Free preview. This lesson is part of a paid course that earns the ETA Specialist: AI for Customer Success and Support credential. About this course
Where AI helps support and success teams
In this lesson you'll
- Map customer support and success work to good AI uses
- Recognize where AI damages trust and creates liability
- Choose first uses for your team
Welcome
Welcome to AI for Customer Success and Support from the Enterprise Technology Association (ETA). This course is for support agents, support leads, customer success managers, account managers, help center writers, operations and anyone who helps customers after the sale. You'll finish with a capstone: an AI plan for your support or success team, reviewed by ETA.
- Step 1: Assess
Is this a good task for AI, and what's at stake?
- Step 2: Direct
Brief the AI clearly: goal, context, constraints, done criteria.
- Step 3: Assure
Check the output before you rely on it.
- Step 4: Protect
Keep people, data and your organization safe.
- Step 5: Tune
Iterate, then save what works for next time.
Use the steps in order on every task, and loop back to Tune when results fall short.
Two teams, one customer
Support solves problems when customers ask. Customer success helps customers get value so they stay and grow. Both live on knowledge, empathy and follow-through. AI helps both, but customers only see one company.
Where AI helps
| Area | Good uses |
|---|---|
| Replies | Drafting responses from your knowledge base for an agent to check |
| Triage | Tagging, routing and prioritizing tickets |
| Knowledge | Drafting and updating help articles from solved tickets |
| Self-service | Chat assistants that answer from approved content, with a path to a person |
| Success | Account summaries, health signals, renewal prep, business reviews |
| Insight | Voice of the customer themes from tickets, surveys and calls |
Where AI hurts
- Confident wrong answers. In Moffatt v. Air Canada (February 2024), a Canadian tribunal held the airline liable after its website chatbot gave a customer wrong information about bereavement fares. The tribunal rejected the argument that the chatbot was responsible for its own words.
- No way out. Customers stuck in a bot loop with no path to a person.
- Empty empathy. Scripted sympathy that doesn't address the problem.
- Decisions without judgment. Refunds, exceptions and account actions made by AI without a person when the stakes are real.
First uses
Start where agents stay in the loop: AI-drafted replies that agents edit, ticket summaries for handoffs and help article drafts from solved tickets.
Workplace example: a software support team
A 12-person support team handled 1,500 tickets a week. They turned on AI reply drafts in their help desk, grounded in their help center. Agents edited every draft. Using ADAPT, the lead Assessed which ticket types were repetitive, Directed the AI to cite the help article it used, Assured quality with weekly reviews of 50 replies, Protected customer data with approved settings and Tuned the help center when drafts were wrong. Average handle time fell, and the team found 40 outdated articles to fix.
Try it
In your approved AI tool:
Goal: Find the best first uses of AI for my support or success team.
Context: Team: [size, channels]. Volume: [tickets or accounts]. Top ticket types or success tasks: [list with time spent].
Constraints: Sort into AI drafts with a person reviewing, AI with a clear path to a person, people own, and not for AI. Flag anything involving refunds, account changes, personal data or regulated information.
Format: A table, then three uses to start with.
Done criteria: Our team lead would agree.Review it with two agents or CSMs.
Key takeaways
- AI helps with replies, triage, knowledge, self-service, success work and customer insight.
- Companies are responsible for what their AI tells customers; wrong answers and bot loops hurt trust.
- Start with AI drafts that people review.
Sources
