Get a feel for the Enablement Edge Operating System.

Revenue Leaders Are Raising Seller Expectations as AI Faces a Proof-of-Value Test

New research on AI, seller productivity, sales training, and GTM trends – and what they mean for revenue and enablement teams heading into 2027.

At a glance

BCG found that 91% of revenue leaders expect higher seller quotas or productivity requirements, even as 65% plan to increase headcount and 72% expect larger sales and revenue budgets.

AI adoption in sales is already widespread, but BCG found a much clearer effect on administrative productivity than on conversion, win rates, or quota attainment.

An 18-month study of 119 new sellers found that AI-heavy practice produced strong execution, but sellers were less adaptable when buyer conversations changed direction. A combined approach performed better overall.

A separate study of 300 B2B software and AI companies found that relationships remain unusually important to GTM: networks and introductions were the most commonly cited core growth channel.

Forrester’s latest Revenue Enablement Platforms research suggests that the technology itself may be moving ahead of organisations’ ability to operationalise it.

Revenue leaders appear to be entering 2027 with a fairly demanding equation.

They expect to invest more, grow their teams, and use more AI. At the same time, they also expect considerably more from each seller.

The question is whether those different bets are moving at the same speed.

New research from Boston Consulting Group suggests that AI is already producing useful productivity gains across sales organisations. Its contribution to conversion, win rates, and quota attainment is much less established.

Elsewhere, new research into AI-based sales training raises a related question: as we make practice easier to scale, are we developing sellers who can execute a known pattern better, or sellers who can adapt when the situation changes?

Those are different problems, but they point towards a similar challenge for revenue and enablement teams heading into 2027.

Technology capability is advancing rapidly, yet turning it into stronger commercial performance remains the harder part.

Revenue leaders want more people, more investment – and more from each seller

BCG’s new 2027 Revenue Agenda surveyed 398 CROs, sales leaders, chief commercial officers, and Revenue Operations leaders across 23 countries.

The growth expectations are striking.

Sixty-five percent expect revenue-organisation headcount to grow over the next year. Seventy-two percent expect sales and revenue budgets to increase. And 91% expect quotas or productivity requirements per seller to rise.

So this isn’t really a story about organisations replacing sellers with AI.

At least according to this research, many revenue leaders plan to add capacity while expecting greater output from it.

AI is clearly part of that calculation. Ninety percent of respondents said they already use AI meaningfully in sales workflows.

A third of respondents reported reduced administrative workloads. The commercial effects were much less common: 14% cited improved conversion as a leading measurable impact, 13% improved win rates, and 8% higher quota attainment. Just 6% said they had reached autonomous, end-to-end AI deployment.

That suggests most organisations are still seeing the clearest gains in the work around selling rather than in sales outcomes themselves.

Removing CRM updates, research, preparation, and other administrative work can create meaningful capacity for sellers. The more important question is what organisations do with that capacity, and whether it eventually results in stronger commercial performance.

Airspeed’s vendor-produced 2026 GTM Maturity Report makes a similar argument from a different angle, putting revenue per head rather than AI adoption at the centre of the discussion. Its underlying point is worth considering even if its customer performance data should be treated as provider-reported evidence: adoption alone tells us very little about whether the commercial system is actually producing more.

BCG’s research makes the issue more interesting because buyer expectations are also moving.

Nearly 60% of the revenue leaders surveyed said buyers now arrive at their first conversation knowing significantly more about products, competitors, and potential use cases than they did two years ago. Fifty-four percent reported greater demand for customer- or industry-specific personalisation, while 51% said buyers increasingly expect customised proof of value before purchasing.

This changes what increased seller productivity needs to mean.

If AI gives someone five more hours, but those hours simply disappear into a busier week, very little has changed.

If that capacity allows the seller to spend more time understanding a buying group, building a stronger business case, bringing relevant proof into the conversation, or working through internal buyer disagreement, the effect could be very different.

