Sync workflows

Compiles the past week of dedupe activity from the Coda queue and both CRMs into a leadership-ready report covering pairs found, merged, rejected, and still pending.

An agent reviews quarantined rows, classifies each failure into fixable versus permanent, repairs the safe ones, re-syncs them, and opens a Linear ticket summarizing the systemic…

Transcribes an inbound demo-request voicemail with ElevenLabs, extracts intent and contact details, creates or updates a HubSpot contact and deal.

Daily matches recently completed Linear and GitHub work against open cross-team dependency tasks in ClickUp, auto-closes the ones that are satisfied.

After access is revoked, this workflow gathers proof from Slack, GitHub, and Google Drive, writes a structured row per system to a Postgres audit ledger.

Sweeps both BigQuery and Snowflake on a schedule, consolidates every newly detected sensitive column into a single Airtable data-governance register.

Watches an Airtable base for new webinar attendee rows, enriches each person and their company, scores them against your ICP.

An agent reads your overnight inbox alongside today's calendar, reconciles meetings against email asks.

On request, an agent reconciles the closing book balance against actual bank and Stripe balances, investigates every unmatched item across BigQuery transaction history.

Two days before a board meeting on the calendar, reconciles the final deck's figures against the Airtable source of record and writes a sign-off page to Notion listing every…

Scans the day's bank and Stripe transactions against a rolling baseline in BigQuery, flags outliers like unusual large debits or duplicate charges, and alerts finance on Slack.

Classifies each webinar attendee's question by buying intent, logs it on their HubSpot contact.

When several people from the same company star your repo, group them as buying-signal evidence, create or update a HubSpot company record.

On a manually triggered backfill, an agent reconciles warehouse rows against Salesforce, investigates why specific rows dropped by reading sync logs.

When a new lead is created in HubSpot, researches the company and contact, builds a sales-ready account brief, and writes it back onto the HubSpot record.

Monthly, an agent pulls Datadog and Honeycomb usage, allocates spend to teams and services by tags, writes the breakdown to Snowflake, and posts a chargeback summary to Slack.

After each reverse-ETL run, reconciles the row count Snowflake intended to sync against the count HubSpot actually accepted and emits a Datadog metric plus a Slack alert…

On a schedule, this scans recent Front and Intercom conversations, fingerprints any errors they mention, matches them to Sentry issues.

Verifies that a single warehouse audience landed completely across multiple destinations (Salesforce, HubSpot.

When a Datadog monitor on Cloudflare cache-hit ratio fires, enriches the alert with the current cache breakdown and the last deploy.

When a HubSpot contact hits a lifecycle stage, generates an ElevenLabs voiceover and Replicate avatar clip in your brand voice, stitches it onto your Loom pitch.

On a webhook from your event tool, scrapes the attendee batch with Apify, enriches into Attio, queues a Loom intro.

Compiles the week's column-classification changes across BigQuery into a written summary and updates a Notion data-governance register so stewards have a single living record…

When an amended RFP version drops in Dropbox, an agent diffs it against the original, identifies added, changed, and removed requirements.

An agent reads each new Dropbox invoice, finds its purchase order in Airtable, and validates line items; when no PO exists, it researches the vendor, drafts a matching PO record.

On a schedule, parses service runbook pages in Confluence and upserts owner, on-call, runbook URL, and SLA into the Postgres service catalog so the ownership bot always answers…

On a schedule, syncs Confluence runbook pages into a Postgres pgvector store and links them to Datadog monitor names so the spike-explainer flows always retrieve fresh.

Fires when a webhook posts a HuggingFace model URL, reads the model card to extract authors and paper citations, resolves each author's affiliation with Brave.

On a schedule, pulls new and updated postmortems from Confluence, chunks and embeds them, and upserts the vectors into Postgres so the retrieval corpus stays current.

On demand, scans a board deck for every figure footnote and verifies each one points to a real, locatable cell in the Airtable source of record.

On a weekly schedule, extracts every cited figure from the latest board deck and checks each one against the Airtable source-of-record table.

On a schedule, scans a Google Drive folder of product screenshots, generates localized alt-text and captions for any image still missing them.

Extracts line items from each scanned PO, checks whether extracted totals and quantities reconcile.

Watches GitLab merges to your API repo, detects changed OpenAPI specs and changelog files, re-chunks and re-embeds only what changed.

Watches your Notion evidence table for rows flagged Low confidence, automatically re-researches each one with fresh sources.

On a nightly schedule, pulls new and updated incident postmortems from Confluence, chunks and embeds them.

When a new deck lands in a watched Google Drive folder, immediately reconciles every cited figure against the Airtable source of record and replies in the uploader's Slack thread…

When the Airtable source-of-record table changes, re-checks previously approved board decks in Drive and alerts the team via Slack to any deck whose cited figures no longer match…

When a ticket is solved without any macro applied, checks whether its answer matches an existing macro and, if not, logs it to Postgres as an uncovered answer so recurring gaps…

When an Outlook email asks for pricing, an AI step summarizes the deal context from the thread, drafts quote inputs into a Postgres pipeline table.

When a receipt image is added to expense storage, this extracts the printed total via OCR, compares it to the claimed line-item amount in Airtable.

Parses a Front vendor invoice and runs a three-way match against the matching purchase order and receipt records in Postgres, posting pass or exception results to Slack.

On a daily schedule, pulls the previous day's Front conversations, detects each one's language, writes the counts to Postgres.

On every closed-lost opportunity in Salesforce, this workflow extracts the loss reason and competitor, summarizes the notes, and writes a clean.

Watches a shared Outlook inbox for vendor invoice attachments, extracts line items, matches them against purchase orders in BigQuery.

For status meetings flagged as low-value, it builds a ready-to-use Notion async update page and posts the organizer a proposal to replace the meeting with that doc going forward.

Before each reverse-ETL sync, verifies the analytics Postgres replica is fresh; if replication lag exceeds the window, it blocks the sync, opens a Linear issue.

Watches a Dropbox folder for new vendor invoice PDFs, extracts each line item, matches it against PO and goods-receipt records in BigQuery.

When a SaaS renewal date approaches via webhook, the CEO agent cross-references SSO usage against contracted seats and drafts a rightsizing brief recommending exactly how many…

When a vendor emails a monthly statement, this matches each line item against your Stripe payment records and posts any missing or amount-mismatched invoices to a Slack channel…

On a fixed schedule, pulls all currently open Sentry issues above a severity floor, writes a single consolidated plain-language status digest.

Polls 1- and 2-star app reviews daily, classifies each by churn driver (bugs, pricing, missing features, performance), and logs every one to an Airtable base with a weekly trend…

On a monthly schedule, this pulls each vendor's latest statement from Google Drive, reconciles every invoice against the AP ledger in Snowflake.

Each week, blends current consolidated cash with historical burn from BigQuery to project runway in months.

When a domain is added to your monitored list, runs a deep Brave Search exposure scan, snapshots every confirmed leak into a Postgres baseline table.

When a vendor invoice PDF lands in the finance inbox, extracts the amount and line items, compares the total to that vendor's historical baseline.

Parses reporting hints and titles from Gmail signatures across an account, reconstructs the implied reporting hierarchy.

When a contract is amended, recomputes the remaining recognition schedule in BigQuery, diffs it against the prior schedule.

Fires when your reverse-ETL tool reports a sync finished, then validates that warehouse and destination row counts match and the data is recent — alerting Slack on any mismatch.

On each completed reconciliation run, ships the full before/after merge audit log from Postgres into Snowflake, generates a reversible change manifest.