> the help center.
Learn how to use Unstuck Engine — from setting up your first ICP to pushing audiences to your destinations.
Tasks
> I want to...
- Build an audience
- Build an audience from yesterday's signals
- Build my first ICP
- Build my first ICP
- Build my first Persona
- Build my first Persona
- Configure field mapping
- Connect HubSpot or Salesforce
- Create a list of leads
- Drill into a signal's detail
- Enrich a lead
- Enrich an entire audience
- Enrich leads from a signal
- Enrich one lead from its detail page
- Enrich via API
- Generate a one-pager for this lead
- Improve an existing ICP
- Look up a term
- Manage API keys
- Manage billing
- Manage signal sources
- Manage team and engagement stages
- Push an audience to HubSpot
- Read a lead detail page
- Read a record's ICP and Persona scores
- Read an account detail page
- Read composition and Reachable %
- Read enrichment results
- Read the Insights dashboard
- Read the Qualification panel on a lead
- Read the signals table
- Run bulk actions on records
- See all 31 signal types
- Set up Unstuck Engine in 10 minutes
- Take a 3-minute tour of the product
- Turn on autoqualify
- Understand AND/OR nesting in the audience builder
- Understand ICP vs Persona
- Understand audiences
- Understand enrichment
- Understand parameter modes
- Understand signal categories
- Understand signals
- Understand the Insights page
- Understand the Signals panel on a lead
- Understand what Copilot does
- Use Unstuck Engine via MCP or CLI
- Use the Accounts page
- Use the Enrichment API
- Use the Leads page
- Use the Signals API
Section 01
Start with Copilot
Copilot is your platform-setup assistant — not just a chatbot. Learn what it can do across signals, ICPs, audiences, and records.
What Copilot can do — your platform-setup assistant
Copilot is not a chatbot bolted onto the side of Unstuck Engine — it's the primary surface for setting up the platform, finding leads, and acting on records.
Setting up your ICPs with Copilot
Two paths to define an ICP — describe your customer to Copilot in plain English, or open the Qualification page and build segments by hand. Copilot is faster; the manual path is for surgical edits.
Setting up your Personas with Copilot
Personas are person-level targeting inside each ICP — who at the company you actually want to reach. Copilot drafts the full Persona from a one-line description; the Qualification page handles surgical tuning.
Building audiences with Copilot
Audiences are dynamic, auto-updating segments — Copilot turns a sentence ("VPs of marketing at Series B SaaS who replied to a demo invite in the last 14 days") into a multi-condition filter, ranked and ready to push.
Copilot actions inside a record — research, write, summarize
Open a Lead or Account and Copilot sees what's on screen. Ask for a one-pager, talking points, a tailored email, or a summary of the signals that flagged this record — all grounded in the workspace data.
Section 02
Getting Started
First-day orientation: what each surface does, what to set up first, the vocabulary you need.
The 3-minute orientation
A whirlwind tour of Unstuck Engine — Copilot at the top, Signals feeding the pipeline, Qualification scoring records, Audiences pushing to destinations. Three minutes to understand the shape.
Your first 10 minutes
A starter checklist. Define one ICP, one Persona, turn on three signals, build one audience, connect one destination. Copilot does most of it; the manual steps take seconds.
Glossary — signal, ICP, persona, audience, score
Every term Unstuck Engine uses, defined once, plain English. Refer back when something in another article assumes you know what an Engagement Score or Parameter Mode means.
Section 03
Signals
Intent signals from first-, second-, and zero-party sources. Configuration, scoring, autoqualify.
What signals are and why they matter
Signals are behavioral and firmographic indicators of buyer intent. Every record in your workspace exists because a signal flagged it. Understanding signals is the foundation of everything else.
Signal categories — 1st-party, 2nd-party, zero-party
Every signal in Unstuck Engine belongs to one of three categories defined by where the data came from. The category shapes which sources you can hook in, how the data lands, and what privacy posture applies.
The 31 signal types — a tour
Every signal type Unstuck Engine ships, grouped by what they tell you about a buyer. Use this as a reference when deciding which signals to turn on for a given GTM motion.
Configuring autoqualify on a signal
Autoqualify runs every captured record through your ICP and Persona rules automatically. With it on, signals produce qualified leads ready for outbound. With it off, you see the captures but the scoring waits for a manual run.
