Inside Agentforce Contact Center: The New "AI Plus Human" Workforce

If your AI agent fails, customers escalate to a human and you “pay double.” In this CX Today interview, Gautam Vasudev, SVP Agentforce Contact Center at Salesforce, breaks down what CX leaders should measure as they roll out agentic service, and why cost alone is a trap.

We cover where AI works best today (high-volume, low-risk queries), where customers still demand a human (complex, high-stakes scenarios like healthcare and finance), and how the frontline agent role changes when routine work moves to automation. Vasudev also outlines how Salesforce thinks about giving customers a clear path to escalate, and why the real differentiation is not basic workflows, but what you build on top of them.

Transcript

SPEAKER_00 0:08

Hello and welcome. I'm Rob Wilkinson and today we're taking a closer look at a recent announcement from Salesforce about the launch of their new Agent Force contact centre platform. The first contact centre solution that's built directly on CRM. We already covered this in an extensive article on CX today. We looked at why this is a lot more news than just another Ccash play. But I really wanted to get closer to the kind of so what questions to dig into the what this really means for organizations, their CX teams, and obviously their customers. And I'm really lucky because today I'm joined by Gotan Fasudev, Senior Vice President of Ancient Force Contact Centre, who's been with the company for over 18 years, so kind of knows what's going on there. And he now directly leads the product for this exciting new platform. Welcome, Gotan. Thank you so much for joining me. Really looking forward to this. Oh, thank you so much, Chris. I appreciate it. So whilst I was writing that article, um, it got me really thinking about the reality of what this change means for the operators on the ground, um, for people who are actually going to have to use this technology ultimately. Um so I'm gonna kind of dig into that with you now. I want to start off though, with just a little of like a bit of a commercial reality check because what's driving kind of this sort of change and because cost pressure is is is one of the big drivers, right, that are behind um AI adoption. So, how do you think or how do you suggest CX leaders can ensure that efficiency goals don't quietly erode the kind of whole customer experience over time, because that's a risk, right?

SPEAKER_01 1:58

Yeah, uh great question and point. I think this is an important consideration because when you think about AI agents, you have to think of them as part of your workforce now. It's humans and agents. And oh, by the way, when your agents don't work, you pay double the cost because your customers are going to escalate to the humans anyway. So I think there's uh a real consideration that needs to be given to how you are tuning and optimizing these AI agents. You can't lean in too much into the cost angle of it. You have to go back to your core metrics of of course, cost is a part of it, but also CSAT and also revenue generation. And all of these are really experience moments that need to come together. I'll also say that at Salesforce, part of our best practice recommendation is to allow customers to have that discoverable moment to escalate to a human rep if they need it. And you'll notice in my response here the thread keeps going back to how do you create that best experience moment? How do you give customers the best uh resolution throughout the process? And when it works with AI agents, it's amazing. But when customers need to speak to human reps, they have the option to do as well as well.

SPEAKER_00 3:19

That's a very good point, and I think it leads us nicely into the discussion around customers because from what you're seeing with your customers, where does this AI-led approach, this AI-led service work well so far today? But where do where do to more to the point, where do customers still expect the human, like you mentioned there, regardless of any gains that they're getting elsewhere, so that we can align with that? Where is that happening?

SPEAKER_01 3:51

Yeah, you know, I always think in this space, we are all the experts because we've all experienced these moments where we've had AI agents or bots helping us and we wanted to speak to a human, or we've just done a Google search and found the resolution ourselves. So, where AI agents are really good is in those uh low calorie, high repetition use cases. You know, where's my order? What are the stores? So, either very standard low-risk workflows or knowledge-based questions. Where we see a continued focus on the human engagement is for those high-calorie, more complex, uh often multi-step scenarios where you as a customer are really looking for that next level of assurance that your issue is being resolved end-to-end. Right now, on one end of the spectrum, healthcare and finance are really good examples of this. You know, oftentimes you want that second degree of assurance that your care team really knows and is informing you what you need to do to prepare for that upcoming operation, for example. Uh, but on the flip side, if you all you want to do is know your order status, it's easy enough to find it on the web or speak to an AI agent about it in your uh your e-commerce app. So we're seeing this spectrum, and we are finding more and more engagements are starting to shift to agentic once the customer has found that great balance, that success in the first set of use cases. So high volume, low calorie goes to air agents, it starts getting more and more mature, you start bringing more workloads onto the air agents, and then you reserve your agent your human rep capacity for really those high-touch moments.

