In this Dreamforce interview, James Watson, vice president of customer excellence at Smartsheet, joins theCUBE’s Gemma Allen and George Gilbert to break down how AI is transforming enterprise support operations – moving from reactive break-fix to proactive, predictive service. Watson details Smartsheet’s “front door / back door” approach: automated deflection for common questions and AI-assisted research for agents. He shares concrete outcomes including cutting complex case triage from six hours to six minutes, handling ~30% year-over-year growth in support volume without headcount increases and boosting deflection from 32% to 60%. The discussion also covers responsible AI adoption, reducing agent attrition (~20% to 6%) through better tools and engagement and early pilots for predictive maintenance across Smartsheet “constellations” of data nodes.
The conversation connects to theCUBE’s broader Dreamforce angles around agentic automation and trust. Watson explains how Smartsheet trained and continually QA’d models (from early Einstein AI work through Agentforce), using both structured knowledge and large volumes of unstructured historical cases, with human-in-the-loop review to prevent hallucinations and refine prompts, terminology and data sources. He highlights how Agentforce 360 and tighter context accelerate high-quality responses and enable more proactive, value-adding customer interactions – turning support insights into cross-sell, up-sell and training opportunities. Watson also cites a survey of 25,000+ customers reporting more than four hours of weekly productivity gains with Smartsheet, even before bringing these new agent capabilities into the workflow.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Dreamforce 2025. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for Dreamforce 2025
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Dreamforce 2025.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Dreamforce 2025. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to Dreamforce 2025
Please sign in with LinkedIn to continue to Dreamforce 2025. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
James Watson, Smartsheet
In this Dreamforce interview, James Watson, vice president of customer excellence at Smartsheet, joins theCUBE’s Gemma Allen and George Gilbert to break down how AI is transforming enterprise support operations – moving from reactive break-fix to proactive, predictive service. Watson details Smartsheet’s “front door / back door” approach: automated deflection for common questions and AI-assisted research for agents. He shares concrete outcomes including cutting complex case triage from six hours to six minutes, handling ~30% year-over-year growth in support volume without headcount increases and boosting deflection from 32% to 60%. The discussion also covers responsible AI adoption, reducing agent attrition (~20% to 6%) through better tools and engagement and early pilots for predictive maintenance across Smartsheet “constellations” of data nodes.
The conversation connects to theCUBE’s broader Dreamforce angles around agentic automation and trust. Watson explains how Smartsheet trained and continually QA’d models (from early Einstein AI work through Agentforce), using both structured knowledge and large volumes of unstructured historical cases, with human-in-the-loop review to prevent hallucinations and refine prompts, terminology and data sources. He highlights how Agentforce 360 and tighter context accelerate high-quality responses and enable more proactive, value-adding customer interactions – turning support insights into cross-sell, up-sell and training opportunities. Watson also cites a survey of 25,000+ customers reporting more than four hours of weekly productivity gains with Smartsheet, even before bringing these new agent capabilities into the workflow.
In this Dreamforce interview, James Watson, vice president of customer excellence at Smartsheet, joins theCUBE’s Gemma Allen and George Gilbert to break down how AI is transforming enterprise support operations – moving from reactive break-fix to proactive, predictive service. Watson details Smartsheet’s “front door / back door” approach: automated deflection for common questions and AI-assisted research for agents. He shares concrete outcomes including cutting complex case triage from six hours to six minutes, handling ~30% year-over-year growth in support v...Read more
>> Hi, and welcome back to theCUBE, here at the Moscone Center in San Francisco. I'm your host Gemma Allen, along with my co-host, George Gilbert, and we are here at Dreamforce, hearing all about just how much AI is transforming the enterprise. Joining us today, we have James Watson, the vice president of customer excellence at Smartsheet. Welcome, James.>> It's a pleasure to be here.>> So, we've been hearing a lot of buzzwords, a lot of hype around AI for quite a while now. You and the team at Smartsheet have a front-row view into how exactly it is transforming enterprise and people's day-to-day, right? Can we just start there? Can you give us some examples?>> I would love to give some examples. AI is really changing the entire industry, especially with somebody like Smartsheet. Smartsheet is this intelligent work management platform, and we're primarily business-to-business. So, our business customer uses our tool to do automations and intake of customers, management of customers, projects and portfolios. So, it's really flexible, but because it's so flexible, the way one customer uses it is different than another customer. So, every time they call and they need some customer service or support, it's like a snowflake. Every single case is a little bit different. So, what we went through was this transformation of, and I'll use an example. We have something called control center. A customer in the past would call and say, "My control center is not working because I have all these nodes of data, and one of those nodes is not working." It's very difficult to understand which one it is, but they say, "I have this problem."
