Summarization workflows

Accepts one uploaded voice memo via webhook, transcribes and structures it into a decision entry, posts a draft to Slack for one-click approval.

Accepts one uploaded voice memo via webhook, transcribes and structures it into a decision entry, posts a draft to Slack for one-click approval.

Scrapes a transcript on a webhook trigger, scores guidance vs actuals, and only when a material miss is detected, escalates a Slack alert plus a detailed Notion brief.

Transcribes a field voice report, classifies severity, and routes routine notes to a Slack channel while escalating anything flagged critical to PagerDuty.

When a deploy webhook arrives, waits for a bake window, compares span latencies before and after the release, and if a regression is detected.

Every morning, gathers all the previous day's production deploys and their Sentry error groups, clusters regressions by root cause across deploys.

Accepts a pasted Loom URL via webhook, transcribes it into a formatted Google Doc with a clickable chapter outline, then posts the doc link to a Slack channel for the team.

On a deploy, checks Sentry for newly introduced high-impact error groups, and if a release crashed real users it clusters the cause, pages on-call via PagerDuty.

On each deploy, summarizes newly introduced Sentry clusters and routes them by severity: high-impact regressions page on-call via PagerDuty, the rest go to a Slack digest.

Identifies error classes that appear for the first time in a release, has an LLM judge their severity from the stack trace and event volume.

On a nightly schedule, gathers the day's field voice notes, transcribes and summarizes each, rolls them into a per-deal digest.

On a Vercel deploy, an agent investigates crash-free trends across the last three Sentry releases, writes a regression narrative, files it as a GitHub release-notes appendix.

An agent rolls up the week's call themes, cross-references them against open support tickets and recent CRM notes.

On a schedule, an agent reads open incidents, alert history, and watch-items across PagerDuty, Datadog, and Sentry.

Chapters a completed Zoom recording, generates a spoken summary clip per chapter.

Accepts an uploaded call transcript via webhook, clusters its objections by theme.

When an incident thread is resolved in Slack, an LLM builds a timeline and root-cause summary, classifies severity, and on high-severity incidents creates Linear follow-up issues…

An agent reads a flagged Slack thread, extracts the decision and action items, then autonomously routes each one — engineering work to GitHub issues.

An operator pastes a Coda doc link into a Slack request and gets back an instant decision-and-open-questions brief for that doc's comment threads.

On a weekly schedule, gathers all CAB Zoom recordings from the past week, merges them into one ranked theme digest showing recurring vs. one-off feedback.

On a weekly schedule, gathers all CAB Zoom recordings from the past week, merges them into one ranked theme digest showing recurring vs. one-off feedback.

Summarizes the week's customer calls into the top five recurring themes with trend deltas versus last week and posts a tight executive brief to a Slack channel.

Scans recorded Zoom calls tied to deals that recently went closed-lost, extracts the unresolved objection that likely killed the deal.

Scans recorded Zoom calls tied to deals that recently went closed-lost, extracts the unresolved objection that likely killed the deal.

Runs daily, clusters new Intercom conversations into themes, and posts a Slack alert only when a theme's volume spikes sharply above its trailing baseline.

Logs every CAB Zoom recording's product themes as rows in an Airtable tracker with severity and segment fields.

After each Zoom discovery call, isolates the objections and lost-deal signals, classifies them by type, and logs them to an Airtable objection tracker.

When a Front escalation closes, drafts a structured handoff document in Notion capturing the resolution, customer commitments, and open follow-ups, then alerts the AE in Slack.

Every morning, compares yesterday's span latencies to the trailing week, narrates the services that drifted slower, and appends a dated digest entry to a Notion performance log.

Transcribes each new Loom standup video, extracts concrete action items and their owners with AI, and creates matching Linear issues while posting the brief summary to the team.

Weekly, ranks the top recurring slow-span clusters across services, writes an executive perf brief to Slack, and opens a Linear issue for each persistent regression that lacks one.

An agent transcribes a crew safety voice report, condenses it into a structured incident record, logs it to a Postgres trend table.

On a daily schedule, gathers the day's field voice memos from Drive, transcribes and summarizes each into a structured report, archives the reports back to Drive.

Right after each Zoom customer call ends, transcribes it, scores sentiment, and only when the call reads as negative or at-risk posts an escalation alert with summary…

After a Zoom sales call, summarizes it into a Salesforce opportunity note and, if the AI detects deal risk or a competitor mention, alerts the account owner in Slack.

When a Zoom recording finishes, it transcribes the call, extracts only the decisions and action items.

After a recorded Zoom call, an agent identifies the unresolved objections, drafts a tailored follow-up email addressing each one, logs the objections to the HubSpot deal.

After a recorded Zoom call, an agent identifies the unresolved objections, drafts a tailored follow-up email addressing each one, logs the objections to the HubSpot deal.

An agent processes a CAB Zoom recording end to end — reasoning about themes, cross-referencing the company's roadmap, drafting a brief.

Each Monday, summarizes the prior week's recorded customer calls into an executive briefing, narrates it as an audio file with ElevenLabs.

When a Honeycomb anomaly trigger fires during an active page, summarizes the slow-span cluster into a triage note and attaches it to the matching PagerDuty incident…

Detects escalations from VIP accounts in Front, generates a one-paragraph executive recap, and pages the named account exec via Slack and Salesforce task within minutes.

Drains a backlog of voice memos by transcribing each, classifying it as decision vs. action.

Drains a backlog of voice memos by transcribing each, classifying it as decision vs. action.

Turns a finished Zoom customer-advisory-board recording into a structured product-themes brief in Notion, then opens one Linear issue per theme assigned to the owning PM.

Turns a finished Zoom customer-advisory-board recording into a structured product-themes brief in Notion, then opens one Linear issue per theme assigned to the owning PM.

Every Friday, transcribes the week's recorded Zoom customer calls, scores each for sentiment.

An agent extracts decisions from resolved Front threads, searches the existing Confluence decision registry for related prior decisions.

On a weekly schedule, gathers the week's Loom recordings, transcribes and summarizes each, compiles a single themed digest in Coda, and posts the digest link to a Slack channel.

When Axiom detects a log signature it has not seen before, this workflow fetches sample lines, has AI explain the likely cause and blast radius in plain English.

Trigger this with two tags and it pulls every PR merged between them, drafts grouped release notes with breaking-change callouts, and publishes a versioned page to Confluence.

After a Zoom all-hands, extract committed action items with owners and due dates, create assigned Asana tasks, and notify each owner in Slack with their items.

Polls Honeycomb on a short interval for trace error-rate spikes, summarizes the affected traces with OpenAI, posts a digest to Slack.

Once a day, gathers all of yesterday's Zoom recordings, summarizes each, posts a combined digest to Slack, and files every follow-up as an Asana task.

On a schedule during a canary rollout, compares the canary release's error-budget burn against the stable baseline and, if it regresses, posts a rollback recommendation…

When someone posts a Loom link in Slack with a command, it transcribes that specific video, summarizes it into a spec entry, adds it to the Coda rollup.

When a Sentry alert fires for an error spike, attributes it to the most recent deploy, summarizes whether the spike is dominated by new error classes introduced by that release.

On every Vercel production deploy, pulls the new and spiking Sentry error groups, clusters them by root-cause signature.

After each recorded Zoom call, extracts the top timestamped highlights and posts a concise jump-link digest to a Slack channel so the team gets the gist without rewatching.

On each new Vercel production deploy, compares Sentry error rates against the previous release and posts a plain-English triage brief to Slack only when a regression crosses your…