> ## Documentation Index
> Fetch the complete documentation index at: https://support.affinity.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Warm Intros in AI Chat and MCP

> Ask Affinity for the strongest path to a target and draft the outreach — in AI Chat or via the affinity-warm-intro MCP skill.

Private capital runs on warm introductions. Affinity already captures the relationships your firm has built directly — through emails, meetings, and calendars — but until now the *probable* paths one step beyond your direct network stayed invisible.

At launch, you can ask Affinity for a way in **wherever you're already working** — in **AI Chat**, where the Warm Intro agent runs on demand, or through the `affinity-warm-intro` **skill** on the **Affinity MCP** in your own LLM client (Claude, ChatGPT, etc.). You describe the target, Affinity surfaces ranked introduction paths from your firm's relationship data and — when you ask — drafts the outreach, then hands it back so you choose an asker and make the intro yourself. The agent runs when you ask it; it doesn't monitor in the background or send on your behalf.

## What you can do at launch

### Ask for ranked paths to a target

In AI Chat (or via the `affinity-warm-intro` skill on the Affinity MCP), ask in natural language:

* *“Who can introduce me to someone at \[Company]?”*
* *“Do we have a way into \[Company]'s leadership team?”*
* *“Find a path to the founder of \[Company].”*

Affinity identifies the target, returns the paths it can find ranked by relationship strength, and — on request — drafts the ask for you to review and send yourself.

### What powers your paths

1. **Explicit relationships** — the relationships your firm has actually interacted with (emails, meetings, calendar activity), scored by strength. The foundation of every path.
2. **LinkedIn connections** — your team's LinkedIn connections, used alongside explicit relationships to widen who might make an introduction.
3. **Inferred connections** — probable relationships Affinity surfaces even without a direct interaction. Two types at launch:
   * **Overlapping work history** — a teammate likely knows an external person because they worked at the same company at the same time.
   * **Investor ↔ Portfolio Company Executive** — a likely **bridge**: a known contact who was at an investment firm when it funded a target company, and therefore probably knows that company's executives.

> **Example:** Looking for a way to Ben P. (a target-company executive)? Affinity might surface: *“Jeremy Klein might know Ben P. — Jeremy was a Partner at Bain, which invested in Ben's company while Ben was CEO.”*

Both inferred types are available in **AI Chat** and via **MCP**; the Investor ↔ Portfolio Company Executive type also appears in the **Connections tab** (Path Finder) on a company profile.

### Use the `affinity-warm-intro` MCP skill

The Affinity MCP ships with the `affinity-warm-intro` **skill** — a packaged set of instructions and context that helps your LLM client generate warm introductions from your firm's relationship data. Both **AI Chat** and the **MCP skill** surface paths and draft the ask on request, and neither sends on your behalf.

## What's expanding after launch

At launch, Warm Intro surfaces ranked paths and drafts the ask on request. Future direction:

* **Background monitoring** of your open asks, surfacing intros proactively.
* **Intro-request tracking** — a first-class, trackable intro-request object (status, asker, bridge, target).
* **Follow-up drafting** — nudges once an ask is out.
* **Inferred connections in global search.**

## How to get started

1. **In AI Chat:** open AI Chat in the web app and ask a warm-intro question naturally. No setup beyond your org having access.
2. **Via MCP:** connect the Affinity MCP to your LLM client (see [Getting Started with Affinity MCP](/s/article/Getting-started-with-Affinity-MCP)), then ask warm-intro questions or invoke the `affinity-warm-intro` skill.

## Tips for better paths

* **Name the target precisely** — a company plus a role (“the CFO at \[Company]”) anchors better than a company alone.
* **Lead with people your firm actually knows** — explicit relationships with recent, two-way activity make the strongest intros.
* **Check ownership before you act** — if a partner owns the bridge relationship, confirm before requesting the intro.
* **Treat inferred connections as leads, not facts** — “might know” means *probably*; verify before you rely on it.

## Related

* [Warm Intro Agent — overview](/s/article/Warm-Intro-Agent) · [Working with Affinity agents](/s/article/Working-with-Affinity-agents)
* [Getting Started with Affinity MCP](/s/article/Getting-started-with-Affinity-MCP) · [Affinity MCP Reference](/s/article/Affinity-MCP-reference)


## Related topics

- [Affinity Ascend agents — Overview](/s/article/Affinity-Ascend-agents-overview.md)
- [Working with Affinity agents](/s/article/Working-with-Affinity-agents.md)
- [Affinity MCP Reference](/s/article/Affinity-MCP-reference.md)
- [Warm Intro Agent — overview](/s/article/Warm-Intro-Agent.md)
- [Prompts for Affinity AI Chat](/s/article/Prompts-for-Affinity-AI-Chat.md)
- [AI & MCP](/s/article/AI-and-MCP.md)
- [Prompts for Affinity MCP (Claude, ChatGPT & other tools)](/s/article/Prompts-for-Affinity-MCP.md)
- [Set up & configure your agents](/s/article/Set-up-and-configure-your-agents.md)
- [Affinity MCP Skills — the 5 prepackaged capabilities](/s/article/Affinity-MCP-skills.md)