Sourcegraph · AI product · Mid · 45 to 60 minutes

Choose pricing for Agentic Batch Changes

Sourcegraph builds tools that help software teams search, understand, and change code across many repositories. Answer this product management case and get it graded line by line against a weighted rubric.

pricing strategy
product judgment
analytical thinking
enterprise GTM
experimentation

The situation

Sourcegraph is preparing the first commercial rollout of Agentic Batch Changes, a product area meant to help engineering teams make large code changes across many repositories. The VP of Product asks you to recommend how it should be packaged and priced before the next round of customer conversations and launch planning. The challenge is that the product uses AI, so Sourcegraph will incur usage related costs, but customers will judge the tool by whether important migration work actually gets completed in their codebase. Different pricing units could make the product feel simple, risky, expensive, or hard to justify, depending on the buyer. Enterprise sales wants something easy to explain and forecast, finance wants healthy margins, and product wants a model that nudges teams toward the highest value behavior. You do not need to pick a dollar amount. Your job is to choose the pricing metric and packaging logic Sourcegraph should test first, and explain how the team should validate that choice with early customers.

What to produce

Write a 1 to 2 page pricing recommendation memo.

Your answer must

Name the primary buyer and primary user, and describe the job they are hiring this product to do.
Compare at least three distinct pricing metrics or packaging approaches, with pros and cons for customer trust, value alignment, margin risk, and sales simplicity.
Recommend one launch pricing metric and package structure, including what is included in the base package and what triggers additional charges.
Define the product events or usage records needed to support your model and how a customer would understand their bill.
Propose a short validation plan with 2 to 4 signals to watch in early deals and one fallback if the first model is not working.

Constraints

Use only the context in this prompt and state assumptions explicitly.
The model must fit Sourcegraph's enterprise sales motion and annual contracts.
The billing unit must be observable from product or code hosting events with reasonable accuracy and low manual review.

Format

Use short sections with headings and include one comparison table.

How Sourcegraph actually played it

Sourcegraph explains why it prices Agentic Batch Changes on outcomes, charging only for changesets that merge, in contrast with coding agents that charge for effort per token. The argument is that code is cheap and following through on large migrations is where teams struggle. A clear pricing-model case for AI products where value is measurable.

Consider outcome-based pricing when your product has a clear success event (a merged change).
Teams buying AI tools worry about justifying the bill, so link price to results.
Look at other verticals that already charge per resolution for pricing ideas.
Aim the product at the hard part (finishing migrations), not the easy part (generating pull requests).

Built from Coding agents usually can't price on outcomes. Ours can. by Dan Adler, Sourcegraph.