AI has moved past the novelty or hype phase in enterprise sales.

Enterprise sales leaders no longer question whether reps can use AI to draft an email, summarize a call, or prepare for a meeting faster. The question has now become how AI will improve how sellers allocate their time, how it'll help managers actively coach performance to business priorities, how it can help RevOps protect forecast quality, and help sales compensation leaders reinforce the behaviors that impact revenue.

This was the bigger conversation inside our Nudge 2026 session moderated by Mollie Bodensteiner, VP of RevOps at ZoomInfo. The session brought together a rare mix of academic research and operator experience. In the session, we heard from:

• Lauren Silvers, Director, Sales Transformation at Okta
• Johannes Habel, Associate Professor at the University of Houston
• Julia Fu, AI Transformation & Adoption Leader, WW Enterprise Sales at Microsoft
• Nathaniel Hartmann, Associate Professor of Marketing & Innovation at the University of South Florida
  

Together, our panelists explored why some sellers embrace AI while others resist it, where AI is meaningfully improving performance today, and what leaders need to understand as AI shifts from experiment to expectation.

The central takeaway we heard on repeat? AI adoption is not really a tooling problem. You need to approach adoption as a trust, data, workflow, and incentive design problem.

Which in turn makes it a sales performance management challenge.

Below we've rounded up our biggest takeaways from the session that should be helpful even if you can't watch the replay right away.

AI trust starts with data, but adoption starts with usefulness

One of the clearest themes from the panel's discussion was that seller distrust of AI is often rooted in something more basic. That is — distrust of the underlying data.

Professors Nathaniel Hartmann and Johannes Habel noted that this innate trust from your sellers depends on multiple layers, from confidence in the data itself to confidence that AI outputs will actually improve decision-making. Lauren Silvers mentioned that if CRM data is unreliable or AI recommendations feel like a black box, sellers are unlikely to engage at all.

So how do you get the trust required to have teams adopt and use AI meaningfully in the sales process? The fastest path to adoption is making AI visibly useful in the flow of work.

Lauren pointed to administrative burden as an obvious starting point for this. When AI helps sellers spend less time on manual updates and more time selling, trust builds quickly because the demonstrated value is immediate.

For RevOps and sales compensation leaders, the lesson's simple: sellers are far more likely to contribute better data when they can see how it helps them prioritize accounts, prepare for meetings, surface risk earlier, and ultimately improve performance. Based on the responses from our GTM leaders on the panel, they insist that trust grows when AI creates better outcomes, not when it creates more compliance.

AI can't be measured as one universal productivity lift

Another insight from our panelists centred around how many AI business cases still assume a very simple model. I.e.: introduce tool, drive adoption, improve productivity. However, Johannes Habel challenged this logic directly.

Before we assume AI will improve every seller’s performance in the same way, sales leaders need to understand how different use cases affect different seller segments. For example:

• A tool that helps mid-performers may do little for top performer
• A training simulator may lift lower performers faster.
• A sales enablement tool may create the most value for experienced, high-performing reps who know how to translate outputs into action.

Here's Johannes on the research backing this:

This research has important implications for revenue leaders evaluating AI-driven sales optimization. Which is that the AI adoption question can't stop at “Did the team use it?” Rather, senior leaders need to ask: who benefited from the AI in application, why, and under what conditions?

For example, conversation analytics may help mid-performing reps because they have enough selling context to act on feedback, while still having room to improve. AI role-play may create more lift for lower performers because it gives them a safer, repeatable way to practice. Predictive churn insights may help experienced sellers most because the output still requires judgment (i.e. do you invest more time in the account, escalate internally, change the renewal motion, or walk away?)

This ultimately changes how RevOps and sales compensation leaders should think about measurement.

Averages are not super reliable for measuring adoption uptake. If overall productivity rises by a few points, but the lift is concentrated in one performance band, tenure group, segment, or role, then rollout strategy should change. So should enablement. So should coaching.

In some cases, compensation design may need to account for whether AI is changing performance potential unevenly across the field like this.

Based on what Johannes shares here, a more sophisticated AI performance review ideally breaks impact down by:

• Seller performance band: Are low, mid, and high performers benefiting differently?
• Tenure: Are experienced reps better able to turn AI outputs into selling action?
• Segment or motion: Does the tool work better in enterprise, mid-market, renewal, expansion, or outbound motions?
• Use case: Is AI improving preparation, prioritization, coaching, forecasting, CRM hygiene, or customer engagement?
• Decision quality: Are sellers acting on better evidence, or just completing more tasks faster?

This is where AI becomes strategically interesting for SPM. That is, not AI adoption for its own sake but rather understanding where AI meainingfully changes the productivity curve, where it amplifies strong judgment, and creates new coaching needs. This is a much more useful conversation than “how many reps logged in and tried the new tool?”

The highest-value AI use cases reduce drag and improve real-time decisions

Julia Fu shared how Microsoft is thinking about AI in enterprise sales, particularly where sellers spend the bulk of their time today based on reasearch.

Highly relevant to every large GTM organization, Julie shared sellers spend too much time on work that is not customer-facing or revenue-generating (think CRM integration, reporting, meeting coordination, and internal team orchestration). All of this consumes capacity that could otherwise be spent with customers.

But Julia also pushed beyond mere admin relief. In enterprise sales, Julie sees that AI can help sellers synthesize account context on the go, and adapt between stakeholder conversations in real time.

As was discussed in the session here, the first wave of AI productivity often gets framed as efficiency (fewer manual updates, faster research, less time lost to internal coordination). But the bigger value is decision speed. A seller who can walk out of one customer meeting, summarize what changed, generate a more relevant point of view for the next stakeholder, and adjust the account strategy before the day ends is operating differently. AI is not just saving time. It is shortening the distance between signal and action.

