Convert a Slack thread into an evidence item

The best customer signal in most companies is in #cs-feedback or a CEO DM thread. It's usually one screenshot, two reactions, and zero follow-through. This use case is about catching it before it scrolls off — turning a 30-second exchange into a permanent evidence item.

The input

Any Slack thread. Could be a customer reply forwarded by the CS team, a screenshot of a support conversation, a Loom transcript pasted by an account manager, or your own #founder-thoughts musing.

[10:42] Sarah (Acme Corp): "Hey, just wanted to flag — we tried to invite our CS team this morning and the invite emails never arrived. Spent 20 minutes troubleshooting before giving up." [10:43] Mike (us): "Sorry about that — checking now." [10:51] Sarah: "It's not blocking but we'll just add them tomorrow."

The flow

1. Open the composer. From any product, switch to the Evidence tab. The composer sits at the top of the surface.

2. Paste the raw thread. The composer handles markdown-formatted Slack exports and plain copy and paste. No formatting required. A short paste becomes one signal; a longer thread is extracted into rows you keep or drop.

3. Watch it get classified. The AI extracts the actual customer voice (Sarah's two messages) and tags the type (friction + request) with a severity suggestion. Classification is not judgment: it sorts the signal, it does not decide whether the signal is true.

4. Add the source and, optionally, the stage. Where did this come from? "Slack #cs-feedback, 2026-05-09." That metadata is what lets future you re-find the original. You can also tag the user-journey stage (for example team setup) if you have a stage taxonomy in this product.

5. Save. It lands in the Inbox marked Unreviewed. It is captured, not endorsed. When you are ready, Use as backing reviews it and anchors it to one exact claim in a spec, stamped with your name and the time.

The output

A structured evidence item with:

• The quote — exactly what the customer said, not a paraphrase
• Type — friction, quote, observation, metric, or request
• Severity — auto-suggested based on the language
• Source — where it came from, so you can verify or follow up
• Timestamp — when it landed

That's a permanent artifact. It survives team changes, tool migrations, and the inevitable decay of Slack search.

Why this beats a #feedback channel

Slack channels are append-only firehoses. They optimize for "did anyone see this?" not "which of these did anyone actually judge?" Pathmode is the second half. Every signal keeps its source, carries a trust state, and stays searchable in Browse long after the thread has scrolled away.

Try it yourself

1. Find any thread in your CS or feedback Slack
2. Copy the messages (or screenshot if it's image-only)
3. Open Pathmode, choose a product, go to the Evidence tab and paste it into the composer
4. Add a source and save; the signal waits Unreviewed in the Inbox

Related

• Use case: Turn 30 support tickets into a prioritized spec
• Use case: Find the friction pattern across 5 user interviews
• Playbook: From Support Ticket to Shipped Feature