The Lies We Tell Ourselves
About AI in B2B Sales
Twenty-five B2B revenue leaders walked into a room in London last month with a shared assumption: AI is changing sales.
The morning was structured around three workshop groups, each given a different provocation. Each group was asked to name the lie, the price, and the shift. What came back wasn’t a tidy set of best practices. It was a candid inventory of the assumptions revenue leaders have been operating on that aren’t holding up under scrutiny.
Three lies emerged. All three were uncomfortable. None of them were new – which is exactly the point.
Lie 1: AI will automatically make your sales team better
The most seductive promise in enterprise software right now is that AI will close the performance gap between your top sellers and everyone else. Deploy the tools, give everyone access, watch productivity multiply.
It isn’t working like that.
What the group observed is more nuanced and more troubling. Top performers are using AI as augmentation, they’re getting sharper, faster, more precise. The middle of the sales team is using AI to automate the wrong behaviours and stagnate. And because CRM fields are now full of beautifully constructed AI-generated summaries, the stagnation is invisible from above. Everything looks fine. The pipeline says otherwise.
Gartner predicts that by 2028, AI agents will outnumber sellers by 10X, yet fewer than 40% of sellers will report that AI agents improved their productivity. Deployment is accelerating. Results are not following.
Danna Gurbaxani, CRO at Vita Mojo, named the buyer-facing version of this problem: “I have very low tolerance for laziness and I think our buyers do too. You can feel inauthentic when there’s that over-reliance on AI, you can see the buyers disengaging.”
The board and investors, meanwhile, are expecting 2–3x productivity gains. That math only works if the middle of the team gets meaningfully better. The tools alone won’t deliver that. Deploying AI without structure, without guardrails, without intentional design around where it augments versus where it replaces, that’s not a productivity gain. It’s a performance gap waiting to show up in your Q3 numbers.
This isn’t an argument against AI. It’s an argument for deploying it deliberately. The difference between AI that sharpens judgment and AI that fills in fields is not the technology. It’s the design of the sales motion around it.
Lie 2: More research equals better discovery
The second group set out to discuss account research. They ended up diagnosing something more fundamental.
AI has not solved the research problem. It has inverted it. Sellers used to say “I don’t have time to prepare properly.” Now they say, if they’re being honest “I have seventeen AI-generated summaries and no actual insight.” The pendulum has swung from one failure mode to another. Volume of preparation has been conflated with quality of preparation, and AI has accelerated that conflation at scale.
The scale of this problem is only going to grow. Gartner projects that by 2027, 95% of seller research workflows will begin with AI, up from less than 20% in 2024. Almost every seller will start with AI-generated research within two years. If the lie isn’t named now, the no-decision rate won’t improve, it will accelerate.
The foundational lie: more research equals better discovery. Knowing more facts about an account does not produce better conversations. It produces longer, more confident monologues that still fail to uncover what actually matters.
Mathieu Charles, VP of Customer Strategy at Temenos, named what’s actually missing: “What makes a good deal and a good salesperson is the magic. A good salesperson is never going to walk into a bank and say ‘I heard you have a cost problem, tell me more.’ You need to come with a point of view. You need to shake the tree a little to be respected.”
AI can do the science. It cannot do that.
By show of hands across the group, 40–60% of forecasted deals still end in no decision. Not lost to competitors. Not disqualified early. Worked, researched, summarised — and never actually sold. The group connected this directly to a discovery process optimised for seller confidence rather than buyer justification. Those are not the same thing, and AI has made it easier than ever to confuse them.
The lie was never “don’t do research.” The lie was that research was the bottleneck. It wasn’t. Judgment was. And judgment is the one thing we haven’t figured out how to automate.
Lie 3: Value is in our DNA
The third group started with a question that nobody had a clean answer to: if value selling works, why does it stop at the top 5% of deals?
Most revenue organisations can say “value is in our DNA” and mean it sincerely. They can also point to a motion where 90–95% of pipeline closes on feature comparisons and hope. Both things are simultaneously true, and almost nobody in the room could explain why without recognising themselves in the explanation.
The group identified the reasons with uncomfortable clarity:
- Sellers think involving the value team slows them down
- Enablement trains on product, not business outcomes
- Value selling only works at the right stakeholder level, getting there early enough is a barrier in itself
- Value resources get rationed to the biggest deals
- Nobody has a proven playbook for making it the default motion rather than a specialist service
The cost of getting this wrong is rising. Gartner research shows that by 2030, 75% of B2B buyers will prefer sales experiences that prioritise human interaction over AI. Buyers are moving back towards trust-based, human-led conversations, particularly in complex, high-value deals. Feature-led selling into a buyer who wants genuine human engagement is exactly the wrong motion at exactly the wrong time.
When value arrives late, it looks like vendor justification. The conversation that should have shaped how the buyer thinks about the problem becomes a slide deck to get the deal over the line. The trust that value selling is supposed to build gets undermined by the moment it arrives.
The honest conclusion the group landed on: this is emerging thinking, not a solved problem. Value selling works brilliantly when done well. It stays locked inside a tiny fraction of the business. And the gap between those two things isn’t a resource problem or a skills problem in isolation – it’s a motion design problem. The sales process as built makes feature-led selling the path of least resistance. Until that changes, “value is in our DNA” remains a statement of aspiration rather than a description of how deals actually get done.
The thread running through all three
These lies persist not because revenue leaders are naive. They persist because the systems, incentives, and tools built around B2B sales make them the rational default.
Automating activity is easier than developing judgment. Generating research outputs is faster than forming a point of view. Deploying value selling on the biggest deals is safer than redesigning the motion for all of them.
The room didn’t leave with a clean set of answers. What it left with was a clearer picture of the questions worth sitting with — and a shared recognition that naming the lie is the necessary first step before anything else can change.
The full event report here, including what the room said the price and shift needs to look like.
