Crawl-Walk-Run Discipline Helps Revenue Teams Separate Scalable AI From One-Off Wins

The enterprise AI market in early 2026 is louder and more crowded than ever. Companies are racing to deploy agentic sales platforms and signal-driven go-to-market strategies, leaving buyers to sort through a growing wave of claims and capabilities. For revenue leaders, success comes down to vendor accountability, integration depth, and the operational discipline to prove value before scaling.

Jim Tennant, VP of Channels at Observe.AI, an AI-powered conversation intelligence and contact center automation platform, brings more than two decades of sales leadership experience across cloud and customer experience. Tennant recently grew Observe.AI’s channel sales bookings from 18% to 42% of total new logo business while scaling the channel pipeline to over $5 million per quarter. Before that, he built Replicant's North America channel program from the ground up, generating a $5 million ARR pipeline within four quarters. Across these roles, he has focused on separating deployable technology from vaporware, which guides his approach to the current agentic opportunity.

"Everybody has agentic, and it's now a race to articulate the real differentiators," Tennant says, describing a market where the underlying technology has matured over the past 12 to 18 months, but rapid monetization has turned differentiation into a visible competition. The gap between what functionality exists today and what’s planned in the future highlights the importance of realistic expectations. Enterprise buyers often assume the most advanced large language models are required, when in fact API integration points and legacy systems often determine production success. Buyers need to ask for live demonstrations rather than take promises at face value.

• The pilot trap: Not every rollout requires the newest, most powerful agentic capabilities. Tennant says the best systems succeed when pilots align closely with a specific use case and measurable goals. “You measure a successful pilot by the metrics you set from the start,” he explains. Teams should establish accuracy benchmarks at 30 days and calibration targets at 90, following a crawl-walk-run approach that builds confidence and momentum before scaling. Pilots that link tightly to real-world use cases are more likely to succeed when they expand beyond controlled environments.

• Futures vs. features: Tennant encourages vendors to clearly distinguish what they ship today from what is still in development. “Be honest with the client. Explain why it’s new technology and why it’s important, because this week will be different from next week,” he says. Transparent communication builds trust and strengthens relationships at a time when new enterprise agent frameworks appear frequently, giving proactive vendors an advantage.

The technology moves quickly, making reactive account management less effective. Tennant recalls an implementation where integration issues between a contact center provider and a UCaaS platform delayed progress, leaving multiple teams debating who owned the responsibilities. These situations illustrate an opportunity for vendors to treat proactive communication as a core capability rather than a reactive service. Buyers benefit from partners who provide updates ahead of competitors.

• No more defense: Relying on clients to discover new features through the grapevine and responding afterward can slow momentum. “That’s not going to fly with this fast-moving technology,” Tennant says. Vendors should proactively brief customers on relevant updates and upcoming capabilities, using each touchpoint to strengthen the partnership rather than reacting to complaints.

• Guarding the data layer: Enterprise buyers want assurance that their proprietary data won’t train models for competitors, while also expecting systems to learn and adapt. Tennant says achieving this balance requires clear guardrails, especially as agentic systems evolve from knowledge retrieval into autonomous execution. “These agents are actually learning from each other without human intervention. Guardrails for that kind of autonomous behavior are different from the ones we used to focus on with generative AI,” he explains. Properly managed, agentic systems can accelerate learning and efficiency while minimizing risks tied to messy data.

Shifting from knowledge specialist to autonomous worker adds a new layer of operational complexity. Tennant says this presents an opportunity to strengthen processes as human disorganization, not AI, tends to create the bottleneck. Teams that link pilots to specific use cases and ideal customer profiles and allocate the right human capital to manage workflows can unlock smoother implementations. Clear ownership ensures the technology thrives rather than stalling.

• Start with ownership: Tennant advises setting clear handoff milestones from day one. Clients should define when they intend to take operational control and include a fallback if the transition does not go as planned. "Give me the keys to the kingdom in 90 days and 180 days. If that doesn't work, I'm going to give you the keys back," he says. Clear expectations enable teams to navigate transitions efficiently without pointing fingers.

• SOPs before scale: Organizations that balance AI and human oversight scale successfully. Tennant says revenue teams should establish standard operating procedures that define responsibilities across vendor and customer teams, facilitate internal knowledge sharing, and maintain oversight as technology accelerates. "Build a business practice that’s not only scalable but sustainable, so you’re not constantly hiring people, but rather using AI within your environment to improve operational efficiency," he says.

The next 12 months will likely bring consolidation among suppliers. Tennant expects vendors that struggle to expand early wins into broader deployments to become acquisition targets, reflecting the wider industry push to streamline enterprise AI stacks from agentic enterprise plug-ins to new guidance on building agents. On the buyer side, Tennant references the technology adoption curve, which maps how different groups of customers adopt new technologies over time. He notes that companies focusing on small, targeted deployments will capture outsized returns, while measured experimentation helps buyers keep pace without feeling overwhelmed.

Revenue leaders benefit from treating agentic AI like any other major software rollout, rather than a revolution demanding an all-in bet. Tennant’s framework keeps execution simple because the complexity already exists elsewhere. “Start slow. Crawl, walk, run. Prove it out, prove the model, scale it, rinse and repeat,” Tennant says. “Be realistic with timelines and hold vendors accountable for their execution. If they’re showing progress, be flexible.” In a market filled with noise, the advantage goes to whoever can operationalize first and iterate fastest. “It’s the old elephant adage. One bite at a time.”