SUMMARIZATION
Intercom Theme Clustering with HuggingFace to Confluence
Uses a HuggingFace embedding model to cluster the month's Intercom conversations into themes and publishes a labeled voice-of-customer report to Confluence.
How it runs
The automated pipeline, trigger to output.
- TriggerMonthly schedule
- ActionFetch prior month's Intercom conversationsIntercom
- ActionEmbed transcripts with HuggingFace modelHugging Face
- LogicCluster embeddings into theme groups
- ActionLabel and summarize each clusterOpenAI
- OutputPublish report to Confluence pageConfluence
What it does
It produces a monthly voice-of-customer report using open embedding models. The flow embeds the month's Intercom conversations with a HuggingFace sentence model, clusters them, labels each cluster with a generated theme name and summary, and publishes the full report to a Confluence page for the product and CX orgs to review.
When to use it
Use it for a recurring deep monthly read of customer sentiment where you prefer self-hosted or open embedding models over a proprietary API, and where the output needs to live in your team wiki for broad visibility.
How it works
- 1A monthly schedule starts the run.
- 2Fetch all Intercom conversations from the prior calendar month.
- 3A HuggingFace inference step embeds each conversation transcript.
- 4Cluster the embeddings and assign each conversation to a theme group.
- 5An OpenAI step labels and summarizes each cluster with size and sentiment.
- 6Publish the assembled report as a new Confluence page under the CX space.
Set it up
What you configure once, before turning it on.
- 1Connect IntercomConversations, contacts, articles.
- 2Connect Hugging FaceModels, datasets, spaces — the open-source hub.
- 3Connect OpenAIModels, embeddings, files.
- 4Connect ConfluenceSpaces, pages, blueprints.
- 5Set each agent's modelWe leave models unset so you pick the tier — fast + cheap, or top-quality.
- 6Tune it to your dataEdit the prompts, filters, and field mappings so it matches how your team works.
- 7Test, then turn it onRun once against a sample, confirm the output, then enable the trigger.
More Summarization workflows
On-submit Loom standup roll-up archived to Confluence
When a standup video is submitted via webhook, transcribes it, generates a per-person written summary, and appends it to a running team standup page in Confluence.
Front Escalation Handoff Doc in Notion
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.
Front Escalation War-Room Brief to Slack
On a Front escalation, posts a concise threat-assessment brief to a Slack channel only when the AI judges the situation high-severity.
VIP Front Escalation Instant Exec Page-Out
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.
Release health note per Vercel deploy
When a Vercel deploy goes live, summarizes the Sentry errors observed in the release window into a plain-English health note that separates brand-new error classes…
Rollback recommendation when a deploy spikes errors
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.
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.