Reading the signals table — column by column
The Signals page is a column-heavy table because it has to answer every question about your signal portfolio in one view — status, capture rate, qualification rate, downstream outcomes, credit consumption. This guide reads it left to right.
Signal detail view — drilling in
Click any row on the Signals table and you land on that signal's detail page. The detail view is where you configure the source, set per-rule filters, see the live capture feed, and audit which records the signal has flagged.
Section 04
ICPs & Personas
Company-level and person-level targeting. Parameter modes, scoring, and the Copilot-driven setup flow.
ICP vs Persona — the difference and the relationship
ICPs target companies. Personas target people inside those companies. The relationship is hierarchical — every Persona belongs to exactly one ICP. The scoring is parallel — every record gets both an ICP score and a Persona score.
Building an ICP — Copilot first, then manual
The complete reference for defining an ICP. Two paths — Copilot drafts from a sentence; the Qualification page handles surgical edits. Both write to the same definition.
Building a Persona — Copilot first, then manual
Personas are person-level targeting inside an ICP. The full reference for defining one — title patterns, seniority, function, tenure, custom questions, person-level signals. Copilot drafts; the drawer handles fine tuning.
Parameter modes — Filter, Score, Exclude
Every parameter in an ICP or Persona operates in one of three modes. Picking the right mode is the single highest-leverage decision in setup. This is the reference.
Reading ICP and Persona scores on a record
ICP and Persona scores are 0–100. Each score is the weighted Score-parameter contribution, given the record passes all Filters and Excludes. Click into the score on a lead detail page to see per-parameter contribution.
Refining ICPs with Copilot suggestions
After 2-4 weeks of outcomes, your ICPs should evolve. Copilot reads your historical reply / meeting / signup / purchase data and proposes specific edits — tighten this parameter, drop that one, add this new Score field.
Section 05
Audiences
Dynamic, auto-updating segments built from ICP, persona, signal, and engagement parameters.
What audiences are and how they auto-update
Audiences are saved filters that update themselves. Define the conditions once; every new record that matches becomes a member, every existing member that stops matching drops out. Destinations sync follows automatically.
Building an audience — Copilot first, then manual
Two paths to define an audience. Copilot translates a sentence to a multi-condition filter. The manual builder gives you fine control over AND/OR nesting and destination field mapping.
Conditions and groups — the audience builder
The audience builder uses groups + conditions with explicit AND / OR logic. This is the reference for nesting them correctly when one condition would be ambiguous.
Audience composition and Reachable %
Two columns on the Audiences page that catch operators off-guard. Composition tells you what the audience is made of by ICP/Persona/source mix. Reachable % tells you how much of it you can actually contact.
Pushing an audience to a destination
Connecting an audience to a destination is a per-audience configuration. Field mapping, sync cadence, overwrite policy — set once, then the audience updates the destination automatically as members change.
Section 06
Leads & Accounts (Records)
Where operators spend most of their time. Lead detail, account detail, and the panel-by-panel anatomy.
The Leads page — overview and bulk actions
The Leads page is the master list of every person in your workspace. Identity-first columns, server-paginated, bulk-actionable. Where you live when triaging the pipeline.
The Accounts page — overview and bulk actions
The company-level counterpart to Leads. Every account in your workspace, scored against ICPs (companies, not people), with rollups of qualified Personas inside each.
Lists — creating, sharing, automating
Lists are static collections of leads or accounts you build manually (or import). Different from audiences — lists don't auto-update. Useful for one-off campaigns, event attendee lists, and conference targets.
Lead detail page — the anatomy
The full panel-by-panel tour. Header, four state panels (Activity / Qualification / Enrichment / Sync), Detailed Profile, Copilot rail. Where most of the operator's day happens.
Lead detail — Signals panel
The Activity panel on a lead detail shows the signals that flagged this person into your workspace. Click through any entry to open the underlying signal source, see the rule that fired, and trace attribution.
Lead detail — Qualification panel
ICP score, Persona score, Stage chip. Click any of the three for the breakdown. Re-qualify from here when you want to confirm scoring after editing an ICP / Persona definition.
Lead detail — Enrichment panel
The Enrichment panel shows the lead's current enrichment status and the `Enrich Profile` button to run the waterfall on demand. Per-channel results (work email, personal email, phone, LinkedIn) appear in the Detailed Profile.
