Mike Boyle: Drew, what would you say the best-fit industry or business is for Agentforce?

Drew Becker: You know, that’s a really good question. And it’s. I think this is a thing where it’s really tough to box it in to say one particular industry because the use cases for it and the capabilities of it are really so broad. But there’s use cases for it in all kinds of industries. I’ve seen manufacturing companies utilize Agentforce. I’ve seen, entertainment companies utilize Agentforce. These AI agents are very adjustable, very flexible, and they can be used in a very broad range of industries.

Mike Boyle: We sort of touched on this lightly a moment ago, but I wanted to ask you to go a little bit deeper on this. What do you think sets Agentforce apart from other salesforce tools, specifically in customer service?

Drew Becker: There’s a couple of key things that set Agentforce apart from some of the other Einstein tools. And I think what comes to a lot of people’s minds when we think of customer service and AI might be, you know, the customer-facing chatbot, or if you’re familiar with Salesforce, the Einstein bot. And so usually when you have a chatbot like this, you can have some basic conversations with it on your website for self-support reasons, self-service reasons. Maybe it can point you in the direction of a help article or something like that. But what people don’t see on the end a lot of the times is that there’s a lot of setup that goes into, setting those things up. You have to account for every single possible type of question that could be asked, every single way that they might be asking through routing to a human afterward. And there’s a lot of things that could go wrong in order to prevent errors. The thing about AgentForce that makes it so useful and so dynamic is that you can actually set it up in a much easier fashion. Since it’s using generative AI and these AI agents that we’re talking about. What it does is it follows some natural language instructions. And so because of that, it lets you be very tailored to, different types of responses. And you can set it up very easily so you can say something along the lines of, you know, if the customer asks for the location of their last order, you should thank them for their time and go retrieve their order. Or you can say, you know, you should follow up every single one of your questions with a has this helped you? Or what can I do next for you? Or something like that. And it’ll be able to take in all of those requests and be able to really dynamically interact with the end users. And it saves a lot of setup time and setup effort on the front end, getting that stood up on say, an experience site or a website or any of these other channels.

Mike Boyle: Well, Drew, as you know, being a proud Salesforce Partner, we’re working with customers, you know, every day with this new Agentforce. So at this point, I realize it’s, you know, kind of early on, but what standout features of Agentforce, are our customers particularly finding helpful?

Drew Becker: There’s one particular thing that comes to mind that people have been finding very impressive that we use it for. We also have a lot of customers that are using it for this, and that’s being able to pull out answers from a very large base of knowledge. So if you have you know, hundreds of help documents on your support website. Agentforce is able to search that more intelligently than you would be able to search it, say from a search bar. And it’s able to pull out the specific chunk, you know, the specific paragraph or three sentences of the right knowledge article that’ll answer your question and be able to summarize those answers to end users. People are also impressed by this because what it can do is if there’s an answer to a more complicated question, say it’s stored across multiple different help docs, then it can grab multiple different help docs and kind of pull them together to summarize and answer for you. So there’s really a lot of capabilities. It really opens up your data and it really activates the ability to use those help documents in the best way.

Mike Boyle: And for those listening, I will be getting to some questions about integration and customization and obviously best practices and some advice as well.

But Drew, I just wanted to change the focus now to some real-world impact questions for you. In particular, how does Agentforce improve efficiency and customer satisfaction for service teams?

Drew Becker: There’s a couple of different things that are going on here with the Agentforce and since you can set it up with a lot of different capabilities, really the sky’s the limit comes to customer satisfaction and also you know, the ability for it to increase the efficiency of your service team. So a lot of people that are a lot of our customers are using this for case deflection. So self-service on their website, customer will log in and they want to find an answer to their question. There’s a lot of knowledge to search through, a lot of help documents to search through. So a lot of times they’ll open up the chat window and try to get in touch with the person. And so like I mentioned earlier, the Agentforce bot is able to summarize a lot of these answers and answer them for the end users. That way the messaging never even has to get transferred over to actual service representative. They’re more free to handle actual complex requests and things that really need human intervention as opposed to say just answering a question from searching a knowledge article that they can pull up and answer the question from. Also from the customer standpoint, if you can get your answer immediately instead of having to wait say five minutes to route down to a, to get in touch with an actual human agent, if you’re able to go on to the website and find your answers that you’re going to spend, you know, the next couple of 10 minutes or so searching through the website, trying to find the right document. You can get a lot of those answers a lot faster with the AgentForce agent just by asking it the question that you’re in there to search.

