5 Uncomfortable Truths About AI Adoption That Most VPs of Sales Don’t Want to Hear: The Data From Laridan

Russ Laridan, CEO of Laridan, walked through how they help companies actually measure AI adoption across sales teams at our last digital SaaStr AI Day.

Laridan builds what is essentially an analytics layer for enterprise AI usage. Their product, Scout, tracks which AI tools employees are actually using across the org, how effectively they’re using them, where there’s waste, and what’s actually driving pipeline. Think of it as the instrumentation layer that sits on top of all your AI investments and tells you whether any of it is working. They deploy via browser extension and desktop agent, and their customers range from growth-stage startups to enterprises with tens of thousands of employees.

Not the hand-wavy “we’re an AI-first company” stuff. The real data on what employees are doing, what they’re ignoring, and where companies are lighting money on fire.

Five things that stood out:

1. More Than Half Your Team Isn’t Even Using the AI Tools You Paid For

Across Laridan’s customers, roughly 60% of employees with enterprise AI licenses are still using their personal accounts instead. Think about that. You negotiated the Claude Enterprise deal. You rolled out ChatGPT Teams. You ran the training sessions. And the majority of your team just logged into their personal account anyway. You’re getting zero data from that usage. Zero learnings. Zero ability to understand what’s working. You spent six figures on enterprise AI tooling and you’re flying blind on most of it.

2. Your Employees Have Found 5-6x More Tools Than You Sanctioned. That’s Not a Problem. That’s Intel.

For every tool Laridan knew about internally, employees had found five or six more on their own. With just 10 people, they had six different AI notetakers. Russ joined a customer call early once and was sitting there with four competing bots before any humans showed up. Most companies treat this as a compliance headache. Wrong framing. Your best people are running experiments for you for free. The question isn’t “how do we lock this down.” It’s “how do we capture what they’re learning before it evaporates.”

3. Most of Your “AI-Powered Sales Team” is Using ChatGPT to Google Things

This one should sting. When Laridan actually measured proficiency, not just usage, they found a big chunk of AI activity was basically search queries. Sports scores. Random lookups. Fine for life, useless for quota. Without some way to distinguish between a rep using AI to build custom pitch decks per account and a rep asking Claude what time dinner is, you have no idea whether your AI investment is generating pipeline or just generating tokens.

4. Stop Saying “Shadow AI.” It Tells Your Best People You Don’t Trust Them.

Russ made this point and I think he’s dead right. “Shadow AI” frames employee experimentation as something shady. It’s the opposite. Your people are grinding to find any edge they can. At a startup, your employees literally can’t win if the company doesn’t win. The incentives are already aligned. So stop treating tool discovery like a violation and start treating it like a free R&D program. Bring the good tools into the fold. Kill the duplicates. Turn what one rep figured out into a playbook for everyone.

5. AI Won’t Make Your Best Reps That Much Better. It’ll Make Your Worst Reps a Lot Less Bad. That’s Worth More.

This was the sharpest insight from the whole session. Every founder who has run a sales org knows the sick feeling of going back to a pile of leads and realizing reps just never followed up. Not your best reps. Your average and below-average ones. The ones who let things slip. AI won’t turn a C-minus rep into an A-player. But it’ll turn them into a B. Across a sales team of 500 or 1,000, compressing that distribution, raising the floor on follow-up quality, speed, and depth, is one of the biggest revenue levers most founders aren’t thinking about. The bell curve shifts right. That’s where the real money is.