The CRM: a contact record that thinks
Mass's CRM isn't a static address book — it's a living record of every interaction a person has with your business, with AI reading that record to tell you what to do next, write the outreach for you, and turn one contact into a personalized campaign. It's wired into the same tracking pixel and self-healing engine that power the rest of the platform, so behaviour on your funnels shows up here in real time.
22 min read · The complete CRM guide
What the CRM is
One contact record that unifies identity, behaviour, deals, and conversations.
Every lead, customer, and prospect lives as a single contact in the CRM. The record holds the obvious things — name, email, phone, company, job title, lifecycle stage, tags, and a lead score — but its real value is what gathers around it: every page they viewed, every email and message exchanged, every deal, order, and course enrollment, all stitched to one identity.
Because the contact record is the meeting point for all of that signal, it's also where Mass's AI does its best work. The same facts that populate the timeline feed the AI features — Next Best Action, summaries, research, and personalized campaigns — so the assistant always reasons from what actually happened, not a generic prompt.
- Unified identity — anonymous pixel visitors are stitched to a contact the moment they submit a form, so pre-conversion behaviour isn't lost.
- Rich profile — identity, engagement, attribution, enrichment, and any custom fields you define, all on one record.
- Lifecycle & scoring — a lifecycle stage and a lead score keep the list sortable by who's hottest right now.
- AI-ready — every interaction is structured so the AI can read it — the record is the context for everything the assistant does.
The unified journey timeline
Every interaction, from every source, on one chronological timeline.
Open a contact and you see their journey — a single, chronological merge of everything that's happened across the platform. Pixel events (page views, clicks, form submits, purchases, and video plays), email sends and opens, two-way messages, deal stage changes, orders, and course enrollments are all loaded in parallel and woven into one timeline.
Video plays get special treatment: a pixel `content_view` event carrying a video id is upgraded to a "Watched: <title>" entry, so you can see exactly which VSL or lesson a contact watched and when. The timeline is the raw material every AI feature reads from.
- Every source — pixel, email, messages, deals, orders, and enrollments merged into one chronological view.
- Behavioural detail — page paths, UTM attribution, scroll depth, and time-on-page come straight from the pixel.
- Video-aware — tracked video plays surface as "Watched: <title>" with the surface and hub they happened on.
- The AI's source of truth — summaries, Next Best Action, and campaigns all reason from this timeline.
Pipelines, deals & activities
Move opportunities through stages and keep the work attached to the contact.
Deals move through a visual pipeline — a kanban of stages you drag cards across — with each deal carrying its value, currency, probability, stage, and expected close date. Those same deal facts are available as merge fields and feed the AI when it decides what to do next or writes a campaign.
Tasks, notes, and activities attach to the contact and the deal, so the full working history — not just the marketing touchpoints — lives in one place. The pipeline is also a signal source: deal momentum is one of the inputs the Next Best Action engine weighs.
- Visual pipeline — a drag-and-drop kanban of stages with deal cards you move as opportunities progress.
- Deal facts — value, currency, probability, stage, and expected close date — usable in merge fields and AI prompts.
- Activities — tasks, notes, and logged activities keep the working history on the record.
- A signal, not just a board — deal momentum feeds Next Best Action so the suggestion reflects pipeline reality.
AI Next Best Action
The CRM reads the journey and tells you the single highest-leverage thing to do next.
Next Best Action (NBA) is the CRM's recommendation engine. It loads a contact's journey summary, deal signals, and course signals, then asks an AI model for one concrete next step — a typed action with a label, a reason, a priority, and an optional due date and template hint. Instead of staring at a timeline wondering what to do, you get "Send a follow-up referencing the pricing page they viewed twice" with the reasoning attached.
NBA is validated against a strict schema so the output is always a usable action, and the model is picked per request by a model-picker tuned for the job. Recommendations are cached and marked stale when something new happens — sending a message, for instance, flags the contact's NBA for refresh — so the suggestion never goes out of date silently.
