Fireside Chat: Harnessing Agentic AI for Smarter Sales Performance Management

Today we have the great pleasure of meeting with Chris Ferguson. He is senior product manager with Xactly. Welcome, Chris.

Thank you for having me. I'm very, very excited to talk with you today.

Are you in San Jose?

I'm not. I'm actually based in Denver. So I've been in Denver for the past seven years now. It's been a good place so far.

Well, what fascinates me about exactly who is the leader? Compensation management and incentive compensation that you have a long history of leveraging the proprietary data for insights that you pass on to sales leaders. So how do your new AI solutions, in particular the agentic AI, combine your decades of compensation data with the client's unique set of data to create better and more actionable insights?

The key for us is because we have, you know, 20 plus years of this anonymized and aggregated data, it's a great resource for our various AI experiences to really pull from and it allows us to really provide sort of key benchmarks throughout the intelligence revenue life cycle when it comes to, you know, different planning processes. You know, if you're looking for different scenarios structuring those compensation plans, how do you do so in an agentic fashion where you know that you're going to be lining up with, you know, the right industry benchmarks and how do you do that in a way where, you know, it's very easy to interact with the data itself. And so by allowing our agents to interact with this core trove of data, we can not only, you know, help provide the next best task to go and reshape a compensation structure, but then we can also provide those key industry benchmarks into what it is they might need to think about so they can instruct the agent to go and make that necessary change.

How does exactly ensure the highest level of data security so there are no breaches, so there's transparency and there's trust on the part of the sales team and also the executive leadership.

Because we're built on a single unified platform, it means you're managing the security and the policy of all of the data and interaction into that data from a single source of truth. Because we are building exactly intelligence and our larger sort of AI experiences are on top of the platform. We're able to provide those same policy sets and same security layer into all of those interactions. What that results in is regardless of the functional area that you're working within, whether you're designing your compensation structures or administering your compensation plans, it's all fundamentally off of a single platform that ultimately allows you to secure those interactions when say a sales rep is interacting with their AI versus a sales leader. And then I think the other important thing is, you know, we want to demystify the fact that AI is a black box and how do we provide insight into the steps that an agent and an assistant are taking on behalf of a user to achieve a task. And so very intentionally we want to provide sort of that, that insight into what's the business logic that's taking place behind the hood and where is the agent going to, you know, achieve that interaction within the platform.

So do exactly the AI powered solutions specifically help a company, sales leaders and of course the finance teams that align on a single source of truth because they need to reduce the friction and speed up the strategic decision making process.

We want to be able to really provide sort of a natural language experience that really minimizes sort of the clicks that they have to go to in a given system to get to that data. We also want to, you know, provide that data and information into that data in an intuitive manner, whether that's graphically or tabularly. And so our experiences are really designed to be flexible in providing that single source of truth because it's built on one platform, but then also providing sort of the multimodality to ingest some of the data that is returned from our generative AI agentic AI experiences.

So it brings up another question. How do companies sort of connect all those little data puddles into a data pool or a data lake so that you don't miss anything?

You know, when you have so many disparate systems that are providing, you know, data into your, say exactly ecosystem or name your compensation system, it's hard to know, you know, what pipeline is bringing what data in and where is, where is the breakdown, if any. And so, you know, because we have the ability to have these, you know, pipelines coming into a single source of truth, we're able to understand very quickly sort of where, where is the data coming in from? Where is the, you know, the data at rest? What are the transformations that are taking place? And those are other questions that, you know, you can use and prompt our AI with as well.

What is the journey like for a company that adopts your AI driven sales performance platforms?

Where we've seen success with our customers is, you know, really getting in there and identifying what are the, you know, different high impact workflows that are really going to be well suited for, you know, different AI experiences. And I think because of the flexibility that we offer at Xactly you know, there is a huge opportunity to really understand, you know, a core business use case. Say it's a, you know, dispute management use case that is really taking up a lot of time for op admins. How do we, you know, provide an agent that's going to really, you know, handle a lot of that traffic that an admin might get from the field? But how do we do it in a way where we're providing the correct language that is used for that company because of the composability that we offer? You know, you can actually define the key instructions and kind of key data that the assistant or agent is going to leverage and then ultimately help reduce that workload for, you know, your admin team as an example. And so I think kind of coming back to it, it's really sitting with the customers and identifying those key workflows that we can then recommend. Hey, we believe this specific AI experience is really going to impact your day to day here.

