On a recent episode of [Un]Churned, Gainsight’s podcast about Customer Success, revenue operations, and emerging technology, host Josh Schachter sat down with Mark Roberge to learn about his framework for AI transformation in go-to-market organizations.
As co-founder of Stage 2 Capital and former CRO of HubSpot, Mark has a front-row seat to how AI’s reshaping B2B sales and Customer Success. His insight? Most companies are still stuck at Phase 1, chasing incremental improvements when step-function gains are actually possible.
Here’s his four-phase framework and how to assess where you are and where you should be headed.
Phase 1: Maximizing Customer-Facing Time
Ask yourself: what percentage of your team’s week is spent actually talking to customers versus doing administrative work, attending internal meetings, and prepping for calls?
If you guessed around 25%, you’re right. Some teams are as low as 10%.
“If you literally go from your rep doing 25% selling time with a customer to 75%, and that rep did a million dollars last year, they will do $3 million next year,” Mark explains. “That’s the step function we’re talking about. And that’s very real.”
Phase 1’s about using AI to eliminate the busywork that keeps your team from customers. Here’s what AI can handle:
- Meeting prep by pulling relevant customer context and account history
- Real-time note-taking during customer calls
- Automatic CRM updates and data entry
- Forecast management and pipeline tracking
- Personalized training modules based on individual performance gaps
Where to start: Track your team’s actual customer-facing time. Identify the top three time-consuming administrative tasks. Pilot AI tools that specifically address those bottlenecks. Focus on achieving 3–10x improvements, not just incremental gains.
Phase 2: AI Does the Selling
In Phase 2, AI agents don’t just support your team—they autonomously handle specific customer interactions. Think of it as extending your team’s capacity: renewals, onboarding, and engagement for long-tail segments can be fully automated while your human team members focus their expertise on strategic, high-value accounts that need that personal touch.
“Phase 2 is the selling and onboarding and renewals done by AI instead of a human,” Mark says.
This isn’t science fiction. Companies are already deploying renewal agents that serve SMB segments at scale and adoption agents that proactively engage customers before issues arise. This frees up their teams to focus on complex problem-solving and relationship building.
The challenge? Change management. “We started to see a lot of resistance from frontline workers who thought they were training their replacement,” Mark notes. The key’s running controlled experiments, measuring both efficiency and customer satisfaction, and ensuring your AI policies are in place before rolling out. When done right, AI doesn’t replace your team, it amplifies what they can accomplish.
Phase 3: AI Talks to AI
Phase 3’s when your AI agents start interacting with your customers’ AI agents. Procurement bots negotiating with sales bots. Customer Success agents proactively engaging with customer AI systems.
“Agents buying from agent sellers, agent customers talking to agent customer support,” Mark describes.
This phase feels closer than ever, but Mark offers a reality check: “Everything takes so much longer than we think.” He recounts listening to a futuristic AI podcast while driving through rural New York and passing an Amish horse and buggy. The contrast reminded him that technology adoption’s never uniform or predictable.
Where to start: Don’t wait for Phase 3 to build Phase 1 and 2 capabilities. Monitor buyer behavior for early AI adoption signals and ensure your systems can interface with emerging protocols.
Phase 4: Reimagining Organizations
The final phase’s the most radical: organizational boundaries start to blur entirely.
“The functional boundaries that we know of today—finance, sales, marketing, Customer Success—they start to blur,” Mark explains. “Those boundaries exist because of limitations of human capability. There’s not a lot of people that get a master’s in finance and then write code. We’re just limited, but those limitations go away when we’re agentic driven.”
This raises a provocative question: Are future go-to-market leaders more like RevOps than traditional CROs? The skills that matter shift from managing people to managing agents, data flows, and automated processes.
The Foundation: Get Your ICP Right First
Before any of this matters, you need to solve a fundamental problem most companies get wrong: identifying your ideal customer profile.
The Common Mistake
“Most people look at a sales funnel and say, okay, which segments do we have the highest close rates? That’s so wrong,” Mark says bluntly. “It’s not CAC, it’s not close rate. It’s LTV. Our job’s to go find LTV—that’s where we have product-market fit.”
It’s common for companies to optimize for customers who are easiest to close instead of customers who’ll actually succeed long-term. AI can transform ICP management from an annual refresh to a real-time, dynamic system that continuously identifies which segments have the highest lifetime value.
Where Should You Start?
Most organizations are still in Phase 1—and that’s perfectly fine. The key’s understanding where you are and building the foundation for what’s coming.
Your Quick Self-Assessment
- What’s your team’s customer-facing time percentage? (If it’s under 40%, you’ve got a massive Phase 1 opportunity)
- Are you optimizing for close rate or customer lifetime value?
- Do you have AI policies in place for legal, IT, and security?
- What would a 3x productivity improvement look like for your organization?
Start by calculating your team’s customer-facing time percentage and identifying your biggest time drains. Focus on one high-impact use case where AI can create a step-function improvement—not just incremental gains. Build the policies and infrastructure you’ll need for future phases.
And remember: optimize for customer lifetime value, not just bookings. As Mark puts it: “I think 2026’s gonna be the year to do it.”
Want to hear more insights on AI transformation in Customer Success and beyond? Listen to the full conversation with Mark on the [Un]Churned podcast, where we dive deeper into the future of AI, the changing role of CS leaders, and real-world examples of companies seeing step-function improvements.
