Skip to content
Verras je gebruikersProduct intelligence

AI-copilot in je app

Een assistent in je app die alleen ziet wat de ingelogde gebruiker mag zien.

Een assistent in je app die alleen ziet wat de ingelogde gebruiker mag zien.

0changes the AI can make on its ownAutoconversation history saved15 minbefore a proposal expires
Bekijk het in werking

Productvoorbeeld

Bekijk het in werking

Dit is hetzelfde visuele component als in de ingelogde applicatie, met voorspelbare voorbeelddata.

Scenario
Voorbeeldscherm
Thema van voorbeeldoppervlak
Voorbeeldmodus

Grounded answer

Inspect the real transcript, citations, saved history, and privacy UI with fixture sources.

Inerte voorbeelddata · geen netwerktoegang

Which launch risks still need an owner?

You

Which launch risks still need an owner?

Assistant

Two risks are still open: webhook retry ownership and provider-backed release certification.

Sources

S1
Launch readinessOpen launch risks and their current owners.
S2
Production certificationProvider-backed verification is not complete.
Current project

Enter sends. Shift and Enter start a new line. AI can make mistakes. Check important details.

Voorbeeld gereed

Kernpunten

Waarom dit belangrijk is

It only reads what the user can

The assistant pulls from data the signed-in user already has access to, and nothing more.

Every answer shows its sources

Answers cite where they came from, so users can tell facts from guesses.

It asks before it acts

Any change becomes a proposal the user must approve, and it expires after 15 minutes.

Het probleem dat dit oplost

Zonder een gedeelde basis bouwt elk team deze randgevallen opnieuw — en anders.

The chat box that leaks data

A copilot that queries the database directly can show a user rows they were never allowed to see. One innocent question, and private data crosses the line. Here, retrieval runs through the same permission checks as the rest of the app.

Answers nobody trusts

Answers with no sources look like guesses. One wrong answer and your users write off the whole feature. Here, every answer cites exactly where it came from.

A model with write access

Give a model direct write access and one hallucinated action corrupts real customer data. Here, the model can only propose. Your user approves. Then the normal, safe code path runs.

Hoe het werkt

Het mechanisme van begin tot eind, zoals het in de repository is geïmplementeerd.

  1. 1

    The server decides who's asking

    Identity and workspace come from the login session. Whatever the browser claims is ignored and never placed in prompts.

  2. 2

    Context is gathered with permissions on

    Context providers return only excerpts the user can already reach, each labeled as a source the answer must cite.

  3. 3

    Answers stream with honest states

    You see streaming, complete, stopped, or failed. An interrupted answer stays visibly partial instead of pretending it finished.

  4. 4

    Writes wait for a yes

    A write becomes a saved proposal that expires in 15 minutes. Approval re-checks access, then runs the existing service, so a retry can never run it twice.

Technische garanties

Wat dit wel en niet belooft

Veiligheidsgrens

Achter de schermen houden autorisatie, tenantgrenzen en foutafhandeling één duidelijke eigenaar. Zo blijft de snelle ervaring voor gebruikers gekoppeld aan veilig herstel als het misgaat.

Implementatiebewijs

De grens hierboven is geen claim maar code. Deze bestanden dragen het contract:

  • src/modules/ai-copilot/MODULE.md
  • src/modules/ai-copilot/policy.ts
  • src/registries/commands.ts
Vragen

Veelgestelde vragen

Verwante functies

Deze functies delen contracten en grenzen met deze pagina.

Built-in AI

Add AI features without babysitting messy model output.

Workspace search

Search everything you're allowed to see, and nothing you're not.

Retry-proof saves

Clicking twice never charges twice.

Bouw verder op AI-copilot in je app

De documentatie beschrijft dezelfde contracten die deze pagina demonstreert.

Lees de documentatieBekijk alle functies