When we get that in the support organization, it would normally take us six hours to go and just do the research and figure out where do we think we can respond? We don't know. It just takes that much effort and then we respond, and then the customer comes back and says, "No, no, that's not my question. My question is something entirely different." And then, we have to go through these iterations. It would normally take two days to be able to get that done. So, we had to put AI into this to be able to help us do those complex administrative tasks so much faster. So, now when a customer comes in, they use that AI, we get context. And then, when the agent gets it, they already know exactly what the question is. We use AI on the back side, we call it the back door, and we do all this research, we generate a response. And all of that takes six minutes. So, we go from six hours to a six-minute experience, and that drives our customer satisfaction way, way up.>> Wow, that's incredible.>> So, I've got the Klarna example on the brain, where they might've backed off a little bit, but they said they were able to downsize their customer support organization dramatically. You were telling us that you've been able to handle 30% volume growth in support incidents with no headcount growth, what was it, the last 15 or 18 months?>> Yeah. Yeah. I mean, just year-over-year, we have a 30% increase in the volume of cases that are coming in. But because we have two things, the front door we call it. Smartsheet uses smart doors. We have a front door for the customer and a back door for the agents. On the front door with this deflection, customers ask the simple questions and were able to answer those without any human involvement at all, which is great. But that went from a 32% deflection to a 60% deflection. So, that takes a lot of that demand, that extra 30% and covers a lot of it.>> Before you get to the back door, maybe unpack a little bit that context that you needed about each problem, how it was collected and organized, so that an agent could reason about it? Because as you were saying, everyone is a snowflake. How did you organize all that data?>> There was a lot of unstructured data. Structured data is very easy. We have a lot of articles. It has headers, it has rows, it has all the information, but the unstructured data, that's all of the cases that we've done in the past. So, we have all of these hundreds and hundreds of thousands of cases where someone asks a similar type of question. They may use a different terminology on what they call the row or the column header, but we can use all that unstructured data. So, we had to train Agentforce to be able to do this. So, we spent about, it was total of about 18 months, but about six months before we started with Einstein AI, which is the precursor to Agentforce. And then, we did another 12 months of just continuous quality and freshness. And then, we turned it on and we did three months just in-house of QA. And anytime something came up and it had that hallucination answer or something that wasn't exactly what it should be, we went straight back to the source. Now, we just had to correct these pockets of sources. And now, we have this great data, unstructured and structured data, and we continue to feed it. Every case continues to feed that loop of learning.>> So, everything that can't be deflected is like a Tesla autopilot disengagement event?>> Exactly.>> And it's a teachable moment?>> Right.>> Okay. So, has the chunking and indexing strategy that, I gather has been improving in Data Cloud, made it easier to prep that unstructured information? In other words, does it feedback from failed outputs or failed responses to natively restructuring the data itself for a better response?>> Yeah, with a layer of people in between. So, anytime something doesn't come back, absolutely correct, it goes into our QA team who says, "What went wrong? Was it the learning model or was it the data?" And then, we go fix whatever that is. We might have to continue to teach the model on what to look for and add additional verbiage or words that might be used in that type of a case. Anytime it spits out and someone says, "No, this wasn't a good-quality answer," then we get it retrained.>> So, those are some very impressive metrics. And I think sometimes when you hear about something being made so simple, you might be a little bit skeptical, not me, James, but some people might.>> Yeah, Gemma.>> But tell me, where's the low-hanging fruit? What industries and verticals are really able to adopt this and avail of the benefits very quickly, and where are the gaps? What makes it challenging for some and easier for others to really get immediate benefit?>> And I would answer that with sometimes any company regardless has some easy, quick wins. The low-hanging fruit is probably going to be something that builds an ROI for the customer and an ROI for the company. And that's where support comes in. Customers get their answers faster and a higher quality so they can get back to their business continuity and get back to what they think is most important. And for us, that level of deflection and efficiency inside of our team allows us to keep our head count working on more valuable things so we don't have to do a lot of backfills because the people that we have are so much more efficient. They can work on what is more valuable. How do I help you go onto your next use case or your next knowledge learning on top of Smartsheet? That's what this does. It helps us to be a better support agent and really give the customer service exactly what they need.