How CROs should use AI in their GTM motion: a playbook (so far)

Picture a typical day for a rep. They research an account before a call, write up notes afterward, update the CRM, draft a follow-up, and sit in a few internal meetings. Somewhere in there, they talk to a customer. By the end of the day, they’ve spent more time around customer conversations than in them.

When we started investing heavily with AI in our go-to-market team at Starbridge, that was the first thing we wanted to change.

I shared where we’ve landed at State of GovTech last month. Afterward, the question I heard most from other revenue leaders was some version of “where do I even start?”

This post is my answer so far. It isn’t theory, and it isn’t finished. It’s what’s working for us right now, what hasn’t, and how I’d think about it if I were starting over tomorrow.

Matt

How should a CRO use AI in go-to-market?

A CRO should use AI to take over the work that surrounds a customer conversation, so the team can spend more time in the conversation itself. That means research, account prioritization, call prep, note-taking, CRM updates, and first drafts of follow-up.

AI shouldn’t replace relationships. Buyers still buy from people they trust, and that trust is built live, one conversation at a time.

Everything below comes back to that one idea.

Pick one metric that keeps you honest

Our north star is the percentage of time our team spends in front of customers.

It isn’t the number of agents deployed or hours automated. It’s time with customers.

AI is great at a lot of things. It can research an account in minutes, summarize an hour-long call, and clean up a CRM record. What it can’t do is sit across from a buyer and earn their trust. Talking live with customers is still a human job.

So every automation we build has to answer one question: does this buy back time for more customer conversations, or better ones?

If the answer is no, we don’t build it. If it’s yes, we check whether the time actually got reinvested. Saving a rep hours every week means nothing if those hours go to more internal meetings.

A metric like this cuts through the useless stuff and zeroes in on what makes an impact. It’s easy to get excited about a clever workflow. It’s much harder to argue with a number that tells you whether your team is spending more time where deals get won.

Which GTM work to automate first

Start with the work that fills a rep’s calendar without ever touching a customer. For us, that’s been:

• Pre-call research. Account background, recent news, open opportunities, and past conversations, pulled into one brief before the call.
• Account prioritization. A ranked list of the accounts most worth a rep’s time this week, with the reason each one moved.
• Call recaps and notes. Summaries of every conversation, so no one has to rewatch a recording to remember what was said.
• CRM updates. Records that stay current without a rep typing them in at 6 p.m.
• First drafts of follow-up. A starting point that reflects the call, which the rep then makes their own.

None of this is glamorous. All of it adds up. And every hour it gives back is another hour a rep can spend with a buyer.

Why the CRO job has changed

The job is still to find revenue and keep revenue. That part isn’t going anywhere.

Everything underneath it has shifted: how you find accounts, how you prep, how you forecast, how you coach, and how you hand an account to customer success all look different now. And in 5 months they’re going to look different again.

Here’s the test I use on myself. If what I’m doing today looks like what I did two years ago, I’m probably doing it wrong.

That doesn’t mean throwing out what works. Discipline still matters. Pipeline hygiene still matters. A great discovery call still matters. It means questioning the why and how.

• Why does a rep spend an hour prepping for a first call when an agent can build the brief in minutes?
• Why is forecasting still a weekly spreadsheet exercise?
• Why does a new account owner inherit a customer through a pile of notes instead of a clean summary of every conversation?

For CROs, the shift is from managing activity to designing systems. It’s less “did you make your calls?” and more “where is my team still stuck doing busywork?”

Start with the experience, then build backwards

The barrier to innovation has never been lower. Things that used to need an engineering sprint can now be built by someone on the sales team in an afternoon.

Most teams approach AI by taking their existing process and asking where to bolt it on. That gets you a slightly faster version of what you already do.

Try a different exercise instead:

1. Forget how you’ve always run the process.
2. Ask what the most magical experience would be for your users, whether that’s a customer, a rep, or a CSM.
3. Build backwards from there.

For a customer, magical might mean every conversation starts with full context, and nobody asks them the same question twice. For a rep, it might mean starting Monday with a short list of accounts worth calling, with the reason to reach out now and the right contact already attached.

Once you can picture the end state, the path gets clearer. It’s usually shorter than you’d expect.

Block time for this. Imagination doesn’t show up in a packed calendar. It shows up when your team has room to ask “what if?” and permission to build something that might not work.

Keep learning, and don’t do it alone

AI moves fast. Things I thought models couldn’t do six months ago now work in one shot. Workflows we built carefully a quarter ago have already been replaced by something simpler.

That means two things.

Revisit your assumptions often. If you tried something six months ago and it didn’t work, try it again.

Find your people. Build a small group you trust, inside or outside your company, and share openly. Talk about what’s working, what broke, and what you’re trying next. The fastest way to get better is to borrow good ideas and give yours away.

Nobody has this figured out yet. Anyone who says they do is probably selling you something. (Yes, I see the irony.)

Where this leaves us

Using AI across GTM is an operational change, not a tool purchase. You pick the one metric that keeps you honest. You hand the repeatable work to software. You picture the best possible experience and build backwards. Then you keep testing, because what works today may look different in six months.

What doesn’t change is the part no system can do, which is showing up for the customer.

If you’re a revenue leader working through the same questions, I’d love to compare notes. Find me on LinkedIn and let’s trade what’s working.