AI-Enabled Entrepreneurship and Small Business Ownership

A qualitative, hypothesis-driven interview study of 50 solo founders, people launching side projects with AI, and small business owners using AI in their existing work. The study looks at how cheap intelligence is changing who runs a small business, how they run it, and what becomes viable to build.

STATUS: Wave 1 complete (N=8) · Wave 2 recruiting · Target N=50

Goal

The study addresses one research question: how are entrepreneurship and small business ownership evolving as generative AI takes hold as the latest General Purpose Technology, and what does the early deployment-phase operator look like?

Each previous general-purpose technology reshaped what it meant to own and run a small business. Electricity gave us the corner store and the small workshop. The personal computer gave us the independent consultant. The internet gave us the e-commerce shopkeeper. The cloud gave us the solo SaaS operator. AI is now doing something comparable to intelligence work, and the early interviews suggest a new kind of small operator is taking shape, both among people starting fresh and among existing small businesses absorbing AI into how they already work.

The interviews document who those operators are, how they run their businesses, how they find customers, where their bottlenecks now sit, and what separates the people who keep going from the people who stop.

About the investigator

Emre Sarbak is the principal investigator. Over the past decade he co-founded four organizations focused on moving people into technology careers, two of them nonprofits: LaunchCode, Kodluyoruz, Patika.dev, and Rise In. Together those programs trained and placed over 5,000 people into technology jobs. He also received NSF SBIR funding to commercialize AI research through Mediate and Supersense, building AI tools for blind and visually impaired users.

The motivation for the study is direct. The bootcamp-to-engineering-job pathway that absorbed those 5,000 people is narrowing, partly because of AI itself. The working thesis is that AI-enabled entrepreneurship and small business ownership is the next version of that wave. Figuring out where the leverage points sit, and who is currently being left out of them, is going to take evidence rather than guesswork. The interviews are that step. What gets built afterward depends on what the data shows.

Methodology

The study is a multi-segment qualitative interview study with pre-registered hypotheses. The full research memo, interview protocol, codebook, and reflection-memo template are versioned in a project repository.

Design specification

Sample. N≈50 semi-structured interviews, 60–90 minutes, across five pre-defined participant segments.
Hypotheses. Eight pre-registered, each with a predicted direction and a specified disconfirming pattern. Frozen before Wave 2.
Pre-interview enrichment. Mandatory briefing from public sources (LinkedIn, product, public archive) before every call.
Coding. Three-pass (open, axial, selective) against a versioned codebook. Only interviewee turns are coded; auto-summaries are not.
Member checking. Each participant receives a personal reflection memo within 48 hours and may correct or redact.
Triangulation. Each subject's interview is cross-referenced against their public archive and a direct evaluation of the product or service.
Saturation. Formal saturation checks at N=15, 25, and 35 per segment.
Critical incident technique. Questions are organized around specific decisive moments rather than general patterns.

The framing draws on a few established lines of work: Bresnahan and Trajtenberg (1995) on what makes a technology a General Purpose Technology; Carlota Perez (2002) on the installation and deployment phases of technological revolutions; Paul David (1990) and Brynjolfsson, Rock and Syverson (2021) on the productivity J-curve; Herbert Simon (1971) and William Ocasio (1997) on attention as a scarce input; and Saras Sarasvathy (2001) on the affordable-loss principle in entrepreneurship. The empirical question is which of these patterns hold for AI-enabled entrepreneurship and small business ownership, and which do not.

Interviews

Wave 2 recruitment is open. Background, geography, stage, and revenue are not screening criteria. Three groups are most relevant:

Solo founders building with AI as the main thing they are working on.
People launching side projects with AI, active or set aside.
Small business owners using AI in their existing work.

Interview protocol, sample items

Walk me through a specific moment when AI was decisive, either making something possible or blocking it.
Where did your last ten customers come from, specifically?
What else do you have going on alongside this one? Walk me through your other projects.
How do you decide, on any given morning, which of your projects to work on?
What is one thing you do now that you could not have done before tools like Claude or Cursor existed?
If someone with no domain experience tried to build what you are building, using the same tools, what would they get wrong?
Who else helped produce the reach you have right now: the algorithm, a community, a specific person, your own identity?
What would you be doing right now if these tools did not exist?

Participant commitment

One video interview of 30 to 45 minutes, scheduled at the participant's convenience.
All quotes are anonymized by default. On-the-record attribution is available on request.

Participate

To take part, email emre.sarbak@gmail.com. Response within one business day.

X/Twitter · LinkedIn · Personal site