6sense’s Kelly Hopping on to Proving AI’s Impact on B2B Pipeline in 2027: The B2BMX Insights Q&A

Published: November 16, 2026

Key Takeaways

  • 6sense CMO Kelly Hopping says CEOs and CFOs will expect proof in 2027 that AI contributes to revenue
  • Because buyers start about 80% of first vendor contacts and 81% already have a preferred vendor at that point, she argues teams should use intent data and buying group signals to decide when to reach out, when to hold back and when a human should take over

Your team can build a full campaign in an afternoon. AI writes the emails, assembles the segments and drafts the follow-ups before lunch. But speed isn’t the same as results. More content, more touches and more automated tasks haven’t reliably produced more qualified pipeline. That’s why sharp B2B marketers are asking harder questions about AI in marketing: What did it actually improve? Where does it make a measurable difference? And where is it just making a weak strategy run faster?

Those questions won’t stay inside the marketing department for long. In 2027, your CEO and CFO will want proof that AI contributes to revenue, not a slide full of hours saved and assets produced. That means tying your AI investments to opportunity conversion, sales velocity, win rates and pipeline per FTE. It also means knowing which tools to keep, which to cut and when more outreach starts costing you deals. Get this right, and you won’t just defend your budget. You’ll set the standard for how AI earns its place in your growth plan.

Few marketers have watched this cycle play out as closely as Kelly Hopping, CMO of 6sense. AI. In this conversation with B2BMX Insights, Hopping explains how to measure AI’s real impact on pipeline instead of reporting vanity metrics, how to recognize the moment a human needs to step back into the buying process and why intent data and buying group signals should change how and when you reach out.

B2BMX Insights: You’ve been a tech CMO five times before joining 6sense. What’s the closest parallel you’ve seen to today’s AI rush, and what did that earlier cycle teach you about the gap between adopting a technology and getting results from it?

Get the latest B2B Marketing News & Trends delivered directly to your inbox!

Kelly Hopping: When marketing automation platforms took off, marketers suddenly had the ability to reach thousands of prospects, automate follow-ups, and generate more leads than ever before. Teams were rewarded for filling the top of the funnel, even when those leads weren’t turning into meaningful sales opportunities. We became so focused on what the technology allowed us to do that we lost sight of what buyers actually needed.

AI creates a similar risk at a much greater scale. We can produce content, personalize emails, and launch campaigns faster than ever. But if the targeting or message is wrong, we’re making an ineffective strategy more efficient.

Across my CMO roles, the lesson has been that technology scales the decisions you make, including the bad ones. The opportunity with AI is to connect what we know about an account, the buying group, and the business problem to a useful next action. That might be a sales conversation, information that helps finance justify the investment, or recognizing that the buyer needs more time. Better context should change what we do, not just how quickly we do it.

B2BMX Insights: You predict executives will demand clearer proof in 2027 that AI contributes to revenue. When your CEO or CFO asks what AI has done for pipeline, what evidence do you show them, and which metrics have you stopped reporting?

Hopping: I think one of the biggest mistakes marketing leaders can make is treating AI adoption as proof of AI’s value. A team might save hundreds of hours or launch twice as many campaigns, but unless that translates into more qualified opportunities, faster sales cycles, or stronger conversion rates, it’s difficult to make the case that the investment is actually contributing to growth.

The question I want to answer is whether AI helped us make better use of our people and investment. If it improved account prioritization, did more of those accounts become qualified opportunities? If it helped sellers prepare for conversations, did opportunities progress faster or convert at a higher rate? The evidence should match the problem we were trying to solve.

A few metrics to look at are opportunity conversion, time in sales stages, win rates, and pipeline per FTE. But be careful about giving all the credit to AI. A stronger quarter after an AI rollout doesn’t prove AI caused the improvement. Wherever possible, compare similar groups over the same period, with comparable sales support.

Productivity still matters. Hours saved can mean lower costs or more capacity for customer-facing work. We just need to show what happened to that capacity. That’s why pipeline per FTE will become the new gold standard, because it’s about efficiency and business results.

I wouldn’t necessarily stop collecting operational metrics. I’d stop presenting content volume, tasks automated, or hours saved as the complete business case. The executive conversation should explain what improved, what it cost, and how confidently we can connect the improvement to the investment.

What B2B Marketing Leaders Should Measure Instead of Content Volume and Lead Counts

B2BMX Insights: Many B2B teams still measure productivity by campaign volume, lead counts and content output. If a marketing leader walked into the next board meeting without those numbers, what should replace them?

