From idea to opportunity

Ideas are cheap; validated problems are valuable

Most new ventures do not fail because the founders could not build a product. They fail because they build something not enough people want to pay for. The antidote is to start with a problem, not a solution, and to test your assumptions before investing heavily.

Problem-first thinking

A strong opportunity usually has:

1. A painful, frequent or expensive problem for a specific group of people.
2. People already trying to solve it, often badly (spreadsheets, WhatsApp groups, manual workarounds, paying someone). Workarounds are evidence of demand.
3. Willingness and ability to pay (or a clear way to monetise).
4. A reachable market: you can find and reach these customers affordably.
5. A reason you can win: skills, access, insight or timing that gives you an edge.

The opportunity statement

Write your idea as a testable statement:

For [specific customer segment]
who struggle with [problem, in their words],
our [product/service]
helps them [outcome]
unlike [current alternatives],
because [key differentiator].

Illustrative example: "For small clothing boutiques in Lahore who struggle to manage stock across their shop and Instagram orders, our inventory app helps them avoid overselling and track bestsellers, unlike spreadsheets and notebooks, because it syncs automatically with their social media catalogue."

Every phrase is an assumption to test.

Mapping your assumptions

List the assumptions your idea depends on, then rank them by importance (if wrong, does the idea fail?) and evidence (how much do we actually know?). Test the important, low-evidence assumptions first.

Sources of good ideas

• Your own experience of a problem at work or in life (founder–problem fit).
• Changes in technology, regulation, demographics or behaviour that create new needs, such as the growth of digital payments in Pakistan, e-invoicing requirements in Saudi Arabia, or new AI tools reshaping workflows everywhere.
• Underserved segments: products that work well for large companies but not for small businesses, or for one country but not another.
• Inefficient industries with many manual steps.

Worked example

Illustrative. Zainab, a pharmacist in Karachi, noticed that small pharmacies frequently ran out of fast-moving medicines while holding expired stock of slow movers. Instead of immediately building software, she wrote an opportunity statement, listed assumptions, and visited 25 pharmacies over three weeks. She learned that expiry losses mattered to owners more than stock-outs, that most used paper registers, and that many would consider paying if the tool saved more than it cost. She refocused her idea on expiry tracking and supplier returns, a sharper problem than her original concept.

Is the market big enough?

Estimate market size roughly, without false precision:

• TAM (total addressable market): everyone who could conceivably use the product.
• SAM (serviceable available market): the part you can realistically reach with your model and geography.
• SOM (serviceable obtainable market): what you could plausibly win in the next few years.

Build estimates bottom-up where possible: number of target customers × realistic price × realistic share. For example, if a city has an estimated number of boutiques you can verify from business directories or associations, multiply by a tested price and a conservative adoption rate. Top-down figures from reports ("a $X billion market") are rarely useful for early decisions.

2026 update: AI-native opportunities

Generative AI has created a new class of opportunity, and a new class of trap. The opportunity: work that used to need a team (drafting, research, translation, first-line support, data entry, basic design) can now be delivered by a small team with AI doing the first pass and humans checking the output. The trap: "an app that uses AI" is not a problem statement. A thin wrapper around a general-purpose model is easy to copy, and the model provider can ship your feature as a free update.

Ask four questions of any AI-flavoured idea:

1. Is there a painful, frequent job? The same test as any idea. AI is a how, not a why.
2. What do you have that a general chatbot does not? Proprietary data, a workflow that sits inside the customer's tools, domain rules, distribution, trust or regulatory know-how.
3. Who checks the output, and what happens when it is wrong? In legal, medical, financial or safety-critical work, the checking process is part of the product.
4. Does it still make money when usage grows? Model usage is a variable cost that rises with every task (we cover this in the unit economics lesson).

Hands-on: an AI-assisted opportunity scan

Use an AI assistant (Claude, ChatGPT, Gemini or similar) as a research sparring partner, not an oracle. Paste your notes and ask it to challenge you, then verify anything factual yourself.

You are a sceptical early-stage investor. I will describe an idea.
1. Rewrite it as an opportunity statement: customer segment, problem, current alternative,
   why now, and what would make this 10x better than the alternative.
2. List the 8 riskiest assumptions, grouped as desirability, viability, feasibility.
3. For each assumption, suggest the cheapest test that could DISPROVE it in 2 weeks.
4. Tell me which facts you are unsure about so I can check them myself.
Do not invent market sizes or statistics. If you would need data, say what data.

Idea: [paste your notes, customer quotes and what you already know]

Then fill in the opportunity scorecard below for your top three ideas. Score 1 to 5, and be honest; the point is comparison, not precision.

Opportunity scorecard            Idea A   Idea B   Idea C
Pain (how bad, how often)          _        _        _
Willingness to pay (evidence)      _        _        _
Reachability (can you find them)   _        _        _
Founder fit (skills, network)      _        _        _
Defensibility (data, workflow)     _        _        _
Speed to first test (< 2 weeks)    _        _        _
TOTAL                              _        _        _
Biggest unknown:                   ...      ...      ...

Worked example: from "AI tool" to opportunity

Illustrative. A developer in Islamabad wanted to build "an AI assistant for accountants". Too vague. After ten conversations with small accounting practices, the pattern was specific: every month, staff spent hours re-keying supplier invoices from WhatsApp photos and PDFs into accounting software, and errors caused rework at tax filing time. The opportunity statement became: "Small accounting practices in Pakistan that process supplier invoices for 20 to 100 SME clients lose several staff hours a week to manual data entry from photos and PDFs; current alternatives are manual entry or generic OCR that fails on local formats; a tool that extracts, validates against client ledgers and flags exceptions for human review could cut that time substantially." Every word is testable, and the human-review step is designed in from day one.

Common mistakes

• Falling in love with a solution before understanding the problem.
• Targeting "everyone".
• Treating compliments from friends as validation.
• Using only top-down market sizes.
• Ignoring existing workarounds, which are your real competition.

Quick self-check

Can you name ten specific people who have the problem, describe how they currently solve it and what it costs them? If not, that is your next task, before building anything.

A note on timing

Good ideas often depend on timing. Ask "why now?": what has changed that makes this possible or necessary today? Examples include new regulations, falling technology costs, new payment rails, changes in consumer behaviour, or a new platform opening access to customers.