Account detail — anatomy and key differences from Lead
The company-level detail page. Same shape as a lead detail — header, four panels, Detailed Profile, Copilot rail — but the panels carry account-level signals (funding, hiring, M&A) and the Detailed Profile is firmographic instead of person-level.
Bulk operations across records
Re-qualify, re-enrich, push to destination, add to audience, export, delete. Bulk operations work on selected rows or all rows matching a filter. The bulk bar at the bottom of Leads / Accounts makes the scope explicit.
Section 07
Enrichment
Reply-rate-optimized waterfall enrichment, invocable from signals, audiences, record detail, and the API.
What enrichment is and the waterfall model
Enrichment fills in the contact details (email, phone, LinkedIn) that turn a captured record into an actionable lead. The waterfall picks providers in reply-rate-optimized order per workspace, so the contacts that land convert.
Enriching from a Lead or Account detail page
The most common entry point. Open a lead or account, click `Enrich`, the waterfall fills missing channels. Use this when you're triaging a specific record and need contact details before reaching out.
Enriching from the Signals page
From the Signals page, enrich the records a signal flagged. Useful for triaging a single signal's output — want contact details for everyone the Keywords rule fired on yesterday? Run enrichment scoped to that signal.
Enriching from Audiences (bulk)
Audiences have their own enrichment scope (which channels to enrich). The bulk-enrich action on an audience runs the waterfall across every member for the audience's configured channels. Used at audience-launch time or when expanding scope.
Enriching via API or MCP
The Enrichment API and MCP tool let you trigger enrichment from outside the app — from n8n workflows, CLI scripts, AI agents. Same waterfall, same credit accounting, programmatic invocation.
Reading enrichment results and provider attribution
Every enriched channel carries metadata — which provider supplied it, when, with what confidence. Use this to debug bouncing emails, audit data quality, and tune the waterfall.
Section 08
Insights
Performance metrics by ICP x Persona x Engagement.
What Insights tell you
Insights closes the feedback loop. Reply rate, meeting rate, signup rate, purchase rate, and LTV by ICP × Persona × Engagement × Signal source. The page where you decide what's working and what to change.
Reading the Insights dashboard
A practical tour of the four views — Overview, by ICP, by Signal source, by Engagement Stage. Where to look first, what to click into next, what to ignore.
Section 09
Destinations & API
Push to destinations (HubSpot, Salesforce, Apollo, more) or consume the Signals, Qualification, and Enrichment APIs.
Connecting a destination — HubSpot, Salesforce, Apollo
First-time setup for the four most-common destinations. OAuth, scope selection, default mapping, sync test. After this, any audience can push to the connected destination.
Field mapping
How Unstuck Engine fields translate to destination fields. Defaults handle the common cases (email, name, company); custom fields handle the workspace-specific cases (ICP code, engagement stage, signals).
Using the Signals API
Push intent signals from your own systems into Unstuck Engine. POST a signal payload with identity + event + source; Unstuck routes it through qualification and into Records.
Using the Enrichment API
A standalone reference for the Enrichment API. POST identity, get back enriched contact channels and firmographic fields. Same engine as the in-app `Enrich` button.
Using the MCP server and CLI
Unstuck Engine ships an MCP server (for AI agents — Claude Desktop, Cursor, custom agents) and a CLI (for shell scripts and quick operations). Same capabilities as the REST API, different ergonomics.
Section 10
Settings & Admin
API keys, sources, billing, team management, and engagement stages.
API keys
Create, rotate, and scope API keys for the Unstuck Engine API, MCP server, and CLI. Each key has a name, scope, and expiration; revoke individually without touching others.
Sources — browser extension and integrations
Sources are the inputs that feed signals — LinkedIn accounts, CRM connections, sequencer integrations, the browser extension. Each signal type relies on one or more sources; manage them here.
Billing and credit usage
Usage-based credit pricing. Every signal capture, enrichment, and qualification costs credits per published rates. Buckets buy you a monthly credit pool with auto-renew; overage tops up automatically or pauses, depending on policy.
Team and engagement stages
Two related-but-separate workspace settings. Team manages who has access. Engagement stages defines the score-band labels (Cool / Warm / In-market / Hot) and their thresholds.