Folio Insights

Your clients' documents hold more information than any template can ask for. A tax return has dates, amounts, and filing history. An insurance policy has coverage details, renewal terms, and beneficiaries. Most of it sits in uploaded files, invisible until someone opens them. Folio Insights lets you ask questions of those files in plain language and get answers — plus concrete suggestions for filling in fields you hadn't even thought to ask about.

Insights tab on a folio — chat panel on the left, proposal cards on the right.

Why Use Folio Insights?

Get more out of documents you already have

Without Insights, a folio's documents are passive attachments — useful only if you open them and read them. Insights reads every uploaded file, builds a working summary, and lets you ask questions against the whole folio at once. "When does this policy renew?" "What's the client's filing status?" "Are there any named beneficiaries?" Answers come back in seconds, pulled straight from the documents.

The AI suggests, you decide

Insights never changes your folio on its own. Every suggestion — whether it's updating a field you already have or adding a new one — shows up as a proposal card you explicitly accept or reject. You see what's being proposed, why the AI is suggesting it, and where the information came from. Nothing sneaks into the folio without your approval.

A Field update proposal — from and to values with the supporting reason.

Real-time across your whole team

Multiple managers can have the same folio open at the same time. When one manager asks a question, the other sees the conversation appear live. When one manager accepts a proposal, the field updates and the card disappears for everyone. Insights is collaborative by default — no page refreshes, no stale views.

Choose the right model for the job

Insights runs two separate AI workloads: summarising every file (cheap, runs in the background) and answering your questions (the conversational part). You can configure a different model for each. Pair a fast, inexpensive model for summaries with a stronger model for conversation — or use the same model for both. The choice is yours, per company.

How It Works

1. Open the Insights tab

On any folio, the Insights tab sits alongside the usual detail view. If the folio doesn't have any uploaded documents yet, you'll see a "Waiting for submissions" empty state — Insights needs files to work with.

Empty state — Insights activates as soon as a field or file comes in.

2. Pick a starter question, or ask your own

The tab opens with a few starter chips to get you going — common questions like "Summarise this folio" or "What's missing?". Tap one, or type your own question. Anything you'd ask a colleague who had just finished reading the files, you can ask Insights.

Starter chips above the chat input — quick prompts to get you going.

3. FolioReady reads the documents

Before the AI answers your first question, it summarises every file on the folio in the background. You'll see a progress indicator as each file is processed. Summaries are cached, so subsequent questions on the same folio are answered instantly — FolioReady only re-reads files that have actually changed.

4. Get a reply and proposals

The AI replies in the chat panel. Alongside the reply, if the documents suggest specific field updates, you'll see proposal cards on the right:

Proposal kind	What it does
Field update	Suggests a new value for a field that already exists on the folio
Field add	Suggests adding a field that wasn't on the folio's template

Each proposal shows what's being changed, the supporting evidence from the documents, and two buttons: Accept or Reject.

Proposals panel — two cards: one Field update, one New field.

5. Accept, reject, or keep asking

Accept a proposal and the field is updated immediately, with an audit entry recorded. Reject it and the card is dismissed. You can keep asking follow-up questions — refining the AI's understanding, pushing back on a suggestion, or exploring a different angle on the same documents.

Configuration

Insights uses your company's AI provider setup, with two separate action slots:

Setting	What it controls
Insights model	The model used for conversational replies and proposal generation
Insights synopsis model	The model used to summarise each file in the background

Both can be configured per company under Settings → AI. If you don't set them, Insights falls back to your default AI model. Usage counts against your company's AI usage limit, the same as extraction does.

Tips

Best for folios with multiple documents. The more material the AI has to work with, the more useful its answers become. A folio with a single name-and-address form won't surface as much as a folio with a full client file.
Proposals are audit-logged. Every accepted proposal writes an entry to the folio's event log, so you can always see which fields were changed via Insights and when.
Rejected proposals aren't resurfaced. If you reject a suggestion, the AI won't propose the same change again in the same thread.
The AI is advisory. It's a reading assistant, not an authority. Verify important values against the source documents before acting on them.

AI Automation — Automatic field extraction from uploaded documents when a folio is submitted
Email Ingestion — Collect documents by email, then explore them with Insights