Mike Boyle: You mentioned case deflection a moment ago and I want to get into that a little bit deeper. But before I do, can you share an example or a story of how one of our customers is benefiting from using Agentforce?

Drew Becker: Yeah, and this is a really cool situation too. We have a customer that has a couple of different sales teams and they also have a lot of different tables. They’re on different PDF documents. They have these tables of pricing data and some other information from each of their different distributors. And their salespeople are usually in charge of grabbing a couple of different documents, cross-referencing different numbers across them, and then coming up with a final pricing number for the customer. But what we’re able to do is send all of those documents and have the aging Force be able to search those and Agentforce can do all of that table cross-referencing for them. And so they can ask their question. You know, they give in a couple of details and essentially ask, how much should I be pricing this much of this product at? Ah. And it can go and cross reference those tables for them and then pull up a final answer for them and save them all of the work of having to find all the different documents. Maybe they’re even paper documents stored in folders somewhere and having to spend all that work cross-referencing them, turning back to their calculator, going back to the pages. Whereas now they can just go into the Agentforce Agent, ask the question about the pricing, it can go pull all of the relevant data and then perform all the calculations for you and then give you the final number.

Mike Boyle: So instead of going to lunch at noon, now they’re going at 8:30 a.m., right?

Drew Becker: That’s exactly.

Mike Boyle: Let’s talk a little bit about use cases for Agentforce. In particular, you used the term a moment ago, case deflection. If you could just tell those who are listening, who aren’t familiar, familiar with the term, what case deflection is, and could you provide some resolution scenarios that are related to support desk personnel when cases enter the queue and then you know, once the person you know gets that to the human agent, how the issue is resolved.

Drew Becker: So when we talk about case deflection, really what we’re talking about is say you’re a customer and you made an order from Amazon and you wanted to go in and cancel your order. And, you know, let’s say that the only way to cancel your order was to go in and submit a formal cancel request. So a lot of the times you’d have to go in there, you know, submit your reason for canceling your order, get in touch with an agent, maybe answer a couple of different questions from the service representative about why you want to cancel your order and if you want a refund and, you know, whatever else goes into canceling an order. And so when you, when you do that, you have a human agent that’s taking up some of their time on your order request. Well, they also have other cases that are stacking up on their backlog while they’re sitting there working, because you’re not the only one who’s, you know, trying to cancel your order or making a return or requesting a refund. And so with AgentForce, what you can do with Agentforce is, as I mentioned earlier, not only can it interact with customers, but it can also take actions for them. So you can go in and request an Agentforce agent, hey, I need to cancel my order. Give it your order number. And the Agentforce agent can go through and handle the order cancellation action for you. And so in this way, the agents that are the human agents that are working on support can continue to work on the other cases that are coming up on their backlog, more complex cases that come in that require, you know, more support and maybe potentially more problem-solving steps to solve. And then some of these more simple issues can come in and be solved directly by the Agentforce agent without having to take up the precious time of the service representatives who are already busy enough.

Mike Boyle: Yeah. I want to ask, you a couple of questions regarding integration and customization specific to integration. How does AgentForce integrate with other Salesforce tools, and why is this integration important for maximizing its value?

