Salesforce’s Kris Billmaier on AI Sales Agents

Hello, and welcome to the latest edition of the Techstrong AI Leadership Insight Series. We're here with Chris Billmeier, who's Executive Vice President and General Manager for Salesforce Agentforce, and we're having a little chat here about, well, what is going on with AI agents and Salesforce workflows because, well, folks are adopting these things like crazy, but success has been somewhat mixed, so we'll get some insights from somebody who knows about it firsthand. Chris, welcome to the show.

Hey, thanks for having me, Mike. Excited to be here. All right.

You guys have been at this for a little while now, and some organizations are much further ahead than others. So from your perspective, what do you wish most organizations knew about this before they got started, and what are the lessons learned so far? It's a big question.

It's a great one, though. I think I always start out with great AI has to come from great data, and great data comes from a lot of places. So I think the first thing that every organization needs to do is make sure that they have a strong data foundation in place.

And what I mean by a strong data foundation is it's not only your CRM data, but it's all the calls you capture, all the emails you capture. That's where the gold is. So having that data foundation to create the context is incredibly important.

Without that, I think you're kind of flying blind. " What I always suggest is, what's your main use case that you want to affect? Do you want to affect your prospecting?

Do you want to work existing leads? Do you want to create account? What do you want to do first?

Prove it, go deep, have goals for it, before you boil the ocean. So those are the two things that I see successful companies doing when they're getting started with AI. A lot of the entries into various CRMs are non-consistent because you see a lot of people using small things.

Companies or misspelled names or titles are not quite the same. How much effort should I put into cleaning all that up before I go launch an agentic AI project? I think it actually depends on what you're trying to do, but I think the point you make broadly is important.

One thing I see that happens in a lot of companies is they start recording their calls. Those calls get auto-updated to CRM, and take the bias and take the inefficiencies out of it by trying to use a tool that can self-update based on a call you have, an email you've written, et cetera. So I think the big things you need to fix like you've talked about, but I think you could create that context going forward really effectively if you have the right data capture tools.

How is the average experience of a salesperson changing as we employ AI agents? There's a lot of things that they do all day long that have very little to do with selling per se, but what are you seeing in terms of how people are using this, and what experience are they having? Yeah, you kind of hit the nail on the head with the first statement you made is sellers are spending more than half their time not selling.

And we run a state of sales report every year, and we found that on average, 60% of a seller's time is spent on non-selling tasks. So when you reduce productivity, you also reduce growth. So when I think about areas of the sales funnel that can really be affected by AI and give back productivity: top of the funnel, prospecting agents, engagement agents to do those initial outreaches to customers, go back and forth with email, book a meeting on your behalf, create account briefs on your behalf based on all that contextual data that they have.

That is a huge time save that we can affect sellers. Why should you ever wake up to an empty lead list when agents can do it for you? And then your job is really get smart on the data that the agents have brought back and have really impactful, effective, either in-person or meetings like this.

So I think that's the change we're seeing right now. Sellers who are growing are doing that. They're affecting the top of the funnel significantly.

And to your point about that, most customers, if they have a relationship with somebody from a sales side, they want it to be more than transactional, and I feel like a lot of the relationships today are still transactional, and what they really want is a sales rep who kind of acts more like a strategic advisor, right? We said something about you're a revenue architect now that you're a seller because we're taking away so many of the busy work things you have to do. You can affect the strategy of your customer, of your team, because you just have more time to spend doing it.

And I totally agree that the relationships are what really matters for the seller and acting as that strategic advisor, that guide who can come in and say, "Now this is how you should be thinking about things," and knowing the business and knowing your customer, I think that's where the gold really is. How will this evolve, though? Because on the other side, the customer who probably has something that looks like a purchasing department somewhere is going to have their own AI agents, and so will the sales rep's AI agents negotiate with the purchasing department's AI agents, and how might that all play out?

That's a version of where we could be headed. But I think at the end of the day, people still want to interact with people, and to the point you made, you want that strategic relationship with the person you trust who's selling you something that's core to your business or core to your life. So yes, I think there will be agent-to-agent conversations and Having that as a deflection point or an adoption point is interesting.

