One of my favorite customers runs a sales team of about 100 people at a B2B SaaS company. Favorite not because of the logo, but because of how hard they push the boundaries of our own platform. They keep challenging us to automate more of how their reps work, and their rollout has become the clearest picture I can give of what Airspeed actually does for a scaled team.
It happened in three levels.
Level 1: Fix the data foundation
This is where we tell every customer to start, without exception. If you don’t build a good data foundation and you try to run AI on top, it will not work.
We sat down with all of their teams, product marketing included, and asked one question: what structured information do you need from your customer-facing conversations? Then we took that data from the calls directly and put it into their Salesforce. Not free-text notes. Structured, reportable fields.
The result: accurate, reportable data across every conversation, sitting in the CRM they already used. No rep typed anything. For the first time, their dashboards reflected what was actually happening on the ground.
Level 2: Make coaching measurable
Like everyone right now, they were going through an agentic transformation. New product messaging, a new way of selling, and 100 reps distributed across the globe who all needed to change how they pitch.
Here’s the problem with rolling out new messaging to a distributed team: you can announce it at every all-hands, every sales kickoff, every weekly meeting, and still have no way of knowing if it’s reaching the calls. You could audit individual recordings if you really wanted to. You could set up keyword trackers. They’d tried; the trackers weren’t delivering.
So we built them fully automated coaching scorecards, each one tailored to their way of selling. A separate scorecard for enterprise discovery calls, demo calls, pricing negotiations. The scorecards checked that MEDDPICC was actually being followed in discovery and demos. That reps were negotiating price correctly instead of handing out discounts willy-nilly. And that the new messaging was genuinely being used, on real calls, not just nodded at in meetings.
Suddenly the leadership team had clear data on how every rep was performing on the things that mattered to them.
Level 3: Bring it together with agents
This is the part that gets me genuinely excited, because it’s where the first two levels pay off.
With structured data underneath, their leadership can now ask agents questions in completely natural language. What happens when this competitor comes up in my calls? The agent digs into exactly the right conversations and comes back with clarity: this competitor appears in every single enterprise conversation, but buyers actually prefer our product; the only gap is brand recognition. So the answer isn’t panic, it’s ABM and enablement. Meanwhile in mid-market, a newer competitor’s product is genuinely beating theirs, so that objective data goes straight to the product team to drive the roadmap.
The same goes for coaching. They configured an agent that compares all their SDRs against each other, looks at scorecards and a sample of calls, and produces a personalized improvement plan for each person. This rep is excellent at personalizing calls but weak on cold-call hooks, so that’s the one skill their coaching focuses on. The manager knows, the rep knows, and you watch each skill tick upward.
Why this works at scale
Small teams can get by on intuition. A leader can listen to a few calls and search a few keywords. But at 100+ reps distributed worldwide, intuition runs out. You have no idea how people are selling your platform or why customers are buying it, unless the data tells you. Quantitatively, reliably, from every conversation.
That’s the sequence: fix the data, make performance measurable, then let agents turn all of it into answers. Each level makes the next one possible. And it’s why we never start with the agents, no matter how much everyone wants to.
Curious what the three levels would look like on your team? Book a demo.