Drew Becker: Yeah, so there’s a lot of great integration with the Salesforce tools. One of the primary ones that we’ll see is the integration with Omnichannel. The Omnichannel integration allows the Agentforce Agent to operate in the Omnichannel as a human agent would. So you can set a presence status for them to be active and then they can go ahead and engage in messaging with the end users. And so this is useful because not only can it be put on, say, an experience cloud site, but it can also be put on a website from another distributor or from another domain. It can also be attached to the Apple Messaging or any of the other advanced messaging channels that Salesforce has. There’s also a couple of cool integrations with some of the other Einstein features in Salesforce since Agentforce can do a lot surrounding cases and work order requests and things of that nature. So let’s say you have Agentforce creating a case for you for your customers who come in and they have advanced support problems. So Agentforce can actually fill in some of those, you know, required fields on the case that needs to be made. So usually that would be something like the subject and the description, maybe there’s some sort of type or something like that. And you can actually have the Agentforce agent fill all of that in based on the conversation they had with the user and then be able to send those down. Those can all be compiled. You can create knowledge articles with them from the Einstein knowledge creation. You can use knowledge creation to create articles based on these conversations that people are having with Agentforce if it resolves their cases. And all of these things can be integrated with Agentforce and it really makes it a, ah, nice end-to-end AI-powered service in Salesforce.

Mike Boyle: Drew, let’s talk about customization. Are there any options with customization within AgentForce that you have found especially useful for tailoring it to our client’s unique needs?

Drew Becker: Yeah, so one of the primary and most useful customization aspects of AgentForce is that it runs based on a series of natural language instructions. So essentially you would speak to it the same way that I’m speaking to you right now. And you just give it instructions like that on how to handle conversations a little bit about your company, maybe to give it some context about who you are and the kind of problems it might be solving. And what that allows us to do is it allows us to handle some pretty specific requests from our customers about how the agent should handle conversations. You know, should it be asking for feedback on its answers? Should it be offering to connect you to a, ah, live agent whenever you say something that goes against what it says? And so there’s a lot of different capabilities right there and it makes it simple with just an instruction here or there. So you can give it a natural language instruction to, say, ask for feedback every time you give a response and it’ll be able to do that. That makes it a lot easier from our side to make it more customized to exactly what the customer needs. Another aspect that I find really fun of customization in AgentForce is that you can put your custom icon on it and also name the Agentforce Agent, whatever you want. And that just makes it more engaging for the end user. And we find that a lot of customers find that to be more fun as well, being able to name their agent, give it a logo image on the chatbot, and have it interact with people as if it’s a, like a character, so to speak.

Mike Boyle: Yeah, but they find that pretty cool… Just two more questions. I wanted to talk about best practices and what you would recommend to ensure a, smooth implementation of AgentForce for businesses looking to adopt it.

Drew Becker: Yeah, and that’s a great question. A couple of best practices for AgentForce that I could think of off the top of my head. Number one is if you’re going to be using your Salesforce objects, so your cases, opportunities, accounts, contacts, make sure to label your fields, make sure to give them YouTube description. When AgentForce interacts with your Salesforce data, it’s reading out all of the descriptions of each of your fields on your object to determine what that is. So if you have something that says name, that’s the name of the field, but maybe it’s a P.O. number, right? And your description, you should say this is a P.O. number. And when you do that, it helps the Agentforce agent discern what each field is supposed to be, meaning when it goes and looks at your data. One other thing is, I mentioned earlier, a lot of people are really enjoying the ability of the Agentforce agent to parse important information out of, you know, a large body of, say, help documents or policy documents. And so one thing is to make sure that you have all of that together. You could either be in knowledge articles or some sort of centralized folder or repository of PDF documents, but either way, getting those all together in a way that it can be ingested into Agentforce is extremely important to be able to have all the data there for it to be able to parse out.

Mike Boyle: And lastly, Drew, what advice would you give to an organization, a business, evaluating whether Agentforce is the right fit for their needs, or considering an upgrade to their current customer service solution?

Drew Becker: I think my best advice would be to stay open-minded with AgentForce and with AI agents. This is such a cutting-edge technology that there’s new use cases, new capabilities for it being constantly discovered, you know, every day that you use it. So what I found is that really the best way to learn about what you want out of your AI agent is to learn what your AI agent can do and being able to kind of go in there and ask questions and see what types of questions people are asking, it really starts planting the seed of some ideas of, hey, what can we do next with this agent? Right, right now, you know, it can answer questions from our help documents. But maybe tomorrow it can create cases for us. Maybe the next day it can, you know, respond to emails, maybe the next day can, you know, do some other capability that you come up with. And so with these AI agents, there’s a lot of explorability and the capabilities of it are really feedback loop between the humans that