SPEAKER_00 5:42

So that's that's it's it's it's a it's a really clear shift, that's really clearly defined. Um, I guess it's around then what's the kind of impact of that? Because as AI is managing to handle more of the routine stuff, um, how do you feel the role of the agents on the front line evolving? Because that's the kind of the next, you know, it's always evolving in this space, the agent's role. With this technology, even more so, right? What do we need to be doing to support that?

SPEAKER_01 6:13

Yeah, and this is one of the most exciting opportunities in the space. Uh, you know, what when this Agentic Revolution happened, I think everyone's mind was immediately drawn to wow, for the first time, I have that scale to support rich conversations with every single one of my customers in a way that human reps could never scale to because uh there is a limiting factor in the number of reps you have on the floor. But something else has happened at the same time, and that is that the human rep, their role is evolving in a few ways. The first is that as AI agents become more mature, the human reps have breathing room. This is one of the hardest jobs that you can have. And so when they have that space to breathe and the context with which the customer is coming in, because the AI agent has listened to the customer, has summarized the conversation, and had a seamless handoff to the human, there's that breathing room moment, and the uh human rep can actually be empathetic on the conversation. And so they're starting off with better compliance, better greetings, and better customer recognition. That space also allows them to then do a better uh overall engagement, which drives up CSAD. But this is just the beginning. Now we are also getting into the space where you can do better role-playing with your human reps because your AI agents can act as coaches, and so now the human reps are getting upskilled. So we are finding more and more customers now embracing their reps, getting into more upsell and cross-cell motions as well. So that's becoming a really exciting space. And then finally, just remember that your AI agents are not just with your end customers, they're also operating as assistants for your human reps. So that new rep on the floor that's three weeks in can have high confidence and is actually operating like a three-year season rep because they both have the same assistant, and the assistants are always up to date with the all of the latest product catalog, all of the latest policies, etc. So it's made a huge difference. In fact, we find that the agent assistant, the rep assistance use cases are actually one of the first and fastest uptake use cases from our customers. They really want to go deep there because it's more contained, lower risk, and starts giving you ROI immediately.

SPEAKER_00 8:54

That's so we're basically saying that um it's providing breathing room. I love that. Um, and it's giving it's giving more support to the humans uh as well as more time to the humans to respond and react and to learn how to um handle the different more complex stuff that they're having to take on board. Yeah, absolutely not. So it's a very different shift to the doom and gloom that was previously kind of a concern in this space, isn't it? Around it's taking everybody's jobs and everyone's going to be too busy to do stuff where actually it's actually very supportive, which is that's awesome to hear that we're seeing that happen. Um I guess we do need to consider though, you know, the impact here in terms of as AI is becoming kind of more responsible and handling more different types of interactions. Um there's an accountability uh question to ask there because uh when things go wrong, and we know that they do, always can always go wrong. Who you know, who ultimately owns that customer experience and how do you kind of mitigate the risk of things going wrong from an AI agent?

