Key takeaways

• Agentic AI can help sales organizations reach long-tail customers who receive limited human attention today
• The role of the salesperson is likely to become more focused on connection, judgment and customer problem-solving
• Sales teams need mechanisms to turn individual AI experiments into scalable, governed capabilities
• Adoption is enhanced by finding “magic moments” tied to real work, not pushing technology for its own sake

AI has already changed how sales teams prepare, prioritize and engage. But agentic AI raises a serious question for commercial leaders: If AI can respond to signals, qualify opportunities and support customer interactions, what happens to the salesperson’s job?

That was the focus of a ZS-hosted webinar featuring Earl Fitzhugh of Microsoft and Arun Shastri of ZS. Their discussion explored where agentic AI can expand sales capacity, where human judgment still matters and how leaders can help teams adopt new tools without losing sight of the customer.

Below is a sample of their discussion. Watch the webinar to get the full insights.

Arun Shastri (AS): Earl, with the fast-paced acceleration of AI and agentic systems, a good place perhaps for us to start as practitioners in the sales ecosystem is, is there a need for salespeople, and do you see that need going away?

Earl Fitzhugh (EF): I don’t think sales reps are going anywhere, particularly in B2B. There is an opportunity to empower salespeople with a new set of tools and reskill them for this new environment. But the human element has never been more important when you’re navigating complexity and ambiguity, especially in enterprise sales. What will change is where technology plays a larger role. Agentic AI can enhance what we would call e-commerce today and make it possible to close more sales, especially in spaces that are hard to cover with traditional sales models.

AS: Okay, that’s a good segue into the agentic part of the conversation. Unlock possibilities how? What do you think happens in long tail?

EF: One clear area is the long tail: customers that sales organizations struggle to reach with full-time teams, such as vendors or channel partners who may be having a suboptimal experience.

Agentic AI can help create richer, more responsive experiences for those customers by scaling engagement capacity to match the unlimited surface area created by the internet and mobile devices. It can respond to signals, answer first-line questions, contextualize the offering and then fast-track the right customers into human-powered sales conversations.

AS: What happens when there are already humans present? What’s the role of agentic in that kind of an environment?

EF: By providing that context and answering those initial questions, agentic AI can bring a more qualified lead to the salesperson. That lets the seller focus on the last-mile conversation.

There’s also the idea of the salesperson in an “Iron Man suit” with the heads-up display. Sales reps can be equipped with more insights in real time. That helps them understand the customer better, prepare faster and connect more effectively.

AS: Does that mean the salesperson is doing a different job, or is it more of the same job but doing it more effectively?

EF: The essence of sales doesn’t change. A salesperson connects with someone who needs to address a challenge and helps them do that, ideally through the company’s product portfolio.

What can change is how much time salespeople spend on that core work. No one becomes a salesperson because they want to enter data into a CRM or prepare for internal reviews. Yet those activities can take a meaningful chunk of a seller’s week—sometimes around 25% on CRM-related work and another 15% to 20% on internal reviews. When AI reduces that work, the freed-up time can be invested in curiosity about the customer and helping customers unlock value.

AS: Now, in your experience for the type of person you might be hiring for, has the competency of these salespeople evolved over the course of the last three, four, five years? Do you find yourself looking for a different kind of salesperson?

EF: At Microsoft, we focus on “learn-it-alls”—people who are agile, willing to take on new challenges and ready to apply what they learn. In cloud and AI sales, some people have deep technical skills. But these tools are advancing so quickly that many nontechnical people can use them effectively.

So the bigger shift is less about requiring every salesperson to become technical and more about hiring and developing people who are hungry to learn, comfortable with digital tools and willing to experiment.

You want people closest to the problem to be able to build solutions. At the same time, leaders need a way for innovators to share what they learn so useful ideas can spread beyond one person’s workflow.

AS: Let me talk a little bit about first-line sales management. In the agentic world, has your philosophy toward first-line sales management evolved?

EF: First-line managers remain coaches first. They need to understand each person’s strengths and opportunities, then help them improve. They’re also change champions, both absorbing and translating organizational change on the front lines.

Where the role evolves is in how managers use tools to become much more effective. For example, managers can look at which parts of the sales management process should be automated or augmented. But they also have to decide where human involvement creates value.

Employee recognition is a good example. You could automate sending kudos for every win. But it may be more powerful for a manager to see all the wins, choose the few where a personal message will matter most and lean into those moments. Efficiency isn’t always the goal. Sometimes the human action is the value.

AS: Do you feel like all functions are operating as one? Do you feel that some of these concepts about intermittent and continuous come to life, or are we still a little bit away from that? What are your observations?

EF: Organizations still tend to operate in functional lanes, but agentic AI is starting to blur those boundaries. Sales, marketing, channel, product and engineering all contribute to business success, and these tools can make it easier for those functions to work from a shared view of the customer.

Agentic AI also raises practical questions about ownership. If an agent engages a customer earlier in the journey, is that a sales tool, a marketing tool or something in between? The important thing is not the label. It’s whether the organization can use the technology to improve the customer experience across the journey.

AS: Oh, so you mean sales doing more marketing-related things and then marketing perhaps doing more sales-related things, what might be traditionally in the sales domain?

EF: The classic tension is lead quality. Marketing may say it delivered 20,000 leads. Sales may accept 5,000. But the real question is what happened with the 1,000 that won—and what both teams can learn from that.

The opportunity is full-funnel visibility and accountability. Marketing should care not only about lead volume but also about which leads become customers. Agentic AI raises an interesting question: Are agents sales tools, or are they marketing 2.0? The answer may depend on where they sit in the customer journey.

AS: We talked about encouraging your salespeople to take chances, to build things and so on. Any lessons from your approach at Microsoft to how you drive adoption? What has worked, what has not worked? Any thoughts in that regard?

EF: First, you need to let innovators run. Then you need a platform for them to share what they’re learning with fast followers. Once 20% or 30% of the organization is adopting the tools and sharing success stories, the rest of the organization starts to follow suit.

Second, make it engaging and fun. Contests and similar activities can help build awareness and encourage adoption.

Third, and most important, help people find the “magic moment.” That’s the use case that solves a real problem for an individual and makes the value obvious. For me, one early example was the meeting summary capability in Teams. It meant I didn’t have to take all the notes. At first, the recap notes were about 75% there; by the third time, they were about 90% there, and I could focus on capturing only the highlights. I’ve had a similar experience with Excel, where I can ask it to do something in regular English instead of spending an hour trying to program it myself.

AS: Fantastic. This has been an amazing conversation. Any other things you’d like to share, things I didn’t touch on, any words of wisdom for our audience?

EF: Embrace the opportunity, but start with the customer. This is a moment to reimagine the customer experience and think about how these tools can help deliver something better.

Where things go off track is when technology is incorporated for its own sake. The better approach is to ask how we can use these capabilities to serve customers better. If leaders stay anchored on that, agentic AI becomes more than an efficiency tool. It becomes a way to improve the quality, speed and relevance of customer engagement.

Watch the webinar here to get the full insights from the conversation.