How to Do Market Research Without a Team
Why solo founders skip research (and why that's expensive)
The most common reason: "I don't have time." The second most common: "I already know the market." Both are usually wrong.
Skipping research doesn't save time — it shifts it. The hours you don't spend understanding the market reappear as months spent building features nobody wants, positioning against the wrong competitors, or pricing below what the market will bear.
The real barrier isn't time. It's that market research feels like it requires a team — analysts, researchers, strategists — and a budget for tools. In 2026, an AI research persona and a structured process eliminate both barriers.
Phase 1: Landscape mapping (2-3 hours)
Competitor identification. Start broad. Ask your AI researcher to map every product that could be considered an alternative to what you're building — not just direct competitors, but adjacent tools, manual workarounds, and incumbent solutions.For a project management tool aimed at architects, competitors include not just other PM tools (Monday, Asana) but also architecture-specific tools (BIM360, Archicad's project features), spreadsheet workflows, and even email-and-folder systems that firms currently use.
Cast the net wide first. You'll narrow later.
Competitor analysis. For each of the top five to eight competitors, document: pricing (all tiers), primary audience (who they target in their messaging), core features (what they emphasise on their homepage), weaknesses (from review sites — G2, Capterra, Trustpilot, Reddit), and funding/team size (Crunchbase, LinkedIn).An AI researcher with web search can compile this in under an hour. Store it as a structured document in your workspace so every persona can reference it.
Market sizing. Rough is fine. You need three numbers: total addressable market (everyone who could theoretically use a product like yours), serviceable addressable market (the segment you can realistically reach), and your beachhead (the specific niche you'll start with).Government statistics, industry reports, and company filings provide the raw data. Your analyst persona can calculate the estimates. For a bootstrapped solo founder, the beachhead number is the only one that matters operationally — the TAM is for investor decks you may never need.
Phase 2: Customer discovery (1-2 weeks)
This is the phase AI cannot do for you, and it's the most important.
Finding people to talk to. Identify where your potential users gather online: industry subreddits, LinkedIn groups, Discord servers, niche forums, Slack communities. Your researcher persona can map these in thirty minutes.Post genuinely in these communities. Not "I'm building X, would you use it?" but "I'm researching how [audience] handles [problem]. Would anyone be open to a 15-minute chat?" The latter gets responses. The former gets ignored.
The interview structure. Ten conversations minimum. Each one: what does their current workflow look like? What frustrates them most? What have they tried and abandoned? What would they pay for a solution? Would they use something like [your concept]?Record the conversations (with permission). Your AI can transcribe and synthesise them later. The patterns that emerge from ten conversations are far more valuable than any amount of desk research.
Review mining. Your researcher persona analyses competitor reviews on G2, Capterra, Trustpilot, and Reddit. The goal: identify repeated complaints, feature requests, and sentiment patterns. What do users love? What do they consistently hate? Where do competitors fall short?This is AI's sweet spot — synthesising hundreds of reviews into actionable patterns. A task that would take a human analyst a full day takes an AI researcher an hour.
Phase 3: Gap identification (2-3 hours)
With landscape data and customer insights in hand, the strategic work begins.
Unmet needs. Cross-reference customer complaints with competitor features. Where are competitors not solving problems that customers consistently raise? These gaps are your opportunities. Underserved segments. Are there customer segments that existing products don't target well? Perhaps every PM tool targets enterprise while small firms are forced to use tools designed for much larger teams. Segment-level gaps often represent the strongest solo founder opportunities because incumbents won't chase small markets. Pricing opportunities. Map competitor pricing against customer willingness to pay from your interviews. Is there a gap between what incumbents charge and what your target segment can afford? Or conversely, is there a premium segment being underserved by cheap, shallow tools?Your strategist persona synthesises all of this into a gap analysis document. With the competitive landscape, customer interviews, and review data already in the workspace as persistent context, the strategist produces a gap analysis that's grounded in specifics, not speculation.
Phase 4: Synthesis and decision (1-2 hours)
The final output is a positioning brief and a go/no-go decision.
The positioning brief: Who exactly this is for. What specific problem it solves. How it's different from the top three alternatives. Why now. What the beachhead market is worth. What the first version must include (and what it must exclude). The go/no-go framework: Proceed if you have: a clear gap in the market, at least three interview subjects who expressed strong interest, a viable path to reaching the beachhead audience, and a differentiation angle that competitors can't easily copy. Pause if any of these are missing.This is where founder judgment matters more than AI analysis. The data doesn't make the decision. You do. But AI-powered research ensures the decision is informed rather than instinctive.
The total investment
| Phase | Time | AI time | Human time |
|---|---|---|---|
| Landscape mapping | 2-3 hours | 80% | 20% |
| Customer discovery | 5-10 hours | 30% | 70% |
| Gap identification | 2-3 hours | 60% | 40% |
| Synthesis | 1-2 hours | 40% | 60% |
| Total | 10-18 hours |
Zerty's research and strategist personas handle the desk research, competitive analysis, review synthesis, and gap identification with shared context — so insights from one phase automatically inform the next. Start your research →
Frequently asked questions
Can AI replace customer interviews? No. AI can analyse existing data — reviews, forum posts, public discussions — but it cannot replicate the depth and nuance of a direct conversation with a potential user. Customer interviews remain the most valuable research activity and the one that AI cannot substitute. How many customer interviews do I need? Ten is the minimum for pattern recognition. By interview six or seven, you'll start hearing the same themes repeated. If you're still hearing new information after ten, keep going until themes converge. What if my market research shows the idea won't work? That's the best possible outcome at the cheapest possible time. Discovering a fatal flaw during a two-week research phase is infinitely better than discovering it after three months of building. Pivot the concept and research again. How often should I repeat market research? A full research cycle every six months is reasonable for early-stage products. Monitor competitors continuously (your researcher persona can do weekly sweeps) and repeat customer interviews whenever you're considering a significant feature or positioning change. What tools do I need for market research? An AI workspace with research and strategy personas, web search capability, and a way to record and transcribe customer interviews (Otter.ai, Grain, or similar). Total cost: under £50/month.Sources
- Harvard Business Review, "To Scale AI Agents Successfully, Think of Them Like Team Members," March 2026 — https://hbr.org/2026/03/to-scale-ai-agents-successfully-think-of-them-like-team-members