The Customer Success
Field Report

Most organizations are pushing AI adoption, so CS leaders and managers are trying it. Almost no one can say what it’s returning. They’ll be asked to justify these tools the same way they’re asked to justify everything else, and right now they don’t have an answer, not on ROI, not on time saved.

Leadership wants them on it, so they’re trying. But they’re worried about leaning on it too much. Most land in the same place: it’s the only realistic way out of the workload. Yet with customers wanting more one-on-one conversation, not less, they’re cautious about what they adopt, and want to keep it as far from the customer relationship as possible.

The AI being sold to them makes big promises and mostly doesn’t deliver. It either doesn’t understand the job, or comes with a learning curve that takes more time than it saves. Every purchase becomes another dashboard, another login that never gets used.

Past about 30 accounts, the limit isn’t care, it’s recall. What each account needs lives in the CSM’s head and a trail of notes, so every meeting starts with rebuilding what happened last time. To answer one question, they open five tools, and still can’t get a straight answer on the health of a customer.

First, the data lives in scattered tools, Salesforce, Mixpanel, that don’t talk to each other, so even with AI, there’s no single source of truth. Second, the signals exist but nobody can make sense of them. The warning signs are there, in tickets, call notes, Slack, just never in one place. And with CSMs managing more accounts than ever, catching those signs as they come is nearly impossible.

Most QBRs just replay usage, tickets, and adoption, numbers the customer already has. The test one CSM uses: if they’d learn the same thing from the deck, the meeting shouldn’t exist.