So what we are hearing here at Selling Power is that sales leaders are very much concerned with forecasting accuracy and they want to create a predictive sales organization. And every good organization is a good learning organization where they have a feedback loop for the intelligent revenue cycle. So could you walk me through how exactly products leverage AI to create and refine that feedback load so that organizations learn faster and better?

That's a really, it's a really good point because I think always seems thus far is that a lot of people have been focused on the generative and agentic AI because it's sort of a sexy new thing. But we're kind of forgetting the larger traditional machine learning and AI components that are so important to this more holistic AI journey. And so, you know, here at exactly if we take more traditional, you know, machine learning workflows like our win rate models, that can help provide some of the explainability that's going to provide better predictability into your pipeline and allow you to use that data to better inform your decisions, whether it's in the next round of planning or even in flight within the given year there. And so I think the really exciting opportunity that we have as technologists is marrying the traditional machine learning with some of the generative and agentic experiences so that it's easier for a comp admin to get insights into the outputs of those machine learning models rather than having to solely rely on a core data science team.

Krista, I'd be interested in hearing your vision when you look at what the buzzwords were last year. Like everybody talk About ChatGPT in 2025, it's all about agentic AI. What emerging trends do you see for 2026 and beyond that you're excited about that will impact the way we create sales, create customers and create better run.

Companies for us and specifically in our space there's this concept of, you know, I think artificial general intelligence. What does that mean in sbm? And I think for us, you know, thinking about what does a truly autonomous SPM worker look like to support a sales team. But then also, you know, as we're thinking about practitioners and you know, our core Persona being the comp admins, how do they need to start thinking in terms of capacity planning and territory planning and you know, just, just larger sales planning around agent sellers. Right? Sales, sales, true sales agents that are kind of working on behalf of, of an organ. So I think we have the opportunity to provide systems that can support both of those experiences and I think other kind of buzzwords to think about. Going back to your initial question around, hey, how do you know or how do you connect these different sources of data? I think something we'll start to see is really agent to agent as a new protocol, meaning, you know, exactly as agents we want to, you know, work with ServiceNow's agents or we want to work with, you know, name your other kind of vendor there. And how does that become sort of a new protocol that, you know, we can really bring to our customers so that it's an easier experience for them to, you know, interact across their entire technology stack. So I think, I think those are, those are a couple. We could go much deeper into, you know, some of the different technologies to bring that to life. But you know, I think that's, that's something that we have top of mind for us going forward.

What is on your drawing board right now for integrating AI into your existing platform to enhance the value of the new products and features that you're developing?

We now have, you know, the market's first SPM agent and the possibilities and the patterns that that have unlocked is pretty, pretty endless, especially around the orchestration and delegation. We're really going to be focused on a lot of the composability aspect. How do we allow our customers and our partners to really put their footprint, put their stamp on some of these agents and assistants while still leveraging Xactly's larger platform. And then also how do we think about both the multimodality and additional tools that these agents are going to need to deliver on their tasks? And so another example there that we're currently working on is a lot of our larger customers have pretty mature data science teams. What that looks like is that they're running these miles kind of off system and then bringing the results into the larger SPM ecosystem to make sure that they can handle the necessary tasks. And so how do we provide a way for that same data science team to bring their models into the exactly ecosystem so they can not only transact with them traditionally, but then also, you know, marry those to an exactly agent.

Running in my prediction.

I would love to.

What about orchestration automation?

That's very much a reality. And you know, this notion of agentic orchestration, it's something that we've brought to life in our first SPM agent, our builder agent. And you know, it's this notion of, hey, you've got this supervisor agent that you are able to prompt into. There's a task group that's created and that supervisor, you know, it's got a hive of agents that it is working with. It's, you know, either it's team members consider them direct reports and it knows the talent of each of those direct reports. And so being able to sort of hand off the task to that hive of agents and you know, handle that sort of orchestration is something that we're doing, you know, in market today. But I think, you know, there's, there's much more, much more to come in terms of, you know, how do you think about different workflows that previously were very manual to automate, whether that's bringing data into the system or different sort of approval flows. I think, you know, agentic experiences give us the opportunity to really start to tie some of those pieces together much more easily than, you know, we've done in the past. Thank you so much, Chris, Kira, it was lovely chatting with you. Thank you so much for having us.