For enablement teams, I think that makes capacity allocation an increasingly important part of the AI conversation.

The conversation after deploying AI shouldn’t stop at: “What did we automate?”

It should eventually move towards: “What are sellers now able to do better because that work has been removed?”

AI practice may improve execution faster than adaptability

A different piece of research published on 1st October looked at what happens when AI moves directly into seller development.

Corporate Visions, working with Dr Leff Bonney and the Florida State University Sales Institute, followed 119 new inside-sales hires over 18 months at a large construction-supply company.

Participants were placed into three training conditions.

One group relied heavily on repeated AI role-play. Another spent more time observing experienced sellers in real conversations and evaluating their behaviour against a defined rubric. A third combined the two approaches.

Towards the end of the study, sellers completed a live role-play in which the buyer either followed the expected conversation or changed direction.

The AI-heavy group performed strongly on execution when the conversation followed the expected path. When the buyer introduced an unexpected change, their adaptability score fell to 25.4 out of 50.

The observation-heavy group performed best on adaptability.

The combined group produced the strongest overall balance, scoring 46.2 on execution while maintaining 34.7 on adaptability. Managers subsequently rated the combined group highest across the two dimensions at both 30 and 60 days.

There are obvious limits to how far we should generalise this.

It is one study inside one company, and Corporate Visions itself is a sales-training provider. The findings are useful evidence, rather than a final answer on how AI should be used in sales development.

AI role-play still solves a very real problem: practice is much easier to scale when sellers do not need to wait for a manager, facilitator, or colleague every time they want another repetition.

The risk is assuming that more repetitions automatically produce greater readiness.

A seller can become very good at handling a scenario they recognise. Real sales conversations require something else as well: noticing that the situation has changed, deciding which part of the learned approach still applies, and adjusting without losing the objective of the conversation.

That suggests a more useful way of thinking about AI practice.

Use AI to create far more opportunities for repetition and feedback. But deliberately expose sellers to real examples, variation, ambiguity, and situations where the expected response is no longer the right one.

If AI makes practice almost unlimited, enablement may need to become much more thoughtful about what kind of practice we are scaling.

GTM is becoming more AI-enabled, while relationships stay critical

There is another interesting counterweight in the 2026 State of B2B GTM report from Maja Voje and Kyle Poyar.

The research surveyed 300 B2B software and AI companies.

AI’s reported contribution has increased significantly: 43% of respondents said AI was having a big impact on their GTM organisation, up from 24% in the previous year’s study.

But some of the strongest channels in the report were distinctly human.

Networks and introductions were the most widely used core GTM channel, cited by 62% of respondents. Founder-led sales was the most common primary GTM motion at 26%. For companies selling above $50,000 annual contract value, networks and introductions ranked first among the top channels, followed by intimate events.

The investment plans for 2027 make the combination even more interesting.

Half of respondents plan to invest more in AI search, 48% in signal-based outbound, and 42% in intimate events.

This seems to point to the fact that as organisations automate more of the work involved in identifying accounts, gathering information, creating content, researching buyers, and coordinating outreach, trusted access and high-quality human interaction may become more valuable rather than less.

There is also a practical implication for enablement.

If networks, introductions, founder involvement, events, and other relationship-led motions continue to matter, seller capability cannot become exclusively about operating increasingly automated workflows.

People still need to build trust, navigate conversations, understand context, make good judgements, and give buyers a reason to involve them when information itself is abundant.

That connects directly back to the BCG findings.

More informed buyers do not necessarily make the seller less important.

They may simply make low-value seller interactions much easier to avoid.

The enablement technology gap may increasingly be operational

One smaller signal from this period came from Forrester’s latest Revenue Enablement Platforms Wave.

Forrester evaluated nine providers and interviewed 22 vendor customers. Peter Ostrow wrote that the biggest surprise was not the level of AI capability now available, but how much functionality customers had still not operationalised.

