AI connections and engagement knowledge

Overshow captures work context on the consultant's computer. Consultants check and approve what goes into the engagement record, then choose when to publish it with its approved source excerpts. Your consultancy decides who can access each engagement and which AI connections it allows.

Two ways to use knowledge in Microsoft 365 Copilot

Connection	What it can use	Who controls access
The consultant's own Copilot	Permitted source excerpts from the consultant's local history, including content they have not published	The consultant explicitly approves the connection, within the consultancy's permitted AI connections
The organisation's Copilot and approved agents	Published records and approved source excerpts for the engagements they have been granted	The consultancy grants access per engagement and approves the AI connections

An engagement grant does not open a consultant's unpublished history. Organisation agents receive the published record and its approved source excerpts, not the original messages and transcripts behind them.

The personal Copilot connection searches the consultant's computer while it is online and relays returned source excerpts through Overshow's gateway to Microsoft. The organisation connection retrieves permitted published content from the hosted organisation repository. These are separate data paths.

On either route, questions, returned source excerpts, answers and citations are processed and retained in the organisation's Microsoft 365 under its policies. Authorised compliance administrators may access those conversations. Disconnecting prevents new retrieval but does not erase earlier content held by Microsoft. See the Privacy Notice for the gateway and retention details.

Microsoft 365 Copilot needs the relevant Microsoft licence and administrator approval. Standard Teams and Microsoft 365 Copilot set-up is included in the paid 90-day set-up on one engagement. The consultancy arranges licences, permissions and policy approval; additional bespoke systems are agreed separately.

Teams transcripts as a source

Teams transcript import is separate from Copilot access. At set-up, we agree selected eligible Teams meetings and the required Microsoft permissions and administrator consent. Only the selected eligible meetings are included in the import. A saved transcript must exist, and Microsoft's meeting permissions and tenant policies govern whether it can be retrieved and whether it includes speaker names.

Imported transcripts enter the consultant's private history for review. They are not automatically published. The consultant checks and approves the engagement record and source excerpts before publishing them. Microsoft's original transcript remains under its own storage and retention controls.

Overshow drafts the day record on Mac and Windows for the consultant to review. Publishing remains a separate choice. Optional meeting audio capture and on-device transcription are available on the Mac.

Siri on the Mac

Siri connects through a separate bridge. Contact us to arrange its set-up for your Mac. It is not one of the desktop app's named MCP setup presets described below.

Other approved AI clients on the Mac

The rest of this guide describes AI apps, the desktop helper that lets approved AI clients search a consultant's personal history. It ships with the Mac and Windows builds. It follows your agreement with us, carried by the mcp_access entitlement, and requires explicit consent and approval for each client. Your consultancy controls which AI connections are permitted. If the setting is unavailable on your Mac, contact us to check your entitlement.

What MCP is

Model Context Protocol (MCP) connects AI assistants to tools and data. An approved assistant can use Overshow's read-only tools to search permitted source excerpts from your local history and cite them in its answer.

An approved assistant can search bounded snippets from your captured screen text and meeting transcripts, look up confirmed people, organisations and projects, and read stored actions and decisions. With source-detail permission, it can also read profile facts and source excerpts. Watched-folder document indexing is unavailable in this release.

Compared with pasting large raw exports into a chat, tools return bounded, filterable snippets by time, source type, and, for screen results, app.

What MCP lets you do

You want to	The assistant can
Recall what you were working on last Tuesday	Search eligible local text with time and app filters
Find where a topic came up in a meeting	Search confirmed transcript snippets inside eligible meetings
Pull context into an IDE session	Fetch bounded matching snippets and opaque source references

The shared thread: you stay in the assistant you already use, and Overshow becomes one of the grounded sources it can cite.

What the tools cover

The local AI apps integration exposes exactly six read-only tools. Assistants cannot execute arbitrary code against your data:

• search_memory searches eligible screen text and confirmed meeting transcripts, with an optional person, organisation or project filter.
• get_memory_source reads a small source window for an unexpired opaque source id, when you grant source-detail access.
• memory_status reports coarse availability, pairing, scope, and index state without paths, tokens, queries, or content.
• search_profiles finds confirmed people, organisations and projects by name or alias.
• get_profile reads a confirmed profile and eligible stored facts with their provenance, when you grant source-detail access.
• list_moments reads stored actions and decisions, with profile, status, creation-time and deadline filters.

The assistant prepares search terms and explicit filters. These tools read stored data without starting an Overshow language or embedding model. Some stored facts and moments were produced by earlier background processing. It does not expose graph data, questions, summaries, full mail messages, clipboard history, settings, or write operations.

The separate personal Microsoft 365 Copilot connector retains its search and fetch tools; it does not gain the local profile or moments tools.

Personal AI client privacy

The integration is designed around a simple rule: where your query and results end up depends on which assistant you connect.