Hopping: I’d lead with the contribution marketing is making to efficient growth: qualified pipeline per FTE, conversion to revenue, and the investment required to produce those results. Win rate, average deal value, and sales-cycle length help explain whether the opportunities we’re creating are becoming valuable customers.

Then I’d use a few leading indicators to explain what’s happening underneath. Are we engaging more of the accounts we want to win? Are we reaching the people involved in the decision? Are those accounts progressing toward a purchase? Those measures are useful because they help us decide where to act.

For example, rising engagement with flat opportunity creation should prompt us to investigate targeting or buying readiness. Growing pipeline with stalled deals might point to unresolved questions about ROI or implementation. Each calls for a different response.

I’d also explain how we’re building future demand. Brand, credibility, and customer relationships don’t always translate into pipeline within a quarter. The board needs to understand how we’re balancing near-term performance with the investments that make future growth possible. AI is driving the need for a re-balance of investments between brand (earned, owned, organic) and demand (paid, attributable), because so much of the buying cycle happens outside of a vendor interaction.

B2BMX Insights: You suggest CMOs stop asking how many tasks AI can automate and start asking where it makes the biggest difference. If you had to cut your AI stack in half tomorrow, which capabilities would you keep and which would you let go without a second thought?

Hopping: I’d keep the tools that help us make better decisions about where to spend our time and money. That includes understanding which accounts are in market, who’s involved in the buying decision, and which investments are creating more pipeline that converts to revenue. Those decisions deserve good data and context.

I’d also protect tools that help people do their jobs better, whether that’s preparing for a customer conversation, coaching a rep, or understanding why a campaign isn’t performing. A content tool can be valuable too, if it helps the team produce stronger work without creating hours of editing.

Where I’d get ruthless is duplication and activity for activity’s sake. Our latest BDR research found that outreach attempts per contact nearly doubled, while average quota attainment was more or less stagnant. Before we invest in helping someone send more messages, we should understand what would make those messages more useful to the buyer.

So I’d cut overlapping subscriptions, pilots nobody uses, and tools that create more work to review than they take away. I’d want a clear answer to what each tool improves and how we know. “We’re doing more” isn’t enough.

B2BMX Insights: Where have you seen an AI investment look like a win on a dashboard but fail to move revenue? How did you spot the disconnect, and what did you change?

Hopping: One example that brought this home was a campaign hackathon with our revenue marketing and product marketing teams. They brought together AI tools to see how quickly they could build a campaign. The technology worked beautifully. They produced the plan, the channels, and the execution pieces. If we’d measured success by how much they produced in a few hours, it would have looked great.

But our revenue marketing leader came back and said, essentially, “The process was awesome. The output, not so much.” The campaign was pretty generic. We hadn’t given the tools a clear enough brief or enough context about what we wanted the audience to think, feel, or do differently.

Her takeaway was that we had a process that worked, but we needed to fix the inputs. The next step was to sharpen the strategy and brief before running it again.

That wasn’t a measured revenue failure, but it exposed the gap between producing a great AI project and producing a great campaign. AI had accelerated the work, but it hadn’t resolved the strategic questions.

How to Spot When AI Improves Output but Fails to Improve Campaign Performance

B2BMX Insights: AI lets marketers reach more buyers than ever, yet those buyers are flooded with AI-generated emails and pitches. At what point does more outreach start costing you deals instead of creating them, and how do you measure that damage?

Hopping: More outreach becomes counterproductive when it ignores the buyer’s timing and needs. Our latest research shows that buyers initiate around 80% of first vendor contacts. They have considerable control over when a sales conversation begins. A buying signal should help us make engagement more relevant, not just trigger an email. That’s why context is so critical. A signal without the context to go with it can cause the AI to be confidently wrong, faster.

I’d watch negative replies, unsubscribes, spam complaints, and deliverability alongside positive responses, qualified conversations, and opportunity progression. I’d also compare different outreach frequencies. If additional touches create more rejection without improving qualified conversations, that’s a reason to pull back.

Those indicators won’t give us a perfect dollar figure for lost trust, but they can reveal when more activity is becoming counterproductive. Sometimes the smartest thing marketing can do is recognize when not to reach out.

B2BMX Insights: You’ve said the next competitive advantage is knowing when to put a human back in the process. In a typical B2B buying cycle, where should that handoff happen, and what signals tell you it’s time?

Hopping: I’d put the handoff at the point where we have enough context for a person to advance the buyer’s decision. A demo request or a specific question is the clearest signal. Before that, I’d look for a combination of behaviors: several people from the same account engaging, research becoming more focused on particular solutions, or renewed activity around an opportunity that previously stalled. One content download wouldn’t be enough.

