You may already pay for ChatGPT or Copilot, and someone on the team uses it to draft emails. Nothing in the business is wired to it yet, and the question from above is what you're doing with AI.
The first thing I ask is whether someone on your team can explain one process out loud, start to finish, including the exceptions. If they can, that process is a candidate. If the answer is "it depends, ask Fran", it isn't one yet.
Agents follow instructions about as well as the instructions are written
At Uber I designed the knowledge base behind R2D2, a Slackbot that took over the repetitive parts of triage. Most of the effort went into the knowledge base: what each kind of request looked like, what to ask next, and where it went. Language models today can do far more than that Slackbot could, and the same rule applies to them. If the steps live in one person's head, an agent has nothing to follow.
Good first candidates
- Sorting the support inbox into the handful of request types you already get.
- Pulling order details out of emails and into the CRM, with a person approving each one at first.
- Answering the same five questions customers ask at 10pm on a Sunday, from answers you've already written.
- Drafting the weekly report from numbers that already exist somewhere.
Poor first candidates
Work where the rules change every time, or where your team can't agree what a good result looks like. Pricing, credit and refund decisions can wait too. A wrong answer there costs money before anyone sees it, so they come after you've watched an agent on easier work.
Keep a person on it until the numbers say otherwise
For the first weeks, the agent drafts and a person approves. Count how often the person changes the draft. When that number is low and steady, let the agent send on its own for that one kind of request. At iAdvize, for Samsung's Galaxy S20 launch, the sentiment analysis in live chat flagged at-risk customers in real time and routed them to senior agents. The software picked who needed help, and a person gave it.