Agentic AI, Robotics, Leadership, Privacy and more part of the conversation at Constellation Connected Enterprise 2025

Out of all the industry events I go to throughout the year Constellation Research's Connected Enterprise (CCE) event is among my absolute favorites.  Each year Ray Wang and the whole Constellation crew create an intimate  atmosphere for a select group of industry execs, thought leaders and influencers to come together, learn from each other, and focus on the future.  And do all of that in a very conversational, organic way.

It's a formula that's been working for 15 years, with a great mix of first-time attendees and multi-year returning veterans. And the seesions don't drag on like they do at most other events because they're usually no longer than 25 minutes, so it's keeps the day moving - providing lots of in-between time for conversations outside of the main panels and presentations. Which allowed Paul Greenberg and I to grab a bunch of our favorite folks for a number of CRM Playaz mini episodes. 

Below is a compilation clip (along with an edited transcript included for your review) from CCE featuring, in order of appearance, the following industry leaders speaking on a variety of important topics:

Ravi Krishnamurthy - Vice President of Product Management, AI Platform and Reponsible AI at ServiceNow

Charlie Isaacs - CTO for Customer Connection at Salesforce

Samir Meharali - Head of US Solutions Engineering & Technical Support at Zoho

Ashley Hart - Global Chief Marketing Officer at Intellect

Vala Afshar - Chief Digital Evangelist, Salesforce

Esteban Kolsky - Chief Distiller, Board Advisor at Constellation Research

Tejas Gadhia - Product Management for Zoho Creator

The Power of incorporating AI into your broader system of record

Paul Greenberg: CRM was built on systems of record and then there was discussions of CDPs which are, you know, systems of record, but they're actually actions being taken in real time, constant back and forth real and run-time activity, et cetera, et cetera.

And  you guys are the first people I've ever heard say this - you are incorporating your AI into your broader system of record.

Ravi Krishnamurthy: You want to discover across the enterprise what is happening. And it could be running, for example, in one of the hyperscalers, Microsoft, Azure, Google. It could be running in ServiceNow, or Salesforce, or any of the other vendors.

So, you want to discover what's happening. Then, typically this is through observability and other platforms, you want to monitor. And then you want to take action depending on, so you'll have thresholds, guardrails, and if they are exceeded or if the guardrails are tripped, then you want to initiate some actions.

So, I actually built an AI steward team. Think of them as a just like there are cybersecurity risk professionals, AI risk professionals, who sit between product and between legal risk, et cetera, who would then look at this and take remedial actions.

The action could be if it's a minor problem, it could be like just reported and it goes through a change process. If it's a major problem, you might decommission the agent, send it back to school, so to speak. Right. So that it gets evaluated and trained and brought back.

And that all requires coordination across many different teams. AI is a cross-functional team sport. So, how do you bring everybody into one place? They're looking at a single system of record from which they can act. That's essentially what we're doing.

Game-changing combination: Robot orchestration and agentic AI

Charlie Isaacs: Robots is the game changer for me, because we're going to be able to take robots assisting and augmenting humans, and orchestrating that in an orchestration layer; agentic orchestration. And so, we could take a robot and a group of robots, for example, and onboard those automatically, dynamically create an agent that orchestrates all those robots, and coordinates activities with humans.
So, when you dispatch a robot—which is something we 0:31: demonstrated at Dreamforce— you dispatch a robot to go check on a gauge 'cause it might be getting too hot. The robot takes a picture of it, and the AI behind that analyzes that gauge. By the way, robots are the 3 Ds: Dangerous, Dirty, and Dull. Those should be doing those types of activities, right? And not necessarily replacing humans, but augmenting human activities that are dangerous and dirty and dull.

So, when you dispatch this robot, it says, "It's way over pressure." It dispatches another robot, creates a work order, it turns the emergency shut-off valve, and then that creates the human in the picture when it's safe for the human to get involved. That is a small use case of how this is going to expand in pretty much everything you deal with.

Agentic AI Ushers in End of the Hyper-personalization Era

Samir Meharali: The CX buzzword I never want to hear again is hyper-personalization. I feel like that's been overkill. People have been talking about it for probably a decade now. We're pretty much no longer in that era. Like, not every single individual customer should be treated as a unique individual customer. Everyone has some context that agentic AI could relate to, that the organization could benefit from and try to serve, and it'll get the job done quicker, faster, and customers would be more satisfied as well.