- One concrete step — a typed action with label, reason, priority, and an optional due date and template hint.
- Grounded in signal — built from the journey summary plus deal and course signals — not a generic prompt.
- Schema-validated — every recommendation is validated so it's always a real, actionable next step.
- Freshness-aware — NBA is marked stale when new activity lands (e.g. a message is sent) so it refreshes.
Journey & thread summaries
Catch up on any contact or conversation in a sentence instead of scrolling.
Two summary passes turn raw history into something you can read in seconds. A journey summary distills the contact's whole timeline into a short narrative the rest of the AI features reuse as their brief. A thread summary condenses a specific inbox conversation — the back-and-forth of an email or message thread — so you can pick up where you left off without re-reading every line.
Both summaries are priced in credits and use the same model-picker as the rest of the CRM AI, so quality and cost stay consistent across features.
- Journey summary — the contact's whole timeline distilled into a brief that other AI features reuse.
- Thread summary — a single inbox conversation condensed so you can resume it instantly.
- Reused downstream — the journey summary is the shared context for Next Best Action and campaigns.
Deep research & outbound briefs
A one-click deep dive that produces a structured outbound brief and a ready-to-send sequence.
Research runs a deep-dive pass on a contact and produces a structured outbound brief: a company summary, ICP-fit read, the buying signal it found, pain points, a custom angle, a recommended offer, an opener, and a three-email sequence. The brief is persisted on the contact so it's there when you come back, and you can add an optional critic pass that reviews the sequence and reports how to tighten it.
Outbound angles are a first-class concept — you pick the angle the research should lean into, and the brief and emails are written to it. It's the fastest way to go from "who is this?" to a personalized, on-strategy first touch.
- Structured brief — company summary, ICP fit, buying signal, pain points, custom angle, recommended offer, and opener.
- Ready-to-send sequence — a three-email outbound sequence written to the angle you chose.
- Optional critic pass — an extra review that scores the sequence and suggests improvements.
- Persisted — the brief is saved on the contact so the work is there next time you open it.
Personalized campaigns: Strategist → Copywriter
Turn one contact into a tailored multi-step campaign with a two-pass AI pipeline.
The custom-campaign builder splits "read the contact" from "write the emails" into two passes. First the Strategist analyzes the contact's journey, deals, courses, and any past messages and produces a StrategyCard — the contact's archetype, the right tone, their pain points, likely objections, proof points, conversion levers, suggested copy frameworks, and CTA candidates. You review that strategy, edit it if you want, and only then run the Copywriter.
The Copywriter (the blueprint generator) takes that tight brief and writes the campaign — a sequence of steps with a goal, a summary, an estimated duration, and exit conditions. Any single step can be regenerated on its own, and the copy is compiled through the merge-field resolver so every {{first_name}}, {{company}}, or deal token resolves against the real contact. Separating strategy from copy means tighter, more personalized output for fewer tokens.
- Strategist pass — produces a StrategyCard — archetype, tone, pain points, objections, proof points, levers, frameworks, and CTAs.
- Review then write — you approve or edit the strategy before the Copywriter spends tokens on the emails.
- Blueprint generator — writes a multi-step campaign with a goal, summary, duration, and exit conditions.
- Per-step regeneration — regenerate any single step without rewriting the whole sequence.
- Merge-field compiled — every token resolves against the real contact and deal at send time.
Why two passes
Handing the Copywriter a finished StrategyCard instead of a blank page produces tighter, more personalised copy for fewer tokens — the strategy call is small and the writing call is focused.
Lead discovery & enrichment
Find net-new leads and fill in the gaps on the ones you have — from one command center.
The Command Center is the prospecting surface. Lead Discovery imports person-level leads (including from a Sales Navigator URL search), the Email Finder fills in missing addresses, the Enrichment Hub completes records, Prospecting Lists organize them, Intent Signals surface who's showing buying intent, and Fathom Calls brings call data in. Imported leads land directly in the CRM as contacts, even when they start as just a name and a company, so later enrichment passes can fill the gaps in place.