The reference customers reported using approximately 75% of the capabilities available to them. Change management, competing priorities, limited administrative capacity, and organisational readiness were among the reasons functionality remained unused. Frontline managers received the lowest adoption rating among the stakeholder groups assessed. 

Obviously, these were vendor reference customers, so they should hardly be treated as a representative benchmark for the entire enablement market. 

But if anything, the fact that even this relatively selected group reported leaving a meaningful amount of functionality unused makes the adoption problem worth paying attention to.

The technology discussion is becoming less about whether a platform can support content, training, coaching, conversation analysis, AI role-play, seller guidance, and analytics.

Many now can.

The real challenge is how much of that capability an organisation can realistically integrate into the way sellers and managers work.

It is a different constraint altogether. 

Buying functionality is comparatively easy. Changing routines, ownership, management behaviour, governance, workflows, and expectations around how the system should be used is much slower.

So as enablement platforms continue to expand, I think the gap to watch may be less about what the technology is capable of and more about how much of that capability organisations can realistically put into use.

What I’m paying attention to

The productivity conversation is moving closer to commercial output

For the last couple of years, a lot of AI measurement has understandably focused on adoption and time savings.

Those are useful leading indicators, but they are not the end point.

BCG’s data makes the gap visible: organisations are already seeing administrative gains, while measurable changes in conversion and win rates remain much less common.

The next phase will require a clearer connection between the capacity AI creates, what sellers do with that capacity, and whether commercial output changes as a result.

Readiness may need to include adaptability more explicitly

The Corporate Visions research raises an interesting challenge for certification and readiness programmes.

It is relatively easy to assess whether someone can execute a defined conversation correctly.

A harder test is whether they recognise when the situation no longer fits the model they practised.

That may mean designing assessments with more variation: incomplete information, changing buyer priorities, unexpected objections, conflicting stakeholder positions, and scenarios in which mechanically following the playbook would actually be the wrong decision.

Human interaction may become more valuable precisely because automation is increasing

The State of B2B GTM findings are only one dataset, but the strength of networks, introductions, and intimate events is difficult to ignore.

As more of GTM becomes automated, buyers may be exposed to an even greater volume of competent-looking outreach, content, research, and messaging.

In that environment, trusted relationships and genuinely useful seller interactions could become more differentiating.

This is a very different implication from saying AI will make sellers irrelevant.

Worth reading

The 2027 Revenue Agenda – Boston Consulting Group – A useful view of how revenue leaders are thinking about 2027, particularly the tension between higher seller expectations, widespread AI adoption, and relatively limited evidence of commercial impact so far.

Practice Makes Permanent: AI and new-seller development – Corporate Visions – Worth reading for the distinction between execution and adaptability, and the implications for how AI role-play should sit alongside other forms of learning.

The 2026 State of B2B GTM – A broad dataset across 300 software and AI companies covering GTM motions, channels, AI adoption, and where teams plan to invest next.

The biggest surprise from the new Revenue Enablement Platforms Wave isn’t AI – Forrester – A useful reminder that growing technology capability does not automatically translate into organisational adoption or operational use.

The bigger picture

Across these developments, revenue organisations are adopting more AI, expecting more from sellers, and giving teams access to technology that can do far more than it could even a year ago.

The important part now is how that capability gets translated into day-to-day work.

Revenue leaders are planning for higher productivity per seller. AI can remove administrative work and create more capacity. Training technology can make practice much easier to scale. GTM teams can automate more of the work involved in identifying and engaging potential buyers.

None of this guarantees better selling.

The real test is what organisations do with the extra capacity, how they help sellers develop judgement alongside fluency, how they create interactions that are genuinely useful to increasingly informed buyers, and how they embed new technology into the routines of sellers and managers.

AI adoption is already widespread. The next stage will be judged much more on whether it helps revenue teams use seller capacity better, improve the quality of customer interactions, and turn that into stronger commercial performance.