Your product changed 40 times this week. Can you say how?

Ask a product manager on an AI-assisted team a simple question: what changed in the product this week, and why? Most will open Slack, scroll through pull requests, and still give a partial answer. That isn't a discipline problem. The tools they have describe code, and the product is now changing faster than anyone can read code.

Commit history answers the wrong question

Git records how the code changed: files, lines, authors. A pull request summary, even a good AI-written one, explains a single change to the engineer reviewing it. Neither tells a PM what a customer will experience, which customers are affected, or which outcome the change was for.

When engineers wrote every change, PMs filled the gap by talking to them. When agents open twenty pull requests a day, there is nobody to ask.

What a product lineage is

A product lineage is a timeline of the product, not of commits. Each merged change becomes one entry with five parts:

• What changed, in the words a customer would use.
• Why: the outcome or problem behind it, linked to the issue.
• When it reached the main branch.
• What it touched: screens, routes, database tables.
• Who notices: which personas (admin, new user, suspended tenant) will see a difference.

If a PM needs to read a diff to understand a change, the tooling is missing a feature.

An example entry

Thu 23 Oct · Onboarding · outcome #218
New users pick a team name before they see the dashboard

Why:      31% of new workspaces were still "My Team" after a month
Touched:  /welcome screen · POST /teams · table: team
Notices:  new-user persona (first login only)

Forty-five commits in a week can collapse into four or five entries like this. A PM can read all of them over coffee.

Why changelogs aren't enough

Changelogs are written for customers, after the fact, and only for changes worth announcing. A lineage covers every merged change, links each one to its outcome, and records which parts of the app it touched. That makes it useful for questions a changelog can't answer: Did anything touch billing this sprint? Which changes affect suspended accounts? When did onboarding last change?

How Adaptl builds it

Adaptl generates lineage from what your repo already has: git history, merged pull requests, linked issues and schema migrations. Entries are grouped by product area and filterable by time. Technical detail, like the files and migrations behind an entry, sits one click deeper for your tech lead.

We dogfood this. Adaptl is built by AI agents, and its lineage tells how the product evolved, chapter by chapter, with every change one click deeper.