Folios

A folio is a subject-scoped workspace. Every folio bundles four things around one topic:

A chat spine - the long-running conversation that grounds everything else
Watches - bills, members, committees, or agencies the folio tracks
Briefings - scheduled digests scoped to this folio
Drafts - working documents the chat has produced for this folio
Folio memory - facts saved during this folio's conversations (positions, strategy, key contacts)

Where an ordinary chat is one-and-done, a folio accumulates context. Open it next week and the prior conversation, saved memories, watch deltas, and recent briefings are all in one place.

When to use a folio

Create a folio when the work is going to come back. Good candidates:

A bill moving through markup, conference, and final passage
An agency rulemaking with multiple comment periods
An appropriations cycle for a specific account
A coalition campaign with a fixed set of stakeholders
A hearing series the office is following

For one-off research ("what's happening on AI this week"), use a regular chat instead. Folios are for subjects you keep returning to.

Creating a folio

Folios are created conversationally. From any chat, ask:

"Make this a folio for the FTC noncompete rule"

"Start a folio tracking HR 1234"

"Open a folio for the FY27 Energy and Water appropriations cycle"

Apogee creates the workspace, attaches the current conversation as the folio's spine, and prompts for a short "about" description that grounds future answers in the folio's scope.

How a folio uses its context

Every time a folio opens, Apogee checks what has changed since the last visit:

New activity on any watched entity (bill moved, hearing scheduled, lobbying filed, news coverage)
New briefings delivered to the folio
Drafts saved during the last session

When more than six hours have passed since the last visit, the chat opens with a delta summary: a bolded one-liner plus three to six bullet points covering the most consequential changes. From there the spine chat resumes - follow-up questions continue the conversation from where it left off.

Adding watches

To track a specific entity in the folio, ask the chat:

"Watch the Senate Environment and Public Works Committee"

"Track Senator Smith on energy bills"

"Add HR 4567 to this folio"

Watches show in the folio's right rail and feed into the next visit-triggered delta and any briefings scoped to the folio. Removing a watch is also conversational: "stop watching HR 4567."

Briefings scoped to a folio

Briefings created inside a folio automatically inherit its scope. Ask the chat:

"Set up a weekly briefing for this folio covering bill movement and lobbying activity"

The briefing runs on the schedule chosen, pulls only the entities the folio watches, and lands in the folio's briefings list (and in the inbox if email delivery is enabled). General-purpose briefings created outside a folio stay user-wide and unaffected.

For the full briefings configuration model, see Briefings.

Folio memory

When the chat saves a fact during a folio conversation - "lead opposition is the Chamber of Commerce" or "position is conditional support if the data threshold is raised" - it scopes to that folio automatically. These show on the folio's right rail and inform every future conversation in this workspace, without polluting other folios.

This is the distinction between folio memory and personal or org memory: role, organization, and mission show up everywhere; folio memories only surface when the user is inside that folio. For the broader memory model, see Memory.

Drafts inside a folio

Drafts the chat produces during a folio session - bill comparisons, comment-letter outlines, talking points, oversight memos - save to the folio's drafts list. Each draft includes the source conversation, so the reasoning trail stays attached. Drafts can be iterated on by returning to the spine chat and asking for revisions.

For the broader drafts model, see Deliverables.

Example folios by role

Congressional staff

One folio per priority bill in the portfolio
A folio for the upcoming appropriations cycle, with watches on the relevant accounts and subcommittees
A folio for each major hearing series the member chairs

Government affairs

One folio per active rulemaking the team is commenting on
A folio per bill the trade association is lobbying
A folio for a coalition campaign with a fixed set of allies and a shared draft library

Advocacy organizations

One folio per issue area the organization organizes around
A folio per state legislative campaign
A folio per federal bill being pushed, with stakeholder memory and member outreach drafts

Journalists and researchers

One folio per running story
A folio per congressional investigation being followed
A folio per policy area with watches on the relevant committees and agencies

Managing folios

From the folios list in the sidebar:

Open a folio to resume its spine chat
Archive folios that are no longer active - history is preserved, but the folio drops out of the active list
Rename or edit the about description to refresh the workspace context

All of this also works conversationally: "archive my noncompete folio" or "rename this folio to FY27 Energy."

Tips for getting the most out of folios

One folio per subject, not per question. A folio works best when its scope is stable. "FY27 appropriations" is a folio; "what amendments were offered yesterday" is a question inside that folio.
Add the about description early. The about line grounds every future answer in the folio. A vague about leads to vaguer answers.
Let watches accumulate. Add entities as they become relevant. The visit-triggered delta is more useful when the watch set reflects the current state of the work.
Use folio memory for positions and strategy. Anything that should bias every future answer in this folio (your position, the key opposition, the timeline you're working against) belongs in folio memory, not in a one-off message.
Archive promptly. Active folios should reflect active work. Archive when a bill dies, a rule finalizes, or a campaign ends - the history stays searchable.

Briefings - Scheduled intelligence reports, available standalone or scoped to a folio
Memory - The broader memory model: personal, org, and folio-scoped layers
Deliverables - Drafts and structured documents the chat produces