SPEAKER_01 10:02

Yeah, I think this is another evolving space because I am finding that I'll I'll tie this back to our first uh discussion, which is AI agents are an extension of your human workforce, right? And so, in terms of responsibility, it is the same board level, C level responsibility to drive your end-to-end customer experience because these are all just part of those same customer journeys, those same moments where you differentiate yourself as a brand. I'll tell you two critical points here. The first is it is becoming more feasible for competitors to catch up with you because they now have all of these rich platforms to deliver the same kinds of uh basic service and basic service workflows. So it's becoming really critical to actually differentiate yourself. And the differentiation is um where all the magic happens. That's how you as a brand are standing out. And so this question about who is responsible for AI agents then becomes a responsibility of the whole organization because the whole organization is responsible for that customer journey. The second point I'll make is AI agents can be pretty clever. They can operate in a sales capacity, they can operate in a service capacity, they can operate in a scheduling capacity. And that emphasis kind of makes it necessary to take this uh customer journey experience not from a lens of, oh, this belongs to the service department and this belongs to the same sales department. It has to be a governance at the enterprise or the business level because these moments are shared across all these departments.

SPEAKER_00 11:56

So um forgive me, but it sounds like you're you're saying there might not be any more silos in operations. It's probably a big wish that, but uh hey, I like the idea of it. I think I can see where you're going with that, and I think I think that should be championed more. I don't think we do enough of making sure that you know embracing this new technology gives us opportunities to drop those boundaries to get rid of those silos internally and all get a comprehensive picture and contribute to what's going on for the customers. So uh I love that. I think that's brilliant. Um okay, I'm just conscious of time. We we do have a good idea of the kind of the business and the human picture now. So um you'll be glad to know we can now start to talk around the technology stuff, which uh which I which I know is uh is is really exciting. So uh, you know what what advice would you give to uh a CX leader today uh to ensure that the experience design uh side of things leads AI and platform decisions rather than the other way around doing the whole technology dictating what's going on.

SPEAKER_01 13:06

Yeah, and it's it's often easy to lose yourself in the excitement of the all the new tech. There's so much good tech out there to play with now. Uh the the basics still remain the same, though. The game is still the same. You know, you want to deliver great customer experiences at low costs and use these validation uh moments to then upsell your customers. You want to give them more of your product because you want to drive more value with your customers. Uh, and in that respect, your KPIs, your CSAT, your uh cost monitoring, your revenue, etc., still play a foundational role. So as you think about this, you know, going back to a previous response, what is your end-to-end journey? What are those really key experiences that differentiate you as a brand? You have to start mapping them out. And when you're able to map them out, you will start getting the blueprint of these are my highest value use cases. This is where automation could make a difference to give your my human reps more breathing room, as we just mentioned. And once you have that mapping, you can then go back and as an organization across departments, have those conversations about hey, how do we transform this experience for our customers? And then bring in the technology. So, really making sure that you're putting the customer at the center of the discussion and the experiences are driving the technology versus the other way around.

SPEAKER_00 14:42

So put the customer at the center of everything. I love that. I think you're right. I think that is so important, and it quite often uh doesn't happen when people get caught up with their internal things that they're trying to tick the boxes for, and um, even like down to the front line, they forget to invite them to the decisions and and and everything. So putting the customer in the middle helps to do that. Yes. Awesome. I I I guess um something that we need to talk around is the flip of this. So where when contact centers are are far more AI-led, um, there must be kind of signals that CX leaders can watch out for um so that they can basically kind of make sure that they go beyond the traditional metrics that the I mean we've talked about some of them today, like HandleTime and CSAT, but there will be an evolution now of the metrics, right? That's what's coming next. So um, what should people look out for or for in that respect?