But I think for sales, both buying and selling, a human is going to be in the loop for quite a while. Mm-hmm. Do you think we'll have better visibility into the funnel and also maybe what the revenue looks like for the next quarter and the quarter after?

Because the main reason a lot of organizations invested in CRM in the first place was to gain that visibility and be more predictable. And so is that going to be something that is the actual ROI of all of this? It's kind of like you're reading my pitch deck that I give to customers a little bit.

I absolutely think it is. I look at the way that a lot of companies forecast today, and it's still based on gut and intuition. And while that all has a place, why shouldn't I be understanding the actual conversation that's going on with the customer to really infer what the likelihood of success with a potential deal is without bias?

So yeah, I absolutely think you're right in that. I think we're getting to a more... What's the best way to put it?

A higher fidelity outcome of observability along the sales cycle. And along that, I think you have to take into account goals and outcomes. I think we're very quickly getting to a place where I want my sales agent not to think about one tiny task.

I want them to think about my quota and the duration of time I have left to burn that down if I'm a rep. I want them, if I'm a CRO, to think about my quarter and how I'm going to make the number and where I need to jump in. So yeah, I think we're going to see that observability and that goal orientation really start to emerge.

How smart can these AI agents get? And I ask the question because, theoretically, if some company buys land somewhere and they're going to build something that looks like a warehouse, chances are they're expanding their operations, and I should be able to correlate that into something that feels like a place that I should go spend more attention. So are the AI agents going to get that smart where they can look through and see there are leading indicators of something that says there's an activity here you should check in on?

I think we're already there to a certain extent. Think about hiring signals. If company XYZ hired 10 new SDR, what are they looking for?

Are they looking for AI to augment those SDRs? Are they looking to grow their business? So I think the signals are already there.

But the point I started with, what's the data foundation that can obtain those signals to make the AI smart enough to say, "Hey, you know what, Chris? " How will the relationship between the sales team and the finance team change in the age of agentic AI? Because I think a lot of times there's tension in that relationship, and a lot of it comes down to just looking at different data from different silos at different times.

So will that get a little bit smoother? Because a lot of times I've been in this meeting where the CFO says the sales outlook looks like X, but then it turns out they didn't have the latest data, so then the sales data outlook is Y. " But I always felt that conversation was disjointed, shall we say.

It's like the Spider-Man meme of everybody pointing at everybody else a little. Right. Yeah.

The interesting thing is the context of sales and the data that you have across the company, I think, is going to shift, and I think there's going to be a more consistent and common view if we do the thing that is capturing the data correctly all up. I actually think what's going to happen is that it's not just sales that's evolving, it's go-to-market teams that are evolving. And that kind of cohesion between sales and marketing and sales and service and sales and finance, it kind of makes everybody accountable to revenue.

And if you're not operating with a consistent set of data, I think that's where the problem will arise. That's where the problem arises today, as you pointed out. But I think we're going to get to a point where we're going to need that common set of data to have the context of the company for agents to work effectively.

And so I think it will change. I think it'll get better. All right.

So now you're reading ahead to my next question, which is to your point about marketing. A lot of times, or historically, salespeople were waiting for marketing to do something and hoping that there would be a bunch of leads generated. But as I look at this agentic AI capability, will each sales rep be able to run their own marketing campaign a little more targeted to specifically the customers that they know?

Because there may be a repository of content somewhere, but they know how to tailor it specifically to a particular client. I think that could happen. I think it's going to be very organizationally dependent.

I think marketing will still be in charge of certain things like inbound on the web, and sales will still be in charge of things like outbound and hunting. But I could see a scenario where you're a seller managing 10 agents that have sales capability but also have marketing and outbound capability. I could for sure see that happening.

We're kind of seeing that in product development right now, right? Product development has become about pods of individuals from design, from product management, and from engineering all working together and all being builders. I think you could have the same kind of scenario occurring across sales and marketing where you just kind of merge and become Go-to-market drivers.

Salesforce, of course, you've been using this stuff internally as well. What do you know now that you kind of wish you knew a year ago before you started running this stuff internally? One, always have a goal for what it's going to do.

Measure it and continue to train it. But yeah, I'll give you an example. Our initial SDR agent that we launched, we needed to continually train it with more and more data because this thing was like a junior employee.

And when you have a junior employee, you train them, you bring them up to speed, you give them more context, and over time, they improve. That's what we've actually learned with our deployments of agents is you got to start narrow with a goal, and then you've got to constantly look at the outcomes and constantly refine what that agent is doing on your behalf. So that's one major learning.

The second learning I'd say is we're already seeing sellers manage people and manage agents to drive crazy capacity. We get more leads inbound than we could ever work. And so now the woman who runs our SDR team, she manages 100 human SDRs who handle the top quality leads as scored by us coming in, and then she manages the equivalent of 400 agents that are handling the other 75% of leads that come in that we just never had the human capacity to touch.

And it's driving growth, which is really awesome. So to your point, every organization has it in their head how many sales reps they need to drive X amount of revenue. Is that math changing?

Will organizations need to kind of revisit just how many sales reps they need to drive a certain revenue goal per quarter? I think what's going to happen is companies are going to look at how they can grow exponentially, and that's going to mean adding headcount in sales. It's also going to mean adding agent headcount.

The way I think about things is I want a sales agent that sits with me, that understands my goal, and helps me become a 10X seller. And helps me get to that point where I'm not working on one deal a week, I'm working on 10, and I don't have to think about updating my CRM because the agent's doing all the blocking and tackling for me. So I think we're going to continue to increase headcount in sales, and I think we're going to continue to increase agentic headcount in sales.

Is there a line between what an AI agent for an individual sales rep is doing, because it's their personal sales helper, and then that which they get from a company who's using your platform, per se, and will those agents need to find a way to kind of collaborate on some level, or should we just have one set agents and maybe each individual sales rep shouldn't have their own personal agents? Is that reasonable? I think, again, I think it's going to be company specific.

" I think if I'm giving my advice to any company out there, I think the best thing to do is have a platform with the agents that you need to do certain jobs, give those agents to your sellers, let the sellers tweak and customize to their voice, so that you have that one centralized repository so you understand the growth of your business and you understand the truth of your business. Because that consistency and that growth is what we want at the end of the day. Out of curiosity, do you think that there's more opportunity to be gained by just focusing on the opportunities that we missed from our existing customers per se?

Because I feel like a lot of times sellers just miss out on things that were sitting quite literally at their feet versus if I go talk to business leaders, they're always like, "Well, we're going to gain market share and address our TAM," and et cetera. And while that may be a noble goal, maybe it's just all about let's get more money out of the clients we have. I think both can be true, but I think the number one signal that's missed is you're on a call with somebody you're trying to sell to and they say, "I'm really interested.

" That is always dropped, and if you're able to capture that and feed it back to enrich existing customers, that is a huge win. We built something called the Prospecting Agent. It does exactly that.

" But you're right. I think it's a prioritization thing, and you could absolutely find gold in existing contacts and opportunities. I think the harder part at this point may not be the technology itself as much as it is just the culture and change management that goes into this.

So based on what you've seen so far, what are people who are doing this well getting right? They're understanding what needs to be assistive and what can be autonomous, and they're kind of building trust in those systems. So when I say that, I think about observability, I think about what is an AI going to help me do and push me to do versus do for me, and I think when you get it right multiple times on that assistive motion, you then moved into the autonomous world, and these two things sit on the same spectrum, and it's just about how much trust I can build with the initial use cases that I've executed on.

I think that is what a lot of companies get right. And then I'll go back to my data point. A lot of companies who are doing this right, they're recording their calls, they're capturing their emails.

They're allowing systems to do deep understanding and almost memory, that they're then using as insights to drive their businesses. All right, folks. Well, you heard it here.

There's no easy button when it comes to AI, but if you put in the time and effort and you think about the culture and the technology, well, the benefits are going to far outweigh any of the effort, that's for sure. Hey, Chris, thanks for being on the show. Thanks for having me.

This was a lot of fun. All right. AI Leadership Insight series.

You can find this episode and others on our website. We invite you to check that out. Until then, we'll see you next time.