Frequently asked questions

A research engine that starts from a business objective — not a search query — and works backwards to the organisations, functions, and specific people most likely to make it happen. Every recommendation comes with the evidence behind it: what was found, where, and how current it is. It’s available as a web app for human researchers, and as an API/MCP server for AI agents and developers.

A lead database returns everyone who matches a title or filter — hundreds of contacts, ranked by nothing in particular. AI Dealigence starts from your objective, reasons about which organisations and functions are actually relevant to it, then identifies and verifies the specific people who matter, scored on relevance, authority, and access. It’s objective resolution, not contact volume — see Use cases for concrete examples.

Every new account gets one real research investigation free — add a payment method to unlock it and you won’t be charged for that first objective. Beyond that, choose a plan (Starter, Professional, Strategic, or Enterprise) for ongoing objective resolutions — see Pricing for current plans. Separately, the Agent API and MCP server (for developers building AI agents) are metered in credits instead: a full objective run costs more credits than a single verify or evidence lookup, since it involves real web search and reasoning calls, and a request is never billed if it gets rejected for insufficient balance. You can check the cost of a specific objective before running it with POST /v1/objectives/estimate, which returns an estimated credit cost, £ cost, and latency based on recent runs at that depth — it runs nothing and costs nothing itself. See Developers for current developer plans.

Every material claim — that a person holds a given role, that an organisation is relevant to your objective — is backed by a specific evidence item: the claim text, its source (a URL, domain, and source type), when it was published and when we retrieved it, and a strength/currentness rating. We never fabricate evidence or present a guess as a finding. If the evidence is thin, the result says so (an insufficient coverage note) rather than manufacturing a confident-looking answer.

MCP (Model Context Protocol) lets an AI agent — Claude, Cursor, or any MCP-compatible client — call AI Dealigence directly as a tool, without you writing API integration code yourself. We run a live MCP server at aidealigence.com/mcp/ exposing eight tools (resolve_objective, find_decision_maker, resolve_company, resolve_function, verify_person, get_evidence, find_access_route, compare_targets), authenticated the same way as the REST API. See the MCP page for client setup.

We never collapse confidence into one blurred score. A person’s role is reported as one of: verified (strong, current, first-party or corroborated evidence), likely (good evidence, not fully corroborated), uncertain (thin or ambiguous evidence), conflicting (sources disagree — shown to you rather than silently resolved), stale (evidence exists but is old), or not_found (no supporting evidence located). Use POST /v1/verify to re-check a specific person against fresh evidence at any time.

No. POST /v1/access/resolve only returns legitimate, publicly-findable routes to a person — a company contact page, a published email address, a LinkedIn profile, a switchboard, a relevant department — never a guessed personal email address. This is a hard product principle, not a configurable option.

Yes — POST /v1/objectives/{id}/cancel stops a running objective for real: it halts the live research task and its ongoing spend, not just a status flag while the work continues unseen. POST /v1/objectives/{id}/refresh re-runs a finished or cancelled objective from scratch for current results.

Email support@aidealigence.com. For API/MCP integration questions, start with /docs or the Playground.

Decision-maker intelligence is identifying, for a specific objective, who inside an organisation actually owns that decision today — not just who holds a plausible- sounding title. It combines organisation relevance, functional ownership, current authority, and a verified, current role, all backed by evidence, rather than a static title-matched contact list.

Yes. AI Dealigence runs a live MCP server at aidealigence.com/mcp/, so any MCP-compatible agent (Claude, Cursor, LangGraph, CrewAI, or a custom agent) can call it directly as a tool — no separate integration code to write beyond configuring the server URL and an API key. See the MCP page for the full tool list and setup.