At our recent digital SaaStr AI Day, Russ Fradin, CEO of Larridin, explained how his company helps businesses accurately measure AI adoption within their sales teams. Larridin essentially creates an analytics layer for tracking enterprise AI usage. Their product, Scout, monitors which AI tools employees actually use throughout the organization, how effectively they are used, where there is inefficiency, and what contributes to pipeline generation. Think of Scout as an instrumentation layer sitting atop all your AI investments, providing insights into what’s genuinely working. It is deployed through a browser extension and desktop agent, serving customers ranging from growth-stage startups to large enterprises with tens of thousands of employees. This isn’t vague “AI-first company” talk. It’s real data revealing what employees engage with, what they ignore, and where companies are wasting money. Five key takeaways: 1. More Than Half Your Team Isn’t Using the Enterprise AI Tools You’ve Paid For Across Larridin’s clientele, about 60% of employees licensed for enterprise AI are still using personal accounts instead. Imagine this: You signed the Claude Enterprise agreement, rolled out ChatGPT Teams, and conducted training sessions—yet most of your team logs into their personal accounts anyway. This means you gain no data, no insights, and no understanding of effectiveness. You’ve invested hundreds of thousands in enterprise AI tools but remain largely in the dark about their actual usage. 2. Employees Discover 5-6 Times More AI Tools Than You’ve Approved — That’s Not a Problem, It’s Valuable Intelligence For every AI tool officially recognized within a company, employees have found five or six others independently. For example, a team of just ten had six different AI notetaking apps in use. Russ once joined a customer call early to find four competing AI bots active before anyone else joined. While many see this as a compliance issue, the right perspective is to view it as free experimentation by your top talent. The key isn’t to clamp down but to capture and leverage their learnings before they disappear. 3. Much of Your “AI-Powered Sales Team” Is Merely Using ChatGPT for Basic Searches This insight is eye-opening.

When Larridin assessed true AI proficiency rather than mere usage, they discovered that a significant portion of AI interactions consisted of trivial search queries—checking sports scores or random facts. While fine for everyday life, this does little to drive sales quotas. Without a way to distinguish reps who use AI to create customized account pitch decks from those asking AI mundane questions, you cannot tell if your AI investment is generating pipeline or simply consuming tokens. 4. Stop Labeling Employee Experimentation as “Shadow AI”—It Undermines Trust in Your Best People Russ emphasized this point strongly and rightly so. The term “Shadow AI” implies something clandestine or inappropriate around employee AI use. In reality, your people are diligently seeking any advantage they can. In a startup environment, employee and company success are closely tied. So, instead of treating exploration of new tools as a breach, embrace it as a free research and development program. Welcome effective tools, eliminate redundancies, and transform individual discoveries into playbooks for the entire team. 5. AI Won’t Transform Your Best Reps Dramatically But Will Greatly Improve Your Weakest Reps — and That Matters Most This was the sharpest insight from the session. Every founder managing sales knows the frustration of sifting through leads only to find reps neglecting follow-ups—not your top performers, but the average and below-average ones. AI won’t convert a mediocre rep into a star but can elevate them to solid performers. On a sales team of hundreds or thousands, raising the baseline quality, speed, and thoroughness of follow-up efforts significantly boosts revenue. Shifting the overall performance distribution to the right is where the true value lies.