When agents meet contact centers - two Salesforce Agentforce early adopters communicate the benefits they're seeing

As part of the wider ongoing push to ‘Agentforce-ize’ everything, Salesforce recently introduced Agentforce Contact Center, pitched by the firm as:

the only contact center solution that unifies voice, digital channels, CRM data, and AI agents natively in a single system, enabling customer self-service at scale, seamless AI-to-human handoffs, and real-time visibility across every interaction.

As usual Salesforce has been Customer Zero, this time with its customer service team as the early adopters, but there have also been 60 external pilot implementations, among them Compass Working Capital and Savant Systems.

They are two very different organizations. Compass Working Capital - aka Compass - is a non-profit financial services operation whose mission is to end asset poverty for families with low incomes and narrow the racial and gender wealth divides, while Savant specializes in smart home and smart power solutions, including energy efficient smart LED fixtures and bulbs for every room of the house.

But both organizations have contact center-related issues and challenges that they have had to address.

Challenges

For Compass, George Reuter, MD, Impact and Innovation, explains:

Compass is a remote-first organization, so we connect with our clients, and have for a long time now, primarily through several channels that haven't changed over time.] - email, text messaging, phone calls, online meetings. We innovate all those with our Salesforce platform using discrete tools. So, we have an SMS app that connects to Salesforce, and all the data lives inside that app. Inside the Salesforce, we have phone calls that are sinking into Salesforce.

But each of these apps have their own unique data models, and so, as an administrator, my team is spending a lot of time making sure that all of those integrations work correctly. And then it's also a training issue for our team. All of our coaches had to learn, ‘How does text messaging work inside Salesforce? Where do I go? I got this reply. Oh, wait, that wasn't an email. That was a text message. How do I get back?’. That kind of thing has been sort of a constant pain for us over the years.

Further to that, he adds:

Historically, I think we balanced two competing goals of meeting the needs of our clients, resolving the things that they come to us with requests for, and nourishing the relationship that we have with our clients. We think about that as balancing resolution with relationship, and we lean into different channels. Just going for resolution, a lot of times we can meet that need quicker with a self help portal or something like that. But we obviously use a lot of personalization in those channels.

If we want to lean into the relationship side, having somebody that's really dedicated to [customers], really knows all their goals, all their barriers, and so can pick up the phone right away. That’s great, except that staff person has a limited knowledge. Maybe they're specialists in home ownership, but if [the customer is] trying to start a small business on the side, they don't have that knowledge at their fingertips.

Meanwhile Savant deals with hundreds of thousand of home custom installs, explains Beth LeClerc, VP Business Systems Architecture & Web Services:

Every environment is different. Every single smart home has different products, different subscriptions, different software, different everything. They might be integrating with different third-party televisions that we don't manufacture or what have you. What that lends itself to bring us is a case complexity that is real and that probably won't change. Historically we have been forced to spend a lot of time where the agents have to summarize what is happening in the home, either right now or in the past.

Agentic impact?

So, what does the latest Agentforce incarnation bring to these two organizations? When it comes to that pull between relationship and resolution, Reuter says:

What agentic AI is really allowing us to do is sort of eat at the margins of both of those problems so that it's less of a competition, so we can make our self help center a place that does have some of that personalization, but maybe does know the client's goals and the barriers before it provides guidance.

As to staffers limited knowledge of topic specialisms:

Maybe their expertise is in home ownership, but they have at their fingertips an agent that can get them started with a customized plan around starting a small business. [That means] being able to say, ‘Well, maybe we don't have to have such a big trade off between these two competing values. Maybe we can do both in both places’.

The Agentforce Contact Center is delivering two main benefits, he goes on:

We kind of feel like we're leaping two [technology] generations. One, we're moving to a modern omni-channel presence, where a text message, a phone call, an email, all end up in the same place so the coach can reply to those things in a rational way. And the data models are really similar.

But at the same time, what Agentforce Contact Center is allowing us to do is embed AI tooling as part of how they manage that. And so for us, it's like a two generation lead where we're both going from sort of the old fashioned [approach], stringing together a bunch of apps to make an experience for our clients and our coaches into that omni-channel presence with AI tooling at the same time.

Agents are stepping in to free up human coaches to spend more time with clients, he says:

Our staff are primarily remote, so most of our appointments are on the phone. What we find when we do time studies with our coaches is that it can take them up to 30 minutes or so to complete the kind of data entry that we need. We know the relationship of trust that our coaches build with their clients over time is the number one reason that clients say they appreciate their experience with Compass, so that means those notes are really valuable because we have long term relationships with them. Our program lasts as long as five years, so it's really important to us to have that data.

But is it the most important use of [coaches] time to be actually entering in all these structured notes, like, ‘What's the current employment status? Are they going to any schools? What school are they going to? Are there any barriers that they have on their goals?’. All that kind of information we really want in a structured format, both for our reporting purposes and so the coach can pick up that relationship where they left off.

So what we're moving from is that 30 minutes that a coach is using to enter all that information. Instead, the coach hangs up the phone, and the agent immediately is on top of all that data entry. And the Salesforce agent knows our data structure. It has at its fingertips all of the field validations, all the conditional requirements that are needed, and can start proposing to the coach ‘Hey, it looks like these things came up in this conversation. Click here and we're going to enter this into Salesforce, and you can move on with your day’.

Benefits

That provides huge productivity gains, confirms Reuter:

We estimate that our 30 coaches, when they're saving 30 minutes on each of their appointments, they each have about 400 appointments in a year, so can save up to 6,000 hours across the agency every year. That's 6,000 hours that they can spend doing things like following up with a client that maybe has a more complex issue, or maybe following up with a client that we've fallen out of contact with in the last six months or so.

Over at Savant, LeClerc points to a focus on how agentic AI and contact center co-existence works in practice and doesn't disrupt, interrupt, or frustrate the customer experience:

We're lucky, because we can leverage a lot of our own cloud data and stuff like that, to influence and help the agents understand. So we're using the technology to do the same exact thing and help summarize and take manual work off of a [human] agent so that he or she can come into a call or a case or a chat, and have a better sort of understanding of where that home is and where it's been, based on summarization that the agentic team is going to do.

We also are using the technology in the same capacity to be predictive, to try to tell us when one of the homes or projects is reaching an escalation point. Typically, it isn't a customer sentiment that equates to the escalation. It's more, if my camera in my chicken coop isn't working, it's probably not as big [a deal] as a customer whose lights won't turn on inside the house or their power is completely down. So we are trying to also use it to prioritize what we would typically also have to manually escalate internally as well.

As for the future, LeClerc sees a lot more potential to be realised:

I think the technology is going to empower any employee to be one who can drive revenue...We’re going to be able to use this technology to say, ‘All right, you know, this call isn't going that bad. This person doesn't seem that upset or that mad. Hey, you know, they look like they've purchased product X in the past. Maybe remind them that we just launched product Y’. I think it's going to make the cliché of making call centers revenue generating that much easier to realize.