SPEAKER_01 15:40

Yeah, there's there's a set of newer signals that we are discovering working with our customers, and I'll start off with the autonomous agent. This is the AI agent you're putting in the channel with your end customer. For those kinds of experiences, you're going to be looking at KPIs that give you a signal around how much containment did you have, right? Was your AI agent able to actually help this customer? Or was there a secondary escalation to a human level? Right? And these are volume correlations that you need to make when uh you really want to gauge the efficacy of your autonomous self-service AI agent experiences because you don't want it to seem like your AI agent is doing really well, but the number of human escalations are rising. That means there's something wrong. They're just escalating on a different channel. Um, other experience um or other signals that you want to track are what are the key areas that your customers are able to be successful with AI agents? And that means AI agents are very good at understanding intent and sentiment. And you should be on top of what are the intents and sentiments that are not working, so that you can have a continuous improvement loop. And this allows you to enhance your AI agents, but also for when your AI agents cannot help, you know when there's a human escalation, there is going to be a seamless handoff and your human reps are prepared to speak about those topics. So that's on the autonomous side. On the human rep side, there are uh newer signals where you have AI agents assisting your human reps. Now you can see how much faster you're able to onboard human reps. You can see what is the level of uh what are the CSAT levels for newer reps versus veteran reps. Because in theory, if your AI assistant is doing a great job, you should start seeing some normalization. And then the third dimension, which applies to both, is you start getting newer signals about if you're a voice agent, what is the tonality of the customer? You know, are they uh more heated, or is the tone of the chat more sarcastic? So there are these fun newer signals that you can play around with. And uh one thing we are discovering is when all of this is on one platform, that's when you get the best returns. Right? Because now you can see are your autonomous agents doing well? What topics are they not able to support? What's getting to the human rep? Are the human reps onboarding? And that drives a continuous improvement hoop.

SPEAKER_00 18:32

It's really exciting. This it's uh it's an it's a completely different way of looking at the data, isn't it? We're getting insights coming from every conversation and and actually both sides of the conversation, you know, not just customer feedback, all at the same time, and that come that that sort of intelligence layer is is gonna be kind of it's the future of where we find our our our solutions to not just our problems but our customers' problems as well. So this is like it's the love, I love this. It's it's so exciting.

SPEAKER_01 19:07

I will I'll summarize maybe in this way. Every enterprise needs three types of agents at least. One is the agent sitting in the channel with the customer, the fully autonomous agent. The second is the human rep assistant, whether it's in a sales or service capacity, they're helping the human rep be more productive. Uh but the third is the background agent, which is picking up all of these signals, as you're saying, the intents and the topics, and then making recommendations for improvement. These three things work hand in hand. Uh, in fact, that's the central thesis of this Agent Force Contact Center product that we've been building. Because the contact center needs to continuously improve.

SPEAKER_00 19:50

There isn't anywhere that um that has has um has more need and more ability to to evolve and improve though, because it's sat on this wealth of information, this wealth of data that's never really been tapped into before, right? We've always recorded the phone calls for training and internet purposes. We very rarely do much with that, and now we can do so much more with it. It's it's really exciting times. Um that that I could talk to you uh about this all day, uh, but unfortunately that is all time, all the time we've got uh today. Uh Gottem, thanks so much for joining me, answering all those questions. Before we do close though, um, for anyone who's watching this uh who wants to explore things in more detail, and what's the best way for them to find out more about Agent Force Contact Center or to get in touch with you guys?

SPEAKER_01 20:38

Um Yeah, I I would say definitely check out our trailhead for Agent Force Contact Center. You know, we we came in with this thesis as I said that we're going to be the first CRM native Agentic Contact Center. And when your channel data is side by side with the customer profile and your agents are continuously telling you how to improve, um, you can get a lot of benefits. But the best way to just understand all of that is go through the trailhead. I think it's a short one, it'll give you a good idea. And then if you want to engage, uh, definitely our Trailblazer community. That is going to be the one place where you will find other customers also who have this interest asking questions and you can engage with them, but you can also receive the answers for um you know our whole community to kind of receive and share and discuss.

SPEAKER_00 21:28

Yeah, the Trailblazer community is is is is awesome for anyone out there who who hasn't been in there before, and especially if when you get to like the think events like Dreamforce later on in the year and you get the t-shirts and everything. So there you go. Um so that's one one way of finding out. Don't forget as well that you can find a wealth of related resources, uh, stories, uh as I've mentioned, videos like this one, uh, all at cxtoday.com. Uh and uh other than that, that that wraps things up now. So I've been Rob Wilkinson at CX Today. Thanks everyone for For joining us, and we'll see you the next time.