Discovery can optionally enqueue a bulk research job for newly imported contacts, so a fresh import can come in already briefed — discovery and the research pass chained into one motion.
- Lead Discovery — import person-level leads, including from a Sales Navigator URL search.
- Email Finder & Enrichment — fill in missing emails and complete partial records in place.
- Intent & calls — surface buying intent and bring Fathom call data onto the record.
- Auto-research — optionally enqueue a bulk research job so imports arrive already briefed.
Outbound sequences & deliverability
Send at scale across warmed inboxes, with deliverability watched per inbox.
Outbound runs multi-step sequences across your sending inboxes. Each inbox has a warmup phase and a daily cap, and messages dispatch through a channel-aware sender. A deliverability dashboard reports per-inbox health — 24-hour and 7-day send volume, bounce and complaint rates, warmup phase, daily-cap utilisation, and the inbox domain's latest DMARC pass rate — so a deteriorating inbox is caught before it drags the rest down.
A per-step campaign dashboard shows the funnel for a sequence: enrollments at each step, plus sent, bounced, complaint, and replied counts, with replies attributed through message classification. Sending a message also marks the contact's Next Best Action stale, keeping the recommendation honest as the conversation moves.
- Warmed inboxes — each sending inbox has a warmup phase and a daily cap so volume ramps safely.
- Deliverability dashboard — per-inbox send volume, bounce/complaint rates, warmup phase, and DMARC pass rate.
- Campaign funnel — per-step enrollments with sent, bounced, complaint, and replied counts.
- Reply classification — replies are classified and attributed back to the sequence step.
The loop with the pixel & self-healing
Behaviour on your funnels flows into the CRM, and the CRM acts on it automatically.
The CRM doesn't sit apart from your funnels — it's downstream of the same first-party tracking pixel. Every pixel event lands on the contact's journey, identity stitching connects an anonymous visitor to a contact the moment they submit a form, and that behaviour immediately enriches the AI's context: a contact who just watched your VSL twice and hit the pricing page is a different Next Best Action than one who went quiet.
It's also a closed loop. Pixel events fan out to automations — a tracked video play can fire a `video_watched` automation, and crossing an engagement threshold can trigger a contact-scoped action — so the CRM reacts to behaviour without you watching the dashboard. Combined with the self-healing engine that optimizes the funnels feeding the CRM, the whole system reads behaviour and responds on its own.
- Pixel → journey — every tracked event lands on the contact timeline as it happens.
- Identity stitching — an anonymous visitor becomes an identified contact the moment a form is submitted.
- Behaviour → AI — fresh behaviour sharpens Next Best Action, summaries, and campaign strategy.
- Behaviour → automation — video-watched and engagement-threshold events trigger CRM actions automatically.
Custom fields, filters & smart views
Shape the record to your business and slice the list by anything.
A field registry is the single source of truth for what's filterable, sortable, and selectable on the contacts list. Static fields map to columns; custom fields you define are addressed as `custom_fields.<key>` and resolved at query time, so the list adapts to your data model without schema changes. The same registry is injected into the natural-language-to-filter prompt, so when you describe a segment in plain language the AI can only build filters from real fields — it can't invent a column.
That makes smart views fast to build: filter by lifecycle stage, lead score, attribution, enrichment data, or any custom field, save the view, and work the list. The Command Center AI chat sits alongside, handling commands like find-leads, hot-leads, analyze-pipeline, and draft-email against the same data.
- Field registry — one source of truth for filterable, sortable, and selectable fields across the list.
- Custom fields — define your own fields, addressed as custom_fields.<key> and resolved at query time.
- NL-to-filter — describe a segment in plain language; the AI builds filters from real fields only.
- Command Center chat — find leads, surface hot leads, analyze the pipeline, and draft emails by chat.