CRM
Postgres Account Dedupe Agent with AI Conflict Adjudication
An agent inspects a Postgres customer table for fuzzy-duplicate accounts, reasons over each conflicting field to recommend the correct value with a rationale.
How it runs
The automated pipeline, trigger to output.
- TriggerWebhook signals candidate duplicate account IDsHTTP webhook
- ActionPull both records and related rows from PostgresPostgres
- LogicAgent adjudicates each conflicting field with rationale
- LogicEscalate low-confidence fields for mandatory human input
- ActionPost adjudicated merge brief to Microsoft TeamsMicrosoft Teams
- OutputOn sign-off, run merge transaction in Postgres and log outcomePostgres
What it does
This agent-driven workflow queries a Postgres customer table for fuzzy duplicates, then for every field where the duplicates disagree it reasons about which value is most likely correct (freshest, most complete, internally consistent) and writes a short rationale. It assembles an adjudicated merge brief and sends it to a Microsoft Teams channel where an owner approves or overrides before the merge is applied.
When to use it
Use it when duplicate detection is easy but choosing the right surviving value is judgment-heavy, and you want explainable per-field recommendations rather than rigid rules.
How it works
- 1A webhook from your data pipeline signals candidate duplicate IDs.
- 2The agent pulls both records and their related rows from Postgres.
- 3It adjudicates each conflicting field, attaching a confidence level and rationale.
- 4A branch escalates low-confidence fields for mandatory human input.
- 5The merge brief is posted to Microsoft Teams for approval.
- 6On sign-off, the agent runs the merge transaction in Postgres and logs the outcome.
Set it up
What you configure once, before turning it on.
- 1Connect PostgresAny Postgres URL — query, write, migrate.
- 2Connect Microsoft TeamsChannels, chats, files.
- 3Connect HTTP webhookTrigger any URL on agent actions.
- 4Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 5Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 6Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More CRM workflows
Backfill Intercom conversation topics onto stale HubSpot contacts
Finds HubSpot contacts with no recent Intercom topic, pulls their conversation history from Intercom, classifies the dominant topic with OpenAI.
Daily orphaned-opportunity sweep with manager Slack digest
Each morning this scans Salesforce for open opportunities owned by inactive users and posts a per-manager digest in Slack listing the orphaned deals that need a new owner.
AI-Personalized Dormant Account Win-Back
On a dormancy anniversary, an agent researches the account's history and recent public signals, drafts a genuinely personalized win-back email.
Dormant Account Anniversary Re-engagement Email
Each day, finds Attio accounts whose last activity was exactly N months ago and sends each owner's contact a personalized re-engagement email from Gmail.
Dormant Account Check-In Call Scheduler
On the anniversary of an account's last touch, books a tentative check-in call on the owner's Google Calendar and emails the contact a scheduling invite.
Weekly Dormant-Anniversary Cohort Digest
Every week, groups dormant Attio accounts by which last-touch anniversary they hit that week, builds a prioritized digest.
Run it inside a business
This workflow drops into a full company template. Import the org, and this is one of the playbooks its agents run.
Run this workflow in your colony.
14-day trial. No DevOps. No Sales call. Provisioned in under a minute.