The Data Problem Killing Your AI ROI

Most conversations about AI in sales start in the wrong place. They start with tools. The better conversation starts with data. 

That was the thread running through our June 3 session with Chad O'Connor (CRO, Backstory) and Nick Keeslar (AVP RevOps & Enablement, Aidoc). AI sales analytics only works when the data underneath it is complete. 

Here are the four things that stuck.

What Makes AI Sales Analytics Actually Work?

AI sales analytics works when three conditions are met: activity data is captured automatically and completely, insights arrive in the tools reps already use, and the system carries real deal context instead of summaries of summaries. When any of the three is missing, the output is confident noise. The tool is rarely the problem. The data foundation is.

AI is working. But not for the reasons most people think.

Where AI is genuinely delivering for revenue teams right now: account research, deal context synthesis, and rep prioritization. Not because the technology is magic. Because someone made it easy enough that reps don't have to think about it. 

Chad put it plainly. The teams seeing results push information to reps in the workflows they're already in, like email and Slack. They don't ask reps to log into another platform. The ones struggling built for the power user. The rep who already knows what they want and will go find it. That's not most people on your team.

Nick's framing: there is more variance in AI literacy on a sales floor right now than in any other skill. More than Salesforce proficiency. More than methodology. If you build for the most sophisticated user, you haven't solved adoption. You've just built something your ops team could have made themselves.

The corpus is the asset. Not the tool.

This was the sharpest idea in the session, and it came from Nick's CCO at Aidoc: "I need you to build the best corpus of data, because every day that goes by that the corpus isn't getting bigger, we're losing value."

The corpus is every email, call, meeting, and stakeholder touch across every deal. It is the raw material AI actually runs on. And here's what makes it urgent: you can always find a better AI tool in the future. You cannot get back the days where the data wasn't captured. That time is gone.

This is exactly the problem Backstory is built to solve. Most teams are asking AI to coach, forecast, and prioritize with massive gaps in their activity data. CRM data only reflects what reps remembered to log. It's not a technology problem. It's a foundation problem. A revenue intelligence platform built on complete capture closes that gap. 

Chad made it concrete. When a rep closes out a deal in frustration, or a deal gets transferred between accounts, the history disappears. What looks like a 90-day sales cycle is often a two-year relationship with no recorded context. You can't replicate your best reps if you don't know what they're actually doing. Sales performance analytics built on partial history will tell you a partial story. 

Stop planning for the year. Plan for the quarter.

Both Chad and Nick came back to this repeatedly. The market is moving too fast for a 12-month AI roadmap. The teams winning are running 90-day cycles. Make a bet, ship it, learn from it, iterate. 

Nick's framing for RevOps leaders who feel behind: run two swim lanes simultaneously. One for building the corpus by capturing more activity data. One for building tooling on top of it. Don't wait for the corpus to be complete before you start building. But don't skip the corpus investment either.

And for CROs: sit down with your RevOps leader this week. Align on what you're trying to learn or prove in the next 90 days. That's the conversation that matters right now. Not the annual plan.

The CRO–RevOps relationship has changed. Most people haven't caught up.

Chad called Nick his "work wife." Nick said if you're not aligned with your CRO, you should probably find a new job. Both were serious.

The best RevOps leaders aren't building dashboards on request anymore. They're anticipating. They show up with the answer before the question gets asked, the same way a great point guard already knows where the pass is going. That requires trust, daily communication, and a CRO who's willing to let the data be the emotion-free third voice in the room.

It also requires shared pipeline visibility. When both leaders see the same picture of deal health and engagement, the conversation moves from defending numbers to making calls. 

Nick's advice to RevOps leaders: put on the chief-of-staff hat. Notice the things your CRO mentions three times in a week. Those are the problems worth solving without being asked. That's how you build the kind of relationship where real progress happens.

Watch the full session on demand →

Follow Backstory on LinkedIn for more from the team.