5.2 Ask an Expert

In Brief

Description

Asking an expert means consulting someone with deep, hard-won experience in your problem domain to shortcut your own learning. Experts are most useful for problems that are complicated rather than simple or chaotic: problems with many factors and components, where the relationships are predictable but not obvious. These problems have many “known unknowns” that cause-and-effect analysis can resolve into a range of workable answers. Founders need to execute quickly, and drawing on someone else’s hard-won knowledge is a fast way to get there.

Some industries require significant expertise to compete. Finance is one example: it is highly regulated and depends on detailed models for valuation, risk assessment, and accounting, all of which are context-sensitive. You can learn some of this from books, but seeing how it plays out in a competitive environment adds insight that reading does not. A founder entering a regulated or technically deep market can use an expert to surface options they would not have considered on their own.

Outside of regulated industries, an expert can still place your situation in the context of many companies that faced a similar problem. If you are planning to test a paid acquisition channel, for instance, an experienced channel specialist can flag the pitfalls before you spend the budget. An LLM can optionally speed up the prelude here — orienting you in the domain’s terminology and common failure modes so you arrive with sharper questions and find the right expert faster — but it is not a substitute for one. It confidently presents outdated information, misses nuances that only come from lived experience, and cannot evaluate your competitive context. A working finance expert knows which regulations are actually enforced versus technically on the books, which investors are currently active, and which approaches have been quietly tried and abandoned; an LLM only knows what has been published.

Use experts to generate options, not to settle them. Asking a third party to evaluate a decision you have already framed tends to backfire: the expert views the situation through assumptions that may not hold up in your data. Expert input sets direction; customer behavior decides.

How to

Prep AI Prompt

1. Find an expert through free channels.
   • Web search: Enter specific terms or questions and look at who is publishing useful content on the topic.
   • Professional networks: Search by keyword, current title, or company to locate working specialists.
   • Local events and groups: Reach out to local experts who present or organize on your topic.
   • Academic publication databases: Find researchers who specialize in a narrow topic and have published in it.
   • Public Q&A sites, blogs, newsletters, and video channels: Long-form public writing or video reveals depth and point of view before you commit to a meeting.
   • Your own sales team: Salespeople talk to customers daily and accumulate insight into objections, feature requests, the competitive landscape, and lost-deal reasons. Treat this as secondary market research, one degree removed from a direct customer interview.
   • Your own customer support team: Support staff hear directly about pain points, frustrations, and unmet needs. Their read on recurring themes can surface problems that don’t show up in the numbers. For quantitative analysis of support ticket data, see Customer Support Analysis.
   • Partners and suppliers: They can verify the feasibility of your approach, reveal hidden dependencies, surface logistics and scalability constraints, and provide cost estimates before you invest fully.
2. Consider paid sources for vetted access. Options include pay-per-minute expert-call platforms with named experts, expert-network platforms ranging from enterprise to founder-priced, and industry-specific forums and trade associations.
3. Reach out and schedule the session. Arrange a call, video meeting, or in-person session. Be specific about how much time you need and the topic you want to cover; vague requests get vague responses.
4. Prepare a written brief before the meeting. Include:
   • Questions you want answered, ranked by priority. Lead with the highest-stakes one in case the call runs short.
   • Areas where you want critical feedback on a draft, decision, or assumption.
   • Topics you want to brainstorm, where you’re looking for options rather than answers.

Before you spend an expert’s hour, use an LLM to sharpen the brief: ask it for terminology you should know, common failure modes in the domain, and the questions a working expert would expect a serious founder to ask. Drafting your own rough model of how the domain works first makes the expert’s corrections far more valuable, because you arrive with a position to correct rather than a blank page.

Analysis AI Prompt

1. Weight experts by the basis of their knowledge. Give the most weight to experts who either have personal experience (success or failure) in the topic area, or have gained significant insight through academic or journalistic research. Discount input that rests on neither.
2. Discount advice for context drift. All advice is context-dependent. Even an expert who succeeded before is describing a situation and competitive landscape that have since changed, so treat their conclusions as dated until you confirm they still apply.
3. Separate first-hand from second-hand claims. Sort what the expert directly experienced from what they read or heard. First-hand claims carry more weight, but both can be wrong for your specific case.
4. Reduce the input to hypotheses, not decisions. End with two or three falsifiable hypotheses you can test with customers. If the experts disagreed, keep the disagreement visible rather than averaging it into a false consensus.