# CRM and data capture

Accounts, contacts, deals, leads, tasks and notes, with custom fields and objects. Records fill in from connected email, calendars, meetings, Slack and other tools, and filled-in values link back to where they came from. Keep your current CRM and sync with it, or move over.

A note in italics after an item says when it's off by default, or needs your own account with another service. Every area: the [product overview](https://www.closefinch.com/product).

## Records, fields and CRM sync

- Native CRM objects (Accounts, Contacts, Deals/Opportunities, Leads, Tasks, Notes, Activities, Meetings, Lists) with a REST schema endpoint per resource.
- Typed custom fields (short/long text, number, boolean, date, URL, single/multi-select, single/multi-user, single/multi-record, currency, percentage), archivable, org-level creation toggle.
- Custom objects and arbitrary relationships between objects.
- Multiple pipelines (sales, partnerships, renewals); colored stages with an immutable stage type (start, in-progress, won, lost) that drives reporting and signals; Kanban board.
- Lead object with lifecycle (new → contacted → qualified/unqualified → converted), source metadata and estimated value; conversion creates Contact, Account and optional Deal.
- Saved and dynamic lists with multi-membership, archive and restore, and auto-refresh as ICP data changes.
- Custom task statuses with category (To Do / In Progress / Done) and color.
- Append-only, source-backed activity timeline: synced items update or disappear when the source changes; custom events posted with `external_id` for idempotency.
- Business or knowledge graph linking accounts, people, conversations, signals and their relationships, used as shared agent memory.
- Object history showing how each value changed over time, with point-in-time reads through the API.
- Account memory: facts that people and agents save about an account, each with its evidence and revision, readable through the API and MCP.
- Auto-create Accounts by email-domain matching and Contacts from email, calendar attendees, Slack Connect members and notetaker attendees.
- Duplicate detection and merge into a survivor record, with redirect semantics for API clients (301 on read, 410 on write).
- A dedup check before an agent creates or updates a record, as an agent tool.
- Dismiss or ignore noise objects so they stop resurfacing.
- AI-defined field: user names the field, the LLM picks the type, writes the extraction instructions and autofills it.
- Backfill a new or changed field across all historical records.
- Computed, calculated and conditional fields.
- Schema-less, conversation-first capture where structured fields are derived from unstructured conversation data.
- Signals as account fields: score, status, signals, and the latest signal with its date and source.
- AI-proposed account matching with source and confidence: a person approves a match once and later emails and imports follow it.
- Canonical master data plus a business glossary/ontology (terms, metric definitions, calendar rules, lineage) read by every connected AI tool.
- Email aliases: users verify alternate addresses (up to 10) so activity attributes correctly.
- Ownership model: deals inherit the account owner, `owner:none` filtering, plain-English bulk reassignment.
- Agent-driven migration from Salesforce, HubSpot, Attio, Pipedrive, Close, Zoho, Affinity, DealCloud or spreadsheets: CSV upload, field-map confirmation, dedup and pipeline import.
- Bulk-create API/CLI with idempotent row keys and bulk operations.
- Run beside your current CRM with two-way Salesforce, HubSpot, Attio or Pipedrive sync, and migrate in phases by object, pipeline and team.
- Overlay mode: agents write through to your existing CRM instead of replacing it, and agent changes wait for a person's approval before they reach it.
- Write-through to Close, Zoho CRM, Affinity and DealCloud (Pipedrive also syncs both ways).
- CRM hygiene agent: checks custom fields, finds duplicates, missing fields, broken associations, stale owners and zombie deals, ranks them by impact, and applies only the fixes a person approves, with a recovery trail.
- Sandbox environments for testing schema and agent changes.
- Two-way sync conflict resolution: field-level precedence rules, a conflict review UI and published sync latency.
- Per-field value provenance: every auto-filled CRM value links to the exact email, transcript span or event that produced it.
- Relationship-strength field per contact (who on the team knows them, how, last two-way touch) usable by agents and gates.

## Capture from email, calendar and other tools

- Gmail sync every minute with a qualification filter: business mail is kept, and personal, marketing and automated mail is left out unless it matches the CRM.
- Outlook / Microsoft 365 mail sync.
- Google Calendar sync from the calendars you pick, by live event-watch and a one-minute sweep; private and declined events are left out.
- Outlook Calendar sync.
- Slack channel and thread capture linked to the right account or deal.
- Slack Connect: each shared channel maps to one account; create channels with a prefix or link existing ones with history backfill; external members become contacts; channel health shown.
- WhatsApp, Telegram and Instagram message capture into the graph. *(Needs your own WhatsApp Business, Telegram or Instagram account.)*
- Notetaker imports from Granola, Notion meeting notes, Circleback, Fireflies, Fathom, Grain, Gong and Chorus, and a signed webhook for any other notetaker.
- Dialer and phone-call capture (Aircall, Dialpad, RingCentral, Zoom Phone, Orum, Outreach).
- In-person capture via a mobile recorder or hardware (Plaud), with a first-use consent screen.
- Billing events as account activity (Stripe webhook matched by customer email; read-only Stripe key for former-customer detection).
- Product-analytics event ingestion (PostHog, Segment, Amplitude, Mixpanel, or a first-party JS/React intent client).
- Support-ticket ingestion (Zendesk, Intercom, Plain) linked to accounts.
- Data-warehouse connections: scheduled read-only queries from Snowflake, BigQuery, Databricks, Redshift and ClickHouse.
- Email-forwarded report ingestion into a dedicated agent inbox (forwarding rule or add as recipient).
- CSV, spreadsheet and file import, including attaching a file in agent chat.
- First-party website tracking script and embeddable forms capturing visitor identifiers and page/form interactions.
- Document-open tracking (proposal opened N times, which sections).
- Automatically permissioned multiplayer ingestion so each teammate's data is shared per policy.
- Read-time freshness: records older than a threshold are re-checked when read; email re-verified before send.
- Knowledge connectors (Drive, SharePoint, Confluence, Notion, Box, Dropbox, Egnyte, Highspot, Seismic, sitemap crawl).
- Permission-aware indexing that respects source ACLs; Slack read in real time and not stored.
- RFQ and order-document parsing from email bodies, PDF attachments and call transcripts into structured line items.
- Bot-free transcript import from Zoom, Google Meet and Microsoft Teams.