AI should do the preparation, pulling together what changed, identifying the relevant people, and explaining why a conversation might be useful now. Then the seller can judge how to engage and bring something specific to the conversation.

Once that conversation starts, the human role becomes even more important. Buyers often arrive with requirements and a preferred vendor already in mind. They need someone who can work through a technical concern, help their champion make the business case, or resolve competing priorities across finance, IT, and the business.

Those are the moments I’d protect human involvement. AI can prepare the team and handle routine follow-up, but when progress depends on judgment, trust, or getting people to agree, a person needs to own it.

B2BMX Insights: 6sense research has found that 81% of B2B buyers already have a preferred vendor at first contact. How should that reality change the way marketing teams use intent data, and what are most teams still getting wrong about identifying in-market buyers?

Hopping: Our latest research reinforces that preference develops well before the first sales conversation: the pre-contact favorite ultimately wins around 80% of the time. But the implication goes beyond paying attention before someone fills out a form.

More than nine in ten buyers keep current on relevant categories between evaluations. They enter a buying process with experience and opinions already formed. Marketing needs to build familiarity and credibility continuously, while recognizing when a buying group needs more specific help.

Intent data helps us adjust the timing and nature of engagement. It doesn’t tell us that we’ve earned the buyer’s preference. An account can be in market without preferring us.

Teams also need to look beyond individual actions. One person downloading a report doesn’t establish that a company is ready to buy. We need to understand the broader account activity, the people involved, and the decisions they’re working through.

And we shouldn’t assume all that research has moved into LLMs. Buyers use them, but they still rely heavily on experience, colleagues, websites, events, and other trusted sources. We need to be useful across that whole process.

How Intent Data, Human Timing, and Sales Context Improve AI-Driven Outreach

B2BMX Insights: Helping sales prioritize the right opportunities sounds simple, but marketing and sales often disagree on what “right” means. How do you build shared definitions and trust so AI-driven prioritization changes rep behavior instead of becoming another ignored score in the CRM?

Hopping: I’ve always believed marketing’s job is to make sales love us. Not tolerate us, accept us, or occasionally listen to us, but love us because we make their jobs easier and help them win. That’s the standard I’d apply to AI-driven prioritization. Does it help a rep have a better conversation or spend their time more effectively?

That starts with agreeing on what makes an account worth pursuing. Marketing and sales need to define that together, using what we know about customer fit, buying behavior, and previous wins and losses. We also need to agree on what action should follow a recommendation.

Then show the rep why the account deserves attention. What changed? Who’s involved? What might they need? A score alone leaves the seller to do the work of figuring out whether it matters. Useful context gives them a reason to act, especially when it’s available where they already work.

And put those insights into the tools sellers already use, whether that’s their CRM, sales engagement platform, or AI assistant. If they have to leave their workflow, log into another system, and piece together the story themselves, we’re adding work. The recommendation and the context behind it should be right where they can act on it.

Finally, sales needs a voice in improving the system. If a rep says the timing is wrong or the recommendation missed something important, investigate it and close the loop. Trust builds when sellers see their feedback reflected in better recommendations, and those recommendations help them create opportunities and win deals.

B2BMX Insights: Looking ahead to 2027, what practical steps should B2B marketing leaders take now to ensure their AI investments translate into pipeline growth, revenue and stronger customer relationships?

Hopping: I’d focus on three things: start with a clear strategy, run it with trust, and scale it with discipline.

First, pick one specific problem to solve. “Transform our pipeline” is too broad. Choose a workflow where you know what decision you’re trying to improve and what success would look like. Set the measures before you launch, whether that’s better opportunity conversion, faster progression, or a more useful customer interaction. And define what the agent can do independently, what requires approval, and what should make it stop.

Second, give it trustworthy information and enough context to act appropriately. An accurate signal can still lead to the wrong action if the agent doesn’t know the account’s history or what another team is already doing. Keep people involved where the stakes are high, make sure they can intervene, and keep checking results after launch. Performance can deteriorate without the system ever reporting an error.

Finally, expand when you have evidence that something works. Know which agents are running across marketing, sales, and customer success, and what information they’re using. If they’re all working from different versions of the customer relationship, you can quickly create more confusion than value.

Give your teams room to learn and adjust. Putting agents to work doesn’t mean handing over responsibility for the results. That’s on us as humans.

Related Stories

dgr event b2bmx2
Campaign Optimization Series
B2BMX SUMMERCAMP Logo resized
Strategy & Planning Series