Best AI Prompts for Go-to-Market (2026)

Prompts for go-to-market#

Act as a go-to-market strategist for B2B SaaS. Using [PASTE YOUR PRODUCT DESCRIPTION, CURRENT CUSTOMERS AND TEAM SIZE], propose a launch plan for a developer middleware product aimed at enterprise engineering teams. Cover who we sell to first, the motion (product-led, sales-led or both), the first three channels, and the order of launch steps. Label each assumption as "from my data" or "assumed", and give one cheap test per assumption.

Act as a B2B growth strategist. Here are our best five customers and why each bought: [PASTE YOUR CUSTOMER NOTES]. Propose a plan for developer-first adoption through open source, plus one outbound channel. For each channel give the audience, the message, the metric that would show it works within 60 days, and the signal that tells us to stop.

Act as a head of sales. Draft an ideal customer profile (company size, industry, buyer role, trigger event, disqualifiers) from [PASTE YOUR WON AND LOST DEAL NOTES]. Then list the ten objections most likely in a first call and a short, honest answer to each, flagging any where the honest answer is that our product is not a fit.

Act as a pricing analyst. For [DESCRIBE YOUR PRODUCT, COST TO SERVE PER CUSTOMER AND CURRENT PRICE POINTS], compare flat-rate, per-seat, tiered and usage-based pricing. Name the value metric each one implies, who it favours, and what could go wrong. Recommend one package structure as a hypothesis, with three questions to ask buyers to test it.

Act as a product launch lead. Build a 90-day sequence for launching [DESCRIBE YOUR PRODUCT] to [DESCRIBE YOUR AUDIENCE]: what happens before launch (private beta, waitlist, proof), launch week, and the 60 days after. For each phase list the goal, the owner role, the dependency that must be done first, and the one number that decides whether to move on.

What to add to any go-to-market prompt#

• Real customer evidence. Anthropic says to give context or motivation behind instructions so the model can target your goals. For go-to-market that means your actual customers, lost deals and sales notes, wrapped in labelled sections (Anthropic suggests XML tags such as <context> and <input>).
• Constraints. Team size, budget, runway, sales cycle and regions. Google's Gemini guide advises stating constraints, including what to avoid.
• An output shape. Ask for a table or a numbered sequence with owner, metric and date per row. Google lists response format as a prompt component.
• One example of the style you want. Both Anthropic and Google recommend a few examples for format and consistency; a sample row from a plan you liked is enough.
• Separation of data from guesses. Ask the model to mark each claim as drawn from your input or assumed. This keeps invented market facts from reading as findings.
• Request-specific context last. OpenAI advises putting changing context near the end of the prompt and reusable instructions first.

A worked example#

Act as a pricing analyst. For a code-review assistant that runs in a team's CI pipeline, compare flat-rate, per-seat, tiered and usage-based pricing.

Product: GitHub-integrated tool; reviews pull requests in about 90 seconds.
Customers: 14 paying teams of 10 to 60 engineers, all on monthly billing.
Cost to serve: roughly $0.40 of model compute per pull request reviewed.
Current price: $20 per user per month, unlimited reviews.
Observation: 3 of our 14 teams generate about half of all review volume.
Constraint: two-person team; billing must be simple to build.

Name the value metric each option implies, who it favours, and what could go wrong. Recommend one structure as a hypothesis and give three questions to ask buyers to test it. Mark each claim as from my data or assumed.

• It should point out that the current plan puts heavy users at the same price as light ones, using only the figures you supplied.
• Each pricing model should be tied to a value metric (seats, pull requests, repositories) and to who pays more or less.
• It should compute nothing you did not give it, or show the arithmetic so you can check it.
• The buyer questions should be answerable in a call, such as how teams budget for tools, not "would you pay more?"
• Assumptions should be labelled, so you can see what to test.

The full prompt library is on our prompt library page.

Limits of this guide#

This page covers prompting for early-stage SaaS planning: who to sell to, how to position, how to price and in what order to launch. It does not cover paid-media budgeting, which our marketing-plan pages handle. The vendor guides explain how to structure a prompt, not whether a go-to-market plan will work, and we did not test any prompt here. We tried to read a Y Combinator Startup Library page on first customers but could not confirm its content, so we cite no YC claims.