Host Richard Ellis interviews Jack Siney, co-founder of Front Race, on why AI needs activity/interaction data instead of biased “opinion” data entered by reps, and how rigid, linear CRM workflows miss the nonlinear reality and micro-details of sales, including timing between actions. Jack warns that automating only documented steps can fail because top performers execute additional undocumented steps, and that LLMs can mis-join data and hallucinate without hard metrics. He recommends leaders: consolidate and standardize company data, map the real end-to-end process (including losses), and use a measurement layer to benchmark and evaluate changing AI tools.
Soundbites
- "All the LLMs are using the same open data set. The magic is in your data. You have the answers. If you have a couple of years of legacy data, you've had some success, you've had some failures. The magic is in that data: uncovering what works and what doesn't."
- "We've been measuring metrics for four decades that have no direct correlation to whether we hit the goal."
- "We get rid of our SDRs to automate 20 steps. But the SDRs are actually doing 32 steps. We train the agent on 20 because we don't even know the other 12 exist. Then we wonder why the pilot failed."
- "The magic's in the micro details. Everyone knows the big things: the culture, the pitch, the pricing, the demo. That's not what separates your best reps from your average ones. It's 20 little things."
- "As soon as you rely on the sales rep to put the data in, we're in trouble. Their job depends on having a good pipeline. They're biased. Garbage in, garbage out."
- "Automate the interactions and you get what really happened. It's not someone's opinion."
- "Some of the magic is the time in between the steps. Letting it breathe. I don't know a system out there that tracks the time between each call."
