You don’t need someone super technical on your marketing team to run AI across your funnel. Over the past year, I’ve built a set of always-on go-to-market plays that run across marketing, sales and customer success, and the real advantage is speed. We can launch a new play, test it, and turn it off the moment it stops performing.
Here are four plays running right now, including the one that’s driven more revenue than anything else we run, plus the rules I follow when I build a new one so you can do the same.
How I build a play
Start with the idea, not the tool. Each play started with a problem I wanted solved, like “the event won’t give us the attendee list.” AI lets non-technical people turn that idea into something that runs.
Filter everything through the ICP. Every play checks against my ICP criteria before anything happens, so reps only see accounts worth their time.
Give the AI context. I mainly use @Swan AI and it’s connected to @Hubspot , @Fathom.ai, @PostHog, our reps’ LinkedIn Sales Navigator and our ad account. That context is why the outreach doesn’t read like AI.
Keep a human where the team wants one. Our AEs wanted right of refusal on anything sent in their name, so they review and approve. Everything else runs on its own.
Build once for the whole team. When I change a play, it updates for sales, CS and marketing at the same time. Nobody has to set anything up on their end.
Play 1: Closed-lost re-engagement
Trigger: A deal closes lost.
What it does: Swan looks at three groups: the people tied to the deal, people in HubSpot who weren’t on the deal but would count as buyers, and movement on LinkedIn, such as new hires or people moving into buying-committee roles. Then it drafts a multichannel re-engagement across email and LinkedIn using everything it knows from HubSpot and our Fathom call recordings.
Who acts: The AE reviews and hits approve. That’s it.
Result: Over a 60% open rate and over a 38% click-through rate. People reply thinking a human wrote it. This play has driven the most revenue of anything we run, and I can tie closed-won revenue to it.
Play 2: High-intent website visitors
Trigger: Someone visits our pricing page or our demo library
What it does: Swan de-anonymizes the visit and checks the company against our ICP. If it fits, it finds the buying committee and their contact info, adds new contacts to HubSpot, updates existing ones and drafts the outreach. Those lists can also become ad audiences.
Result: This is the play that runs the most.
Play 3: Event attendee monitoring
Trigger: We sponsor an event, and the organizer won’t share the attendee list in advance.
What it does: Every morning, Swan checks LinkedIn for anyone mentioning the event. If they’re in our ICP and attending, it slacks the rep who’s going a direct link to the post and a recommended comment. It also keeps a running list of who’ll be there.
Who acts: The rep comments and connects before the event. A week or two out, we use the list to decide who to invite to dinner.
Result: I get a slack every morning telling me it ran and found a couple more people.
Play 4: Expansion signals for CS
Trigger: Customer visits the in-product search page
What it does: Alerts our CSMs about which customers to pitch additional functionality to with customer research and drafted messages complete.
Result: The CSMs get a package to make upselling easy at the time when cusomers are already thinking about it.
What changed
Our AEs don’t spend time digging through closed-lost deals or researching website visitors. They review, approve and show up to the conversations that matter. It’s a big reason our head of sales says each person on her team can hold more pipeline.
Key metrics
60%+ open rate and 38%+ click-through rate the closed-lost re-engagement campaign
4 always-on plays across marketing, sales and CS
0 manual research for AEs on closed-lost deals and high-intent visitors