Sixteen Agents

Brief looks like a single assistant. Behind it runs a team of sixteen background agents, each responsible for keeping one part of your product picture current.

August 12, 2026 6 min read

When you use Brief, you are talking to one thing. Behind that one surface runs a team of sixteen agents, each a background worker with a single job: to keep one part of your product picture current, drawn from the tools you already use. Most people meet two or three of them and never see the rest of the team. This is the whole roster, and what each one is for.

These are not chat. Chat answers a question while you wait. An agent runs on its own, on a schedule or when something changes in a connected tool, and keeps one strategic primitive up to date: your personas, your decisions, your delivery speed, your competitors. You do not prompt it. It watches a source, synthesizes what changed, and writes the result back where the rest of Brief, and any coding agent reading from Brief, can use it. Think of it less as an assistant and more as a standing team, each member responsible for one slice of what a product organization has to keep track of.

The ones that track your work

• Work Agent maps everything in flight, from an idea you are only considering through to shipped, out of Linear, GitHub, and the rest of your delivery tools.
• Feature Agent keeps the list of what you are actually building, pulled from that work pipeline, your docs, and your website.
• Velocity Agent computes how fast you are delivering from the work items themselves, so throughput is measured rather than guessed.
• Decisions Agent finds the strategic decisions scattered across Linear, GitHub, docs, and Slack, the calls that usually live in a thread nobody will find again.

The ones that track your users and market

• Signal Agent extracts research signals and themes from calls, documents, and research sources, the raw material of what your users are actually telling you.
• Persona Agent turns those signals into user personas, then waits for you to review them.
• Behavior Agent pulls product analytics, today from PostHog, so what users do sits next to what they say.
• Competitor Agent enriches your competitor set, crawls their pages, and flags when one of them drifts.
• Positioning Agent proposes how you are positioned, drawing on competitors, deals, and signals together.

The ones that track the money

• Revenue Agent syncs billing and subscriptions from Stripe.
• Pipeline Agent syncs deals from your CRM, starting with HubSpot.

The one that holds the strategy

• Strategy Agent pulls your vision, goals, and OKRs out of the documents where they live, so the top of the plan sits in the same picture as the work at the bottom.

The ones that synthesize and speak

• CPO Agent reads across every domain above and surfaces the cross-cutting insight no single agent would catch on its own.
• Weekly Brief Agent synthesizes the whole picture into a strategic weekly brief, the one read that tells you where things actually stand.
• Proactive Brief decides when Brief should speak first rather than wait to be asked: a decision about to expire, a sudden surge in a signal, a new fact that contradicts an old call.
• Entity Resolution Agent links every one of these objects across sources, so a persona, a decision, a feature, and a deal that are about the same thing are recognized as the same thing, and the whole becomes a graph rather than a pile of separate lists.

Fourteen of these you can open and inspect today inside Brief. Two of them, Proactive Brief and Entity Resolution, run further back in the system, doing their work without a page you visit, which is rather the point of them.

Which calls the system makes, and which it leaves to you

There is a distinction worth naming here, because it is a design choice and not an accident. Some of these agents write directly. The ones that deal in observed fact, Revenue, Pipeline, and Velocity, do: a subscription exists or it does not, a deal is at a stage or it is not, so the agent simply keeps the record true. The ones that deal in judgment do not get to decide. Strategy, Positioning, Persona, and Feature deal in exactly that, and judgment is yours. They propose, and the suggestion sits in a review queue until you accept it, edit it, or throw it out. The system knows the difference between a number it can read and a call that is yours to make, and it never quietly makes the second kind on your behalf.

That is the thing worth seeing. Brief presents as a single assistant, but underneath it is a team, each member keeping one slice of the product picture current and correct, from the tools you are already using, so that the picture as a whole stays live instead of aging into a document nobody trusts. It is also what a coding agent reads from when it asks Brief what the team has decided. So the question is less how many agents there are, and more what it would change to have every part of your product picture kept current by something whose only job is to keep it current: what would you stop having to reconstruct by hand?

Frequently asked questions

What are Brief's agents? Brief's agents are background AI workers, sixteen of them, that each keep one part of your product picture current. Unlike chat, which answers a question while you wait, an agent runs on a schedule or when a connected tool changes, synthesizes what is new, and writes the result back into Brief. Each one owns a single primitive, such as personas, decisions, delivery velocity, competitors, or revenue.

How many agents does Brief have, and are they all visible? Sixteen. Fourteen are visible in the product today, where you can open and inspect each one: Behavior, Competitor, CPO, Decisions, Feature, Persona, Pipeline, Positioning, Revenue, Signal, Strategy, Velocity, Weekly Brief, and Work. Two more run behind the scenes, without a page you visit: Proactive Brief and Entity Resolution.

Do Brief's agents change my data automatically? It depends on the kind of data. Agents that deal in observed fact, such as Revenue, Pipeline, and Velocity, write directly, because there is no judgment to make. Agents that deal in strategy, such as Strategy, Positioning, Persona, and Feature, only propose: their suggestions wait in a review queue until you accept, edit, or reject them.

Where do Brief's agents get their data? From the tools you already use. Work and Decisions draw on Linear, GitHub, docs, and Slack; Revenue on Stripe; Pipeline on your CRM, starting with HubSpot; Behavior on product analytics such as PostHog; Signal on calls and research sources; Competitor on competitors' own pages. Each agent watches its sources and keeps its slice of Brief current as they change.