies, by field. Each article lists its original sources.

Related articles