This has direct implications for forecasting and sales performance management.

Forecasting often suffers because reality changes faster than the operating cadence captures it (for example, a champion loses influence or a buying committee expands). If those types of signals remain trapped in call notes, Slack threads, emails, or seller memory, the forecast becomes a lagging artifact.

AI can help close that gap, but only when the organization has the right data flows and decision processes around it.

For RevOps leaders, this means the AI agenda should be tied to specific operating improvements:

• Can AI reduce the time it takes for customer signals to appear in the CRM?
• Can it improve the quality of stage progression by tying movement to customer evidence?
• Can it surface deal risk earlier, with enough context for managers to intervene?
• Can it help sellers personalize outreach or meeting preparation without removing human judgment?
• Can it make forecast reviews more evidence-based and less dependent on rep narrative?

For sales compensation leaders, it opens an equally important question: when AI improves visibility into seller activity and deal quality, how should incentive design evolve?

AI-driven SPM should not simply automate payouts faster. It should help leaders understand whether incentives are driving the right behaviors in a more complex sales environment.

This could mean better visibility into leading indicators. Better signal around multi-stakeholder engagement. Faster feedback loops between plan design and field behavior. More informed decisions about when to use activity-based incentives, how to evaluate quality of pipeline, or where coaching should complement compensation.

The best AI use cases aren't just about making sellers faster, but making the revenue system more responsive.

Don't make the seller the integration layer

One of Lauren Silvers’ strongest points came when the discussion turned to workflow adoption.

She warned that many organizations are now throwing AI tools at sellers without intentionally designing how those tools should work together. In this type of environment, the seller becomes the integration layer. They are being asked to figure out which tool to use first, what to trust, how to connect outputs, and how to fit it all into the selling motion. And it's also where adoption starts to break according to Lauren:

A seller’s day is already fragmented. Adding more AI tools without workflow design creates cognitive overhead disguised as innovation. The result is inconsistent usage and leadership frustration that the field “isn’t embracing AI.”

Lauren’s recommendation was to test, pilot, iterate, and validate AI workflows with the teams expected to use them. Johannes added that this is especially important for autonomous agents, where much more can go wrong than leaders may expect. Getting an agent to the point where it can operate with limited supervision can take weeks or months, not an afternoon.

This is where AI adoption needs to be treated more like product management than software deployment. Before scaling a tool or agentic workflow, leaders should be able to answer:

• What seller behavior are we trying to change?
• What existing workflow does this improve or replace?
• What data does it need to work?
• Where does human judgment remain essential?
• What failure modes are acceptable, and which are not?
• How will we measure lift beyond usage?
• What will we stop asking sellers to do once this workflow is live?

That last question is often missed.

If AI is added on top of existing process without removing friction, it becomes another obligation. For sellers, the value proposition has to be obvious. For leaders, the operating model has to be deliberate.

This also matters for sales compensation because incentive plans already shape seller attention. If AI introduces new recommended actions, prompts, scoring models, or activity signals, leaders need to understand whether those signals complement the plan or compete with it.

A rep cannot be told that strategic account development matters, paid primarily on short-term bookings for example, and then nudged by AI toward dozens of micro-actions with unclear value. That is how systems create confusion.

The future of AI-enabled SPM requires alignment between workflow, measurement, and incentives. Otherwise, AI just becomes another source of mixed messages.

AI stands to raise the value of human judgment, especially in edge cases

A recurring theme across the panel was that AI doesn't eliminate seller judgment. Rather it changes where judgment matters most.

Nathaniel Hartmann noted that AI performs best when there's sufficient training data. In novel situations, complex stakeholder dynamics, or customer-specific scenarios, humans still need to lead. Lauren Silvers and Julia Fu echoed this, emphasizing that enterprise selling still requires empathy, relationship-building, and the ability to navigate ambiguity (something only reps can do with their context).

Johannes Habel suggested that as AI automates routine work, sellers will spend more time on edge cases where judgment creates value. This will mean organizations may need to rethink what they train, coach, and reward.

Future sellers will need to excel at interpreting AI insights, managing complex buying groups, and knowing when to trust (or challenge) AI recommendations. For RevOps and sales compensation leaders, as more tasks become automated, performance systems should place greater emphasis on the outcomes and behaviors that still require uniquely human decision-making.

AI may automate more work, but it will not automate accountability for the quality of seller decisions.

Building trust in AI for adoption is the focus for RevOps and sales compensation leaders

Ultimately, AI is forcing revenue teams to confront an uncomfortable truth. Many organizations are trying to automate systems that were already difficult for sellers to trust. From bad data, to fragmented workflows, to misaligned incentives, sometimes all of this can feel more useful to leadership than the field.

And AI doesn't make these problems disappear. It only amplifies them.

As our panelists shared — used well, AI can reduce seller drag, improve data fidelity, accelerate decision-making, strengthen forecasting, and give leaders a more dynamic view of what actually drives performance. It can make incentive compensation more responsive to the behaviors and signals that matter. It can help RevOps move from reporting on the business to actively improving how the revenue engine operates. The catch is that adoption has to be intentionally designed around trust.

Sellers need to trust the data. Leaders need to trust the outputs. Customers need to trust the interaction. The business needs governance around what AI can access and what it can act on.

This is the real work ahead. Not simply adding AI to the sales motion. But building a sales motion that AI can actually improve.

Want to see the full discussion? Watch the full Nudge 2026 session, “AI in the sales motion: The psychology of adoption and trust,” to hear how leaders are thinking about AI adoption, seller performance, and the future of revenue execution. One form fill gets you access to all five sessions on-demand.