From guesswork to gospel: How we made our forecast boring (and 15% more accurate)

Forecasting feels more like astrology than science for too many revenue teams, and our reps are drowning in manual tasks that have nothing to do with selling. With every board member and executive asking, "What's our AI strategy?" the pressure is on not just to adopt AI, but to get it right.

Throughout my career, including my current role leading Revenue Transformation at Gong and previous experience in RevOps leadership roles at companies like Hearst Newspapers and Wpromote, my approach has been a bit more measured. This year, my goal had been to make the forecast boring, because predictability should be the norm, not the exception.

Here’s five tips for how we made it happen:

Right now, every company is wrestling with the same question: To build or to buy?

After coming from a company with the capacity to build, my experience is that you should go through the evaluation exercise regardless of your resources. But the truth is that not every company needs to build its own AI stack. Building can make sense if you have the resources, but it comes with tech debt and long implementation timelines. Buying, on the other hand, can deliver speed and leverage, but only if you choose wisely.

The filter I’ve used is simple: What specific problem are we solving and how does this solution ladder up to our company's OKRs? If the solution can eliminate a manual process — data entry or forecasting roll-ups — then it’s worth serious consideration."

The smartest leaders I know aren’t trying to reinvent the wheel; they’re building credibility by partnering with proven solutions that solve real problems now and will still matter five years out. For enterprise companies, that means building a roadmap and iterating as the AI landscape continues to evolve.

A culture of experimentation and learning is just as important as any technology implementation, but one of the greatest challenges to building that culture is adoption.

And the truth about adoption is that it’s about more than just tracking logins.

From a RevOps perspective, usage data is important, but the real signals are cultural. Are reps leaning in during team meetings? Are they voluntarily signing up for pilots? Are they peer-coaching each other on how to use new platforms and solutions?

We’ve worked hard to foster a culture where trying (and failing fast) is celebrated. That means tailoring incentives to individual motivations. Some reps are driven by career growth, others by recognition, and others by their bottom line. If you can align your AI initiatives with what motivates them, adoption will follow.

One of the most practical applications of the Gong Revenue AI Operating System I’ve rolled out in the past is something we called the “defend your deal” ritual.

Each week, we used AI Briefer to surface our reps’ five riskiest deals based on engagement patterns — not just how old an opportunity is, but what the customer is actually doing (or not doing). Each rep then has three minutes to defend their flagged deal in front of their peers. But there’s a catch: Their defense must be backed by actual customer signals. They can’t lean on hope or “my champion loves us.”

After ten weeks of going through this exercise, our monthly forecast accuracy improved by 15 percentage points. And perhaps just as important, our forecast conversations are no longer about gut feelings; they’re about real customer behavior.

This ritual has also been a powerful onboarding tool for new reps. Every week, they’re hearing how deals actually move through the funnel in real-world scenarios, accelerating their ramp time and strengthening our sales culture.

The best revenue teams aren’t just reacting; they’re orchestrating. But how do you give reps confidence that they know the next best step to take?

Historically, that’s with playbooks, but they often get built in an ivory tower. Marketing, sales, or RevOps creates something they think is helpful, tosses it over the wall, and hopes for the best. Meanwhile, reps are drowning in resources they may never use.

We took a different approach by rebuilding our playbooks with AI at the center.

For example, we created a brief library of closed-lost deals with different AI-generated prompts that reps can run depending on the stage. A new hire inheriting lost deals can quickly understand the deal momentum, see who was involved in the previous cycle, and pull in external signals like earnings reports, new hires in relevant departments, and organizational changes that indicate renewed interest.

The last thing you want is for a prospect to go through the same discovery process twice because your new rep doesn't have context. These AI-powered playbooks give reps immediate success and help them understand not just what happened, but what to do next.

The single biggest shift in buyer behavior we’ve seen is that the buying committee has grown significantly, and it's not shrinking anytime soon. When we analyzed our closed-won deals, we discovered that successful deals now involve eight or more stakeholders.

This insight forced us to emphasize multi-threading from day one — whether it's an MQL or an outbound opportunity. We knew from conversion rates that by the time we're presenting our pitch, we need that extensive buying committee engaged. Waiting until later in the cycle to expand our reach is a recipe for stalled deals.

AI helped us track that engagement across stakeholders and flag risk when too few decision-makers are active. It’s not just about volume of meetings — it’s about breadth of influence. For leaders, this means teaching reps to view their opportunity health through the lens of stakeholder engagement and not just stage progression.

The question isn't whether AI will transform revenue operations. It's whether you're using it to amplify the right things.

Revenue AI is no longer a shiny object; it’s a competitive necessity. But it’s also not magic. The leaders who will win aren’t the ones chasing every new vendor or trying to build an AI empire in-house. They’re the ones who:

• Eliminate manual, low-value work so their teams can focus on strategy.
• Establish clear baselines and align executive teams on measurement.
• Use AI to surface evidence over opinions when forecasting.
• Rebuild playbooks that empower reps to take the next best step.
• Drive adoption through culture and motivation, not mandates.

If we do this right, the forecast won’t be a stressful guessing game. It’ll be boring — and boring is exactly what revenue leaders need.