Why AI in go-to-market has crossed a new threshold

Let me start with something that probably sounds familiar.

A few months ago, I bought a Whoop. Every morning I'd wake up and get a little report. Heart rate, REM sleep, recovery score. Seventy-three percent recovery. And I'd sit there thinking, okay, what does this actually mean? What am I supposed to do with seventy-three? How do I get to ninety?

The information was there. The action wasn't.

Now think about your GPS. I turn mine on every single morning, even though I've driven the same route to work hundreds of times. I do it because if there's an accident or a road closure, the system reroutes me automatically.

It doesn't just tell me there's a problem. It makes a decision, gives me five seconds to override it, and then gets on with it.

And then there's autopilot. A system that's simultaneously managing speed, routing, and lane positioning. You just provide the destination. It handles the complexity.

Go-to-market is going through exactly this kind of transition right now. We've moved from AI that observes and reports to AI that actually learns and executes. That shift is more significant than most people realize, and it's happening faster than most teams are prepared for.

The uncomfortable truth about where we are today

Here's a number worth sitting with. According to Salesforce research, sales reps spend only around 28% of their time actually selling. The rest goes to administrative work, meeting prep (which alone accounts for roughly 15% of their time), internal pipeline reviews, deal reviews, and various other activities that don't directly move revenue.

We've bought more tools than ever. Ten tools on average. And yet win rates across B2B have largely stagnated somewhere between 17 and 20%.

So what's been happening? We've been taking technology and using it to help reps do more of what they're already doing. Faster outreach. Quicker note-taking. More dashboards. The underlying motion stays the same, just slightly accelerated. That gets you 10 or 20% efficiency gains, maybe. But it doesn't move the needle on the things that actually matter.

That's the trap a lot of organizations are sitting in right now.

Three generations of go-to-market technology

To understand where we're headed, it helps to understand where we've been.

Gen one was Salesforce. Revolutionary at the time. It gave CROs and sales leaders a system of record. How many deals are in flight, what's set to close, what's the pipeline. Genuinely useful, but fundamentally passive.

Gen two is what most of the world is operating in today. Tools like Gong, Clari, and sophisticated dashboarding platforms moved us from record-keeping to insight generation. A call happened; here's what was discussed. Based on your CRM data, here's your forecast and whether you're likely to hit it. This generation gave us visibility we didn't have before.

The problem is that most organizations are still trying to build on top of Gen Two. They're investing in amplifying insights. Meanwhile, the technology has already moved somewhere else.

Gen three is what I'd call a system of execution. It doesn't observe. It analyzes and acts. People talk about agentic AI and AI-native platforms, and that's essentially what this is. The distinction matters because it fundamentally changes what's possible.

What a system of execution actually looks like

A genuine system of execution has three distinct components.

Context. This is the foundation. It means connecting with all your go-to-market systems, not just call recordings, but sales enablement content, product knowledge, technical documentation, emails, everything. The system builds a deep understanding of your business, your motion, and how your reps actually sell. Without this layer, everything else is surface-level.

An agentic layer. This is where the actual execution happens. Let me give you a concrete example.

A rep gets on an early discovery call. They're relatively new. They're not sure which questions to ask, they struggle to navigate toward the real pain, and they default to feature selling. That's the old scenario.

In the new model, AI joins the call. It's listening to what the customer is saying in real time, and it's guiding the rep on the side. "Ask this question; it'll surface the pain point this product addresses." Or, "they just mentioned compliance, that's a major use case for us, follow up on that."

Or take deal execution. A rep finishes a call where the customer mentioned budget concerns twice and went quiet when the enterprise tier came up. The rep logs a note saying "all good, following up in two weeks" and moves on. A system of execution notices what the rep didn't.

It flags the budget signals, recognizes the pattern from similar deals, and knows that in situations like this, an ROI document or a relevant case study tends to be what moves things forward. It drafts one and asks the rep to approve it. That's the difference.

The ability to learn. Think about a team running 500 meetings in a week. What's actually happening in those conversations? Where are deals getting stuck? What objections are coming up repeatedly? What approaches are working in certain segments?

Today, most of that knowledge evaporates. It lives in individual reps' heads, or it requires someone to know the exact right question to ask a search tool to surface anything useful.

A system of execution feeds all of that data back into the shared context. Every agent is learning from what happened across every deal. The insights from a deal that closed in Europe this week can be in a US rep's meeting prep within the hour. That's a fundamentally different kind of organizational learning.

Where AI can actually move the needle across the funnel

Right now, most of the energy and investment in AI is going toward top-of-funnel activity. Hyper-personalized outreach, better reply rates, more pipeline generated. And those results are real. Getting from a one to two percent reply rate up to something meaningfully higher is worth doing.

But there's a part of top-of-funnel that tends to get overlooked: discovery. Most CROs I talk to will tell you, without much prompting, that their reps aren't doing great discovery. And that problem compounds. Weak discovery means deals that shouldn't be in the pipeline end up there, and then they drop later, dragging down win rates across the board.

AI can make a real difference here. Discovery is consultative by nature. You're trying to understand whether you can genuinely help the customer. AI is well-suited to surfacing the right questions at the right moment, in a way that actually moves the conversation forward.

Mid-funnel is where the most expensive leakage happens. Reps are scheduling recurring meetings, prepping for every call, trying to figure out their next move. Deals stall. Momentum gets lost.