>> So, is it fair to say that customer services in your shop because you're able to deflect so many of the routine calls? It's not so much as you're automating them away, as you're redeploying them to higher-value tasks, which instead of reactive, it's almost like proactive on the journey to help them get more out of the product. Just the way many organizations now try and turn customer service into a revenue-generating, you're into a product success organization.>> No, I think that's a great point. Moving it towards this higher level of service and this proactive motion. Traditionally, a support organization is all about break fix. Something's broken, they go fix it, but they're very technical. They are very expert in how to have empathy with the customer, as well as the product. Those are the right people to put in front of the customer and say, "Now, that we have resolved your issue, what can we do next? How can we expand it within the customer?" And that becomes a revenue-generating or cross-selling opportunity. Something we're working on right now, it hasn't come out yet, but we're working on this idea of how do I use AI to really look at the entire solution? So, we call it constellation. You have this constellation of nodes of data. And if something breaks here, it could be three or four nodes away from the dashboard that a customer sees. It used to take all this effort to go and find out what was broken, but now I can start to look at this and the refresh rates and the data transfers and predict when your solution is going to break.>> You're describing AIOps in a little microcosm.>> Right.>> Okay, that's beyond break-fix.>> Right.>> That's proactive predictive maintenance, in a sense.>> And we're already starting to do that. We're already in that pilot stage of creating cases before something breaks. I always think of the idea of I'm driving my car and when that check engine light comes on, that's when I have a break-fix. But what if I could have AI tell me a week before that check engine light's going to come on? That's the magic.>> It's interesting because I think what a lot of people still think about customer support, you think about tickets. It's the old-age resonance, but it's really now more about efficacy at scale. And I know it can sound like a sloppy term, but it's about partnership, I'm sure, right? And you gave us some great examples there in the pre-chat around Smartsheet itself as an exemplar in efficacy, especially as it comes to talent and how you think about that from an overall organization perspective. Can you chat to us a little bit about that?>> We want to make sure that we use AI in a very responsible way for our customers and for our employees. We want to make sure that it's helping empower everybody. And I think that's really the magic. When you take a human, especially an expert human like in support, and you combine it with a tool like AI that can really do all the administrative work that makes it a better work experience for this employee, they have more fun doing their job because they have the right tools to do it well. They want to be in support because they have the opportunity to serve a customer and help a customer get over a sticky situation. So, our attrition rate, when I first took on this transformation, it was around 20%. Now, it's 6%, and that's because we have such a high employee engagement score. People want this. They're so excited. And our customers love it because now they have a better personal relationship with that support agent. And it's not that you get the same person every time, but because we have this back door that all this history when you call so Gemma, if you called in as a support need, we would know how many times has Gemma called in, what has been her history? What does she use the product for? So, I don't have to be the same person, but I have all of that information and context, so that I can serve you and empathize with you and get your stuff resolved.>> And we're here at Dreamforce. It's a very exciting time and it's such an exciting time too, in the industry broadly. But there's also, I think a little bit of trepidation around what's coming next. What are your early impressions here in San Francisco? What excited you most about this morning?>> There's a lot of great things in Salesforce. And we chose Salesforce, I want to say, as a trusted partner. Not only do we use them as our native CRM platform, but we trusted them, and they came along on this journey with us to make sure that in our AI journey and their AI journey can merge together and we can both benefit. And some of the great things that are coming out of the announcements today, like the Agentforce 360, it's what we currently do, and it's just going to get beefed up, they just have greater context. And that's the biggest business problem, is context and communication. And that is sped up and more effective with all of these announcements coming out of Agentforce today.>> Absolutely.>> So, if you look at the trajectory of the technology, what you're hearing from other vendors, but the trajectory of which Salesforce is integrating it all and simplifying it, what do you think some of your new workloads might look like and new outcomes that you might achieve when we're back here next year?