Skip to main content

For product builders

Hand engineering a pull request, not a ticket.

Ask the code how your product behaves today, check your specification against it, and try the change in your own running app. Then hand engineering a pull request to review. An engineer still reviews it, and a person merges.

Is it worth it?

Everyone asks for something. Decide with the evidence.

Customers ask in calls, in support tickets, in your app and in interviews. Each ask joins the request it matches, so every request shows who wants it and for which account, what your own teams think of it, and the market signals behind it.

  • Every channel, one list.

    An AI notetaker on your Google Meet, Zoom or Microsoft Teams calls, the feedback widget in your app and, on plans that include support desk import, your Freshdesk or Zendesk conversations.

  • Your teams weigh in.

    A teammate adds their role and their position on a request. A conditional yes records its condition.

  • Ask the customers back.

    Set the questions you want answered, and an AI interviewer emails the customers you choose, follows up on their answers and can switch to a voice call. Each interview comes back summarized.

  • Decisions on the record.

    Ask your team to record a decision when a request changes status, or require one.

Read moreFeature requestsCustomer feedback

Meet Software Inc, a fictional company that sells invoicing software. Every name in these pictures is sample data.

How it works today

Stop waiting for an engineer to tell you how it works.

Open a feature, choose Code in the PM agent panel, and ask. The agent answers from the code, names the files and lines it relies on, and says so when the code doesn't settle a question. When the specification is wrong, ask it to fix it: the correction waits for your review.

  • Every feature, in your words.

    Feature discovery names each feature your code implements in business language, with the business rules its code enforces.

  • It never touches the code.

    A code chat only reads. It never changes your code, starts a preview or opens a pull request.

  • Your questions stay yours.

    Only you can read your code chats, until you draft a plan from one. The corrections they propose appear on the feature, for everyone who can see it.

  • See what moved.

    A dot on the Features list marks each feature that is new or changed since you last looked.

Read moreFeature discoveryCode conversations in the user guide

Sample data

What it should do

Write the spec.
An agent checks it against the code.

Work out a feature's specification with the PM agent beside it. In Review mode, comment on the passages you doubt and send them to an agent: it checks each comment against the code, revises only what your comments point at, and says so when the code can't settle one, such as a product decision. Nothing changes until you apply the revision.

Sample data

  • Written together.

    A feature's specification is one document your team edits together, live, with comments and mentions.

  • Tasks from the spec.

    Generate tasks reads a feature and its context and suggests the tasks that would build it, for you to edit before they are created.

  • What every session reads.

    A coding session starts from the specification and business rules of each feature it works on, so what you agreed is what the agent reads.

  • It keeps up.

    Each session proposes the specification updates its change calls for. Nothing changes until someone approves them.

Read moreSpecs and documentationReview mode in the user guide

Try it

Show the change in real-time, not a mockup.

Produtyping is production-ready prototyping in your applications. Start a Live session from the Dashboard or from a feature's page, and your own app runs from the session's branch in a live preview. Describe a change in plain words, and the preview reloads with it. When you're happy, Open PR turns it into a pull request.

  • Your app, not a copy of it.

    The preview runs your applications from the session's own code, the web client and the API side by side.

  • Nothing you finished is lost.

    Every change the agent finishes is committed and pushed to the session's branch, so stopping the preview keeps it.

  • A bug, straight to a fix.

    Have AI describe a bug report from its screenshots, then start a bug fix session from the report.

  • In our cloud, or on your computer.

    A preview runs in Productmind's cloud, or on your own computer with the Productmind CLI.

Read moreProdutyping and coding sessionsLive preview in the user guide

Sample data

Hand it over

Your engineers review and merge.

Choose Ask for human review and pick the people or the team who should look. Productmind gives them a review task linked to the session and each pull request. They comment on the lines and approve or ask for changes. Approving doesn't merge: a person still decides when it does.

  • Bigger ideas start with a plan.

    Launch a coding session from one request or several. Each becomes a task, and the agent drafts a plan that nothing is built from until someone approves it.

  • Comments go back to the agent.

    Reviewers send their comments to the agent, which works on the same branch and pushes to the same pull request.

  • The risky files, flagged.

    Data migrations, API contracts, sign-in and security code, and CI files are flagged high risk, so reviewers know where to look first.

  • A product review, from our team.

    On the plans that include reviews, ask a Productmind Architect, a person from our team, for a product, code or architecture review of a session with an open pull request, billed by the hour.

Read moreAsking for a reviewFor software developers

Sample data

Close the loop

Tell the customers who asked.

A release lists the features it ships and the sessions that built them. On plans that include lineage, follow it back to the requests and to the customers who asked for each one.

Sample data

  • On your Dashboard.

    Needs your attention lists the customer conversations behind requests promoted into features you own.

  • Release notes, drafted.

    Productmind drafts the release notes from what the release ships.

  • Tell them it's live.

    Turn the release notes into a post for your customers, and Open LinkedIn opens LinkedIn with it filled in.

  • Why are we building this?

    From any feature, its lineage answers the question with the requests and the conversations behind it.

Read moreRelease managerLineage

Just build it.

An idea or a backlog. Either way, you're not building alone.

Get started