Assistant type	What leaves your machine
Local LLM (for example, Jan.ai or LM Studio)	Nothing. Queries, tool calls, and responses stay on device.
Cloud assistant (Claude, Cursor, others)	Query text and any tool results returned to the assistant are sent to that provider's servers.

Tool results can include verbatim source excerpts from captured text and transcripts, confirmed profile identities, eligible facts, actions and decisions. Returning them to a cloud assistant sends that content to its provider. Overshow therefore treats cloud assistants as an explicit opt-in, within the consultancy's permitted AI connections:

• A local client (for example a local LLM connected through the custom flow) keeps the entire exchange on your machine.
• Connecting Claude, Cursor, or similar cloud clients requires an explicit opt-in.
• That choice is auditable in the desktop app so you can see what was enabled and when.

The desktop helper connects to Overshow's local server on the same computer. The optional Microsoft 365 Copilot relay described above is a separate connection that carries approved requests and returned excerpts through Overshow's gateway.

How it fits into your machine

When AI apps are enabled, an assistant connects to the packaged local helper on demand and reads only what each tool call returns. The desktop app must be running because the helper never opens the encrypted database itself.

Setup

Connect a client from Settings → AI clients. Overshow gives you a helper configuration to add to your assistant; there is no token to copy. The client appears for approval after it first connects.

Step	What you will do
1. Install Overshow	Install the desktop app and complete onboarding as normal.
2. Turn on AI apps	Turn it on in Settings on your Mac or Windows PC. If the toggle is unavailable there, your account lacks the entitlement.
3. Configure and restart	Copy the generated client configuration and restart the assistant.
4. Approve the pairing	Approve the pending client, choose its history limit, and optionally grant source detail.
5. Ask your first question	The assistant can now search your eligible history and cite returned sources.

Every setup requires explicit Overshow consent and per-client approval. A cloud assistant also sends the query and returned results to its provider. If you approved the earlier three-tool local integration, review the updated disclosure and approve your clients again before access resumes.

What to expect in practice

A few realistic scenarios with the local AI apps tools, once a client is connected:

Scenario	What happens
You ask Claude about yesterday's standup	The assistant calls search_memory for that meeting and replies from the returned snippets.
Cursor needs the cited passage	After a search hit, the assistant calls get_memory_source (if you granted source detail) for a redacted window.
A local LLM helps you write a follow-up	The assistant searches your history for recent mentions of the person or topic and drafts from those snippets.
You check whether AI apps are ready	The assistant calls memory_status to confirm consent, entitlement, and connection state.
You ask who Sarah is	The assistant finds her with search_profiles, then reads eligible facts with get_profile if you granted source detail.
You ask what you owe Acme	The assistant resolves Acme with search_profiles and filters list_moments for related open actions.

In each case the assistant sees only what the tool returns for that specific call, not the entire database.

Settings lets you change an approved client's history limit later. The effective window is the lower of that client limit and your account limit. Privacy settings can additionally redact email addresses, phone numbers and IBANs from client responses; secret, credential and payment-card patterns are always scrubbed. You can also hide an individual person, organisation or project from connected AI tools from that profile's detail page.

Trade-offs worth knowing now

• Tool results are bounded. Each call returns a limited window (for example, a page of search hits or a capped number of events). Assistants may need to call more than once for broad questions, which costs latency and, on cloud assistants, tokens.
• Grounding depends on retrieval. The assistant's answer is only as good as the captures surfaced by the tools it chose. Vague prompts tend to trigger broad searches with mediocre results.
• Cloud assistants can be verbose. Some clients will read large tool outputs into their context repeatedly. If that matters for cost or privacy, prefer a local LLM for day-to-day work and reserve cloud assistants for tasks that clearly need them.
• MCP is still maturing. Clients differ in how they display tool calls, cite sources, and handle errors. Overshow follows the protocol closely, but the experience also depends on the assistant you pair with it.

What it does not do

• It does not give assistants write access to your Overshow data.
• It does not stream captures in real time; it answers on-demand queries.
• It does not replace the desktop Ask feature. Ask stays the primary grounded-answer surface inside Overshow itself.
• It does not bypass capture controls. Paused or excluded content is not indexed, so MCP cannot retrieve it.

Client support

AI apps work with MCP-capable clients. Settings includes named setup presets for Claude Desktop, Claude Code, Cursor, OpenAI CLI, and Perplexity Mac, plus a custom flow for any other MCP-capable client, including local LLMs; the contact page is the best way to tell us which presets to add next.

See also

• Ask: the grounded-answer surface inside Overshow that uses the same local retrieval signals.
• Search: the filters and modes the integration's search tools map onto.
• Privacy: on-device processing, capture controls, and encryption, which together define what the integration can ever see.
• Security: how cloud assistant integrations are treated as an explicit opt-in.