AI can generate genuinely useful meeting prep based on everything that's happened in the deal so far, combined with patterns from across your entire deal history. It can flag stakeholder dynamics that need attention, identify risk signals the rep might be rationalizing away, and suggest concrete next steps.

Reps tend to be optimists. That's often what makes them good at the job. But it also means they sometimes need someone (or something) to show them honestly where a deal might be heading south.

Bottom-of-funnel is where the gap between your top performers and everyone else becomes most visible. Your best reps close at roughly twice the rate of average reps. They know how to build a compelling proposal. They've done the work to understand the customer deeply.

They frame business value in a way that resonates. With AI, that capability can be democratized. The system can analyze the conversations, understand what the customer cares about, and draft proposals the way your best reps would write them. You're essentially spreading what your top performers do across the whole organization.

The honest assessment: top-of-funnel is where most of the focus is today. Mid and bottom-of-funnel is where the real opportunity sits as you look ahead.

Three ways to approach building this capability

When you're thinking about how to actually acquire or build a system like this, there are three paths worth understanding.

AI-native CRM. This means replacing your existing CRM with one that has AI functionality built in from the ground up. It works well for smaller organizations, particularly those coming off a Series A who are moving from spreadsheets to a real system for the first time.

For larger organizations, it's a harder case to make. Your existing CRM holds customer history, institutional knowledge, legal and procurement workflows. Ripping and replacing that is a significant undertaking.

Point solutions on top of your existing CRM. This is where most organizations are today. You pick best-in-class tools for deal coaching, forecasting, conversational intelligence, and layer them on top of Salesforce. It's a legitimate approach.

You can show wins quickly and get from zero to sixty fast. The ceiling you hit eventually is that these tools don't share context. Your coaching tool doesn't know how the deal is progressing. Your forecasting tool doesn't know what happened on the call this morning. At a certain scale, that fragmentation becomes limiting.

The platform route. This means going with an AI-native platform that centralizes your data and breaks down those silos. Every agent is working from the same context. What the coaching agent learns informs what the deal execution agent does. The insights compound over time. This is the approach that, in my view, delivers the most durable productivity gains.

The caveat worth being honest about: it requires adoption. A bad product with 90% adoption will outperform a great product with 20% adoption every time. So whatever platform you choose needs to integrate with where your reps already live. It needs to sit in Slack, in email, in interfaces that feel familiar. The lift to start using it has to be low, or it won't happen.

The difference between conversational intelligence and execution AI

This comes up a lot. When we talk to prospects, we often hear something like, "We already have [conversational intelligence tool], isn't this similar?"

There are a few meaningful differences.

Many of the established conversational intelligence tools weren't built AI-native, which limits how quickly they can evolve and what they can do. More importantly, they don't have the full context of your business.

Here's a real example. A customer asked me how they'd explain the difference between what we do and a conversational intelligence tool to their team. I told them to go ask the other vendor this question:

"Can you analyze a deal I'm currently running, and based on what's happening, tell me which product to position given the competitors in play and how to position it?" They went and asked. The answer was no. Because the tool doesn't know what the products do, doesn't know the sales enablement content, doesn't know the technical documentation. So the guidance it can offer stays at the surface level.

If you use a conversational intelligence tool for coaching, it might tell you that you spoke more than the customer, or that you should pause more, or listen better. That's hygiene. It's useful. But it's a different category of insight from, "when they raised compliance concerns, you should have positioned our governance product, here's why."

The other difference is timing. Conversational intelligence tools work after the fact. You review the tape after the game. The new generation of AI can be live on the call with you, in the moment, before you miss the opening.

We experienced this directly a few weeks ago on a call with the worldwide head of enablement at a large Fortune 500 company.

The conversation was going nowhere. They had Gong, they'd spent significant money on it, adoption was low, and they weren't sure what to do with it. Our live assist was running during the call, and it surfaced a prompt: ask about the renewals team.

That thread had come up in an earlier conversation with someone at the company. So we asked. Turns out the renewals team was full of relatively new sellers who struggled to position the product on calls. By the end of the week, we were meeting their head of renewals.

That's the kind of moment that happens on every call. Having something in your ear that's tracking the full context of the relationship, not just the current conversation, changes what's possible.

A few things to get right as you move forward

If you're heading down this path, a few practical things are worth keeping in mind.

Start with the data you have. Don't wait for your CRM to be perfectly clean. It probably never will be. Start with your conversation data, your emails, whatever product information you have in your systems. Within a week, you'll have a clearer picture of what objections are coming up, how they're being handled, what's working. That flywheel starts turning faster than you'd expect.

Sequence your use cases. There's a natural temptation to try to implement deal intelligence, live call coaching, forecasting, and rep coaching all at once. Prioritize one, show success, then move to the next. That sequencing matters more than most people think.

Define success before you start. Win rate. Ramp time. Deal progression from stage two to stage three. Pick the metric that matters, establish a baseline, and measure against it. If the tool isn't moving that number, have an honest conversation about why. Too many implementations drift into "well, we're getting some insights" without anyone being able to say clearly whether the investment is working.

Where this is all heading

You might have come into this thinking about AI primarily as a productivity tool. Something that helps reps do more, move faster, handle more volume.

The more accurate framing is that AI is becoming the operating system of your go-to-market motion. It can help you execute deals more tightly, spread the capabilities of your best reps across the whole team, and build organizational learning that compounds over time rather than evaporating after every call.

Every motion that transitions from analyzing to executing is going to have a structural advantage. That transition is already happening. The question is where your organization is on that curve, and how quickly you're moving along it.