Right now, on Slack

A customer just asked your CSM something nobody knows off the top of their head.

It's 2:47pm. Sarah just joined as a CSM last month. This just landed in her DMs.

customer-success-asana

JM

Jordan M. 2:47 PM

Hey Sarah — our CTO is asking about how we'd handle AI workflow failures in Asana when the agent misroutes a task. Is there a runbook for that? We have an exec review in 40 mins and I don't want to wing it 😬

💬 Sarah is typing...

That answer exists somewhere. Probably. The question is — where? And how long will it take Sarah to find it before Jordan's exec review starts?

The next 23 minutes

The answer is in there. Somewhere.

Sarah opens every tool she has access to. Here's what happens.

💬

Slack search: "AI workflow failure runbook"

847 results. Mostly unrelated threads from 2023.

Nothing useful

📄

Confluence: "asana agent misrouting"

3 pages. Last updated 14 months ago. None match.

Outdated

📁

Google Drive: "AI runbook customer escalation"

Found a doc. It's a template with no content filled in.

Empty template

🙋

DMs Alex (senior CSM, 4 years tenure)

Alex is in back-to-backs until 4pm. Away status on.

Blocked

⏱️

Jordan's exec review starts in 15 minutes

Sarah sends "checking on it!" and hopes for the best.

Dropped

Same question. Knowledge base.

Sarah types the question. The answer appears.

Type the question below — exactly how Sarah would ask it. Watch what happens.

AI Workflow Failure — Escalation Runbook Escalation Playbooks

When an AI agent in Asana misroutes a task, the response depends on severity. Here's the decision tree your team uses in customer calls.

Step 1: Confirm whether the misrouting is a rules-engine issue (config) or model inference error (bug).
Step 2: For config issues — walk customer through Custom Rules editor. Resolution time: <10 min.
Step 3: For inference errors — log via internal escalation form, set customer expectation at 2 business days.
Step 4: Send follow-up template #7 ("AI Workflow Clarification") within 1 hour of the call.

Found in 4 seconds

That answer was always there. It just wasn't findable. A knowledge base is the difference between institutional knowledge and tribal knowledge.

60 seconds

Your KB structure. Three categories. Start there.

Don't overthink this. Every CS org needs the same three buckets. Click to add each one.

One toggle

Enable AI. Your KB now understands questions — not just keywords.

Click the toggle. That's the whole setup.

AI-powered search + chatbot

Answers from your KB only — always shows source, never guesses

✓ AI enabled. Searches by intent, not keywords. Chatbot only answers from your content — architecturally cannot hallucinate.

You're done

Invite your team. Every reader is free.

CSMs, support agents, leadership — all free as readers. Only the people writing content pay. Your whole team has access before the end of the day.

⚡

Sarah finds answers in 4 seconds

Not 23 minutes. Not "I'll check and get back to you." The answer is there before the customer finishes typing.

🤖

AI Teammates get KB context via API

Your agents stop flying blind. Clean Markdown via REST. MCP-compatible. Session amnesia solved.

📊

You see which questions have no answer yet

Analytics surface searches that returned nothing. You write the article before the next CSM hits the same gap.

🚚

Already on Confluence or Document360?

Migration is included — we move your content, build your theme, onboard your team. 30 days to live.

Start for free — no CC required Request a demo

This tutorial runs on KnowledgeOwl — $79/month, transparent pricing, readers always free. Awarded for 5 years on G2.