The need to separate signal from signal

Paul Greenberg:  The industry trope has always been you have to separate signal from noise. The actual problem is there's a lot of signal. The real issue is, how do you separate signal from signal? How do you do it?

Ashley Hart:  And a lot of times you'll see that marketers are are using what they they hear the buzzwords of AI and everything, but that to me is not the signal. The one thing that has stayed the same, that I have not seen as a change when it comes to product and the customer is, I still have to give them the right information for them to be able to go to their committee to say, "Here's what I need in order to be able to either drive more adoption of customers, I'm trying to expand my feature sets, I'm trying to solve what our customers need."

And that's always about purely information. Everything that we do is about how can we give them information and not overload them, but give them the right information for that time as they're going through their process.

I think in B2B tech, it doesn't matter if it's SMB, mid-market, or enterprise, there's an overemphasis of how much content do we need to give those different buyers at their different points of their journey. I don't think that's ever been the most efficient way to educate your prospect or even your customer to do an expansion of their of other products that they're already utilizing with you, but are you aware that you could be more efficient if you also add on this other app or if you do this extension? It's really more understanding their needs better so that you give them that just crucial critical information at that moment in order for them to make a decision.

Corporate boards expexct CEOs to be well-versed in AI

Vala Afshar: On a weekly frequency, certainly monthly—I find myself in front of boards. I've been in tech for 30 years. I didn't have the open access to CEO and their extended leadership, and then certainly boards, I didn't have that for cloud, I didn't have that for mobile or social or IoT or data revolution, quantum, distributed ledgers, you name all the new ingredients, new technology ingredients last 20 years. They were mostly IT conversations; CTO, CIO, CDO. Um, and now, uh, and it and right now, more than 60% of CIOs report directly to CEO. Two years ago, that number was like in the low 50s. So, we've seen almost a 15-point shift because I find that I believe that when CEOs are in front of boards, and the board members are asking about, "What's all this AI stuff all about?" Uh, if the CEO is not well-versed, um, explaining machine learning, computer visioning, smart robotics, natural language processing, generative, agentic, all these derivative sciences under this umbrella we call AI, then he or she would would would would feel uncomfortable. And so, now they're closer to the technology leaders within the enterprise, one-degree separation, learning along the way. So, the biggest shift for me is my audience are the most influential, biggest decision-makers at companies, regardless of size, regardless of geography.

The importance of creating conversational frameworks for executives

Esteban Kolsky: I actually took a couple months to build some agents to collect information. I call it Charles because I think ChatGPT is stupid. I built Charles based on a series of prompts and instructions on the to collect the information that I need. And every day and every week I get information from Charles. I get data points, I get insights, I get research, I get pointed to things that are new. And uh, what I do is I process all that information and I distill that into a monthly report that hopefully will be between six and eight pages when I get done. I'm trying to build a community for the people who receive this report. This to me is the most important part. It's not about writing or distilling information, it's about, like, you know, it's a way to actually say, uh, "This is what we're trying to do." The report, by the way, is not a report as much as a conversation framework. And the conversation framework is really what the report is about. It's like, you know, I give you the key talking points for the topics of the day and the questions that should be answered. So, the idea of this whole thing is to create not another report, but a tool that board and executives can use to have better conversations.

Agentic browsers are bringing next level invasion to privacy

Brent Leary: Is privacy worse off now, or is AI actually making privacy better off, in your estimation?

Tejas Gadhia:  Privacy's way worse off. A couple of days ago, OpenAI announced their browser, Atlas. Perplexity has a browser.  I want to say Opera is an AI browser. Agentic browsers are generally what they're going for. There's a small wave of privacy browsers for a bit, now it's kind of going into a little bit more agentic browsers. 

I'd recommend anybody go to Perplexity or any of these agentic browsers' privacy policies, probably just throw it in an LLM and say, "What are the privacy implications of using this browser?" And it's next level wild.

it's nice to think that the browser is going to be smart for you, but that thing can see everything you do in your browser. Already Google can probably see it in Chrome, but this is like next level invasion 'cause it also can see like the passwords you're typing and every input box you put of stuff. At least Google can only track like where you're visiting and things along those lines.

So, the privacy implications are out of control, especially when you go to like finance websites or maybe some kind of personal pages or even sensitive work pages. You don't know what actually is being sent and how that information is being used. And so, yeah, It's definitely a privacy nightmare.