>> Yeah, I think it's going to be how do I help someone on their customer journey to get maximum value? And I'll share another really good statistic. When we surveyed a bunch of our customers, surveyed more than 25,000, and they said on average they save more than four hours of productivity every week by using Smartsheet, and that was before we introduced something like Agentforce. Now, that we have the AI tools on the front door for service, we have the back door to get your answers faster, and we've embedded AI into this intelligent platform. Now, imagine how much time you can save. So, for a customer jumping into this Smartsheet experience, now it's all about efficiency. And if it's four hours or five hours or six hours, that's time to do what they think is most important. It might be doing more work, it might be having dinner with your family, but that's important time. That's the value of Smartsheet, it gives that back. And I think you're going to have customer success and professional services or sales will have a better context on when I talk to somebody, is it the right person? And when I do talk to them, what do they need? Without having to go through two or three weeks of discovery, you know going into the conversation?>> So, it sounds like it's a reinvention of the product and the processes around it.>> Absolutely.>> So, that it's a much more fundamental transformation than the GUI, which just made simplified and standardized interaction with the user experience. But it's also much, much more profound than mobile in the cloud because those were tweaks. Mobile was a richer user experience that was on the go, but the real change was that it was a back-end operating model, which was invisible to the user. So, that allowed a lot more apps to be created and distributed, but now you can change the experience of the app, and then the efficiency with which you can deliver this value-add services.>> And it's that process on the back-end that makes a difference. If I'm a customer support rep, if I'm an agent and I have a conversation with you, George, I'm not just fixing your here-and-now problem, I'm trying to fix what your next problem is going to be by giving you, is it additional training? Is it we call a plus-one program? Is it giving you some value enhancement to that conversation? Because calling in your greatest time of need, it's the greatest time to be able to help you on your next step, but that has not traditionally ever been the support organization. That has been for training and development or for the customer success team to work on, but now we are all in that boat. So, even if a CSM is talking to a customer, they can have the same tool and they can say, "I understand where you are. I can help you with training. I can help you with your next step use case, or I could even help you understand if it's a simple fix," because they're knowledgeable enough to do that. So, now what we did is spreading across the entire organization and every interface with the customer.>> Because you mentioned CSM, I think just to talk a little about the human element to this process for a second, right? Because one thing we also know about customer success and service is you want somebody to blame at the end of the day if everything goes haywire, right? You want a head on the chopping block.>> Right.>> So, how do you think about that in a world of agent-to-agent, and how do you think about the strength of the business relationship too? You have agents managing agents and you have maybe a CSM or how... Actually, what do you visualize? I know what I visualize, but I'd love to hear it from you.>> Yeah. The CSM has a slightly different remit and responsibility than the support organization. They are working at the account level and we're working at the individual user level, so often, the two don't cross paths very often as far as the audience that they have, so we don't have much conflict. But I think you still have the opportunity at that CSM to know the full spectrum of that account's experience. How many users called in? What did they call in about? Was it an access question? Was it a billing question? Whatever it might be, but they can see that because they use that backdoor AI to help them out.>> So, the opportunity is huge for some thesis and for, as George said, a more predictive model, where you're trying to always predict what the next big bet is for a customer and where the opportunity might lie. Absolutely.>> Yeah. And that helps with those cross-sells and up-sells. If I'm a support person and, Gemma, you're calling every other day, I'm like, "Well, do we need to do some training? Let's sign you up for that. Or maybe you need a technical account manager? Let's sign you up for that." So, there's a lot that we can do from the support perspective that has never been done before because we just didn't know what that next step should be.>> Well, James, it was wonderful to have you on. I think I've certainly learned a lot about the next gen of customer excellence. Thanks so much.>> I appreciate it. Thank you for the time.>> I'm Gemma Allen, here with my co-host, George Gilbert. We're live at Dreamforce 2025. We'll be back with our next guest after this short break.
