Our next session is with Amit Zavery and Sarah Franklin. Amit, Sarah and I have known each other for at least a decade. So consider this as a less formal introduction and more me finally getting a mic in front of an audience that they can't interrupt. So fair warning, this is going to be a fun and not just a panel of AI vocabulary. On the technical side, Amit leads product and engineering at ServiceNow. And the reason he can talk credibly about running thousands of agents without losing the plot is that he spent decades building the governed systems that make workflow trustworthy at scale. First at Oracle and then at Google and now with AI Control Tower. That's not a side project for him. ServiceNow's AI ACV crossed a billion dollars in Q2 with agent customers in production up ninefold in nine months. So he's living the problem in real time, not theorizing about it. Please welcome someone who will make the case that control, not autonomy, is what everyone in this room is actually buying. Amit Zavery.
And my thank you, sir. My dear friend Sarah that I worked very closely with when she was the president at Salesforce. Sarah is CEO of Lattice where she's leading the company's mission to make work meaningful at precisely the moment AI is reshaping what work looks like. Before Lattice, she spent more than 15 years at Salesforce, including as president and chief marketing officer, and helped build platforms and communities that brought technology closer to the people using it. So when Sarah talks about AI governance, she brings the perspective we need, not only how to deploy powerful technology but how to do it with trust, accountability, and people at the center. Please welcome Sarah Franklin.
Thank you. Good to see you.
I'm doing well, thank you. How are you?
I'm pretty well. How is everyone doing?
Middle of the day. Good day.
It's good energy. So, thank you all for being here.
I have to follow you, just so you know.
I know. We have to have better stories, right? You know, with the proliferation of agents and it's a good teed-up for the workforce, you know, being more than just people, it's also people and agents. And what are you seeing at ServiceNow in terms of the agentic workforce?
Yeah. No, I think this idea of agentic workforce and AI agents has become quite prevalent over the last few years for sure. Every customer I speak to wants to figure out how they will do automation, how they'll improve the efficiency of their business and business processes using AI. So it's a very natural thing for sure. The question comes up with many of these customers is like, how do I know I'm doing the right thing? What are the agents up to? What is, what, how many agents am I running? Who owns them? What, what kind of systems they have access to? So the whole, this issue around security, governance, trust has become very, very high on my, in their mind, right? Initially everybody was very excited to building workflows with AI agents and suddenly, you know, what, I have 30 agents, then the worry started very quickly about what is happening inside the business. So we have been very clear since last year we delivered this idea of something we call AI Control Tower. The idea behind that is very simple, right? How do you discover everything you're running with AI inside your company? And ServiceNow's heritage has been about tracking all the enterprise assets, hardware, software assets. So it's very natural to add AI agents, which is another software in a way, into that particular architecture. And that's what I think has really helped customers understand that they have to really get full control around all their investment and figure out what's happening with the systems and the life cycle around it, the cost around it, the governance around it, so that nothing wrong can go on even though there's tons of activity. Without having that kind of control, you can't really get there. So that's usually the top question. So ServiceNow has always been about like, how do we really help you manage your enterprise assets? And AI agents is one of the enterprise assets and working with various different third-party systems. How do you find that information, put it into the control tower and really help you track everything in a much more integrated manner. So very difficult times for customers, but I think it's important to really continue to help them solve that problem.
And what are you seeing in terms of the governance and the way that your customers are adopting all of this technology and getting value from it?
Yeah, I think the governance, I know everybody talks about governance for every type of software. It was never that exciting and it's not exciting, right? It's always the last thought you put in. Oh my god, I have to do some governance and I had to maybe create some reports, show it to my audit committee and make sure things are not going wrong. So it's always an afterthought. So I think with AI it has become the first thought more and more. Maybe it was not that last year, but this year for sure after all the recent attacks with the AI agents into various websites and companies as well as the volume of things they're adopting from so many third-party systems is becoming very difficult. So governance is becoming top of mind and I think what the work you guys are doing in terms of creating AI workforce, it's not just easy to go and use it if you don't know what and all the information around it. How do you bring that into a governance layer? And the risk management and trust, trustworthiness is very important as well.
Yeah. And at Lattice, what we're seeing and what we've done, for those of you that aren't aware, Lattice is like an HR tech platform, manages people and their performance. And about two and a half years ago, we went to innovation, we had an innovation that was shared around the concept of digital workers. And what's very interesting is that two and a half years ago, this is timestamped on LinkedIn, you can go see, we shared that we were going to create a record for an agent as a digital worker in the workforce. So you could have a human manager for that work and you would be very clear for the work that it was assigned. Not from the standpoint of is it secure or is it governed and who's the IT person taking care of it, but actually if you're going to take headcount dollars from a sales representative, an SDR, and assign an agent to do the work, is that SDR going to create good leads? Is it going to create good quality pipeline or is it just going to create, you know, a lot of bad pipeline? And just like every human has a performance review, how do we make sure that the agents are also held accountable, especially when we're using headcount dollars to fund them. And the reaction was not expected, because the reaction was clown emoji. What is going on?
Yeah. It was, it was like, no, this is crazy. What crazy talk? What are you saying? And this was, you know, 2024 we said the digital workforce would be about 30 billion dollars in spend. And we are 2026 and it's presently about like 15 to 20 billion. So well in that, that, that head, that pace.
And expected by 2030 to reach like half a trillion or plus dollars.
