Enterprise AI strategies fall apart in the space between a successful demonstration and daily work. Tom Gersic has spent much of his career working in that gap. During his 12 years at Salesforce, where he served as VP of Product Adoption, and later as SVP of AI & Data GTM at an OpenAI implementation partner, he helped organizations turn new technical capabilities into tools people would trust, adopt, and use. He now brings that experience to YouEx.ai, the AI-native lead-to-revenue company he founded. Gersic argues that leadership teams often frame AI too narrowly as a technology investment when the larger challenge is changing how work gets done. But the pace of technical change is not just a source of disruption. It is an opportunity to make the organization better at adapting. Every workflow an organization redesigns becomes practice for the next.
Bringing Frontline Teams into Strategy
Boardroom discussions tend to spend far too much time focused on sprawling lists of use cases and impressive demos. Gersic finds this focus misguided. Capable AI is becoming broadly accessible. The real challenge is integrating it into actual workflows, managing ongoing organizational change, and shaping the user behavior that turns technical capability into business value. “Honestly, the notion of an ‘executable’ strategy is part of the problem. It assumes the strategy is done and the teams just have to carry it out. The ones that actually work aren’t finished when they leave the boardroom,” Gersic explains. When plans are finalized in isolation, they rarely account for the real friction that staff face every day.
Real change begins by looking at a process from end to end and identifying where human judgment matters most. As Gersic notes, “The real value of AI comes from redesigning the workflow around it, and the people who own that workflow are closest to where judgment, friction, and workarounds actually live.” During a major rollout for a consumer brand, his team ran workshops where employees scoped their AI projects and pitched them live to executives. Leadership defined the outcome, priorities, and guardrails. The people closest to the work helped redefine the workflows. Giving workers a direct role in creating the solutions meant they didn’t have to be persuaded to use them later.
Recognizing Behavioral Signs of Adoption
Basic login telemetry tells leaders whether people opened an application, but it doesn’t reveal how well the application has integrated into daily habits. Gersic points out that teams must first define a primary business outcome before trying to evaluate any usage metrics. Once that target is in place, adoption shows up across four key behaviors: retention, expansion, organic demand, and subtraction. When employees stick with a tool for months and apply it to tasks beyond their initial training, the software is delivering real value. Organic demand appears when people begin asking for access or new capabilities before leadership pushes them. Subtraction is often the clearest indicator that an organization has embraced a new tool. Healthy adoption leads teams to retire old manual spreadsheets, stop running duplicate reports, and cancel unnecessary update meetings. The tone of day-to-day feedback changes as well. “When you stop hearing ‘I don’t know how to use this’ and start hearing ‘Why can’t it do X yet?’, that’s a great sign,” Gersic says. “It means AI isn’t something they’re being asked to adopt anymore. It has become a part of how they work.”
Designing for Human Judgment
User experience is often misunderstood as visual design, but it really represents the complete interaction someone has with a system. While deploying an AI sales assistant inside a financial services firm, Gersic observed that the original bot answered questions accurately while missing the real context behind them. To fix the issue, the team spent time studying how the firm’s top seller handled customer conversations. They then built that practical judgment directly into the tool so the rest of the team could draw on that expertise. By centering the software on how top performers actually work, the company created natural interest among staff. “The experience layer isn’t the interface. It’s where you encode the judgment of the people who are already best at the job. Find them first, then design around what they know,” Gersic says. He took that same approach with YouEx.ai, designing the web agent and revenue automation as one lead-to-revenue workflow rather than several disconnected tools. Context captured in a visitor conversation follows the lead into research, qualification, routing, follow-up, and opportunity management.
As companies move from single tools to fleets of autonomous agents, fragmentation rapidly becomes a serious risk. Sales reps often spend their time manually copying notes between disconnected systems, essentially working for the software instead of letting the software handle the routine tasks. “The goal has to be one workflow, not ten tools,” Gersic warns. “The user should experience the fleet as one coherent system that gets the job done.” Without clear ownership over the handoffs between tools, the entire process breaks down, and staff revert to manual habits. Success in this environment comes down to how quickly a business can take workflows apart and rebuild them around new capabilities. Spending six months planning a rollout guarantees you deliver something the market has already moved past. “Speed is the part people tend to underestimate,” Gersic says. “An organization that has redesigned five workflows this year is faster at the sixth.” That doesn’t mean bypassing governance. Gersic notes that “clear governance allows teams to experiment and improve faster because they know where the boundaries are”. Teams that build this habit gain a lasting advantage, because they can adjust their operations as fast as the underlying technology evolves.
Follow Tom Gersic on LinkedIn for more insights on AI product adoption, workflow design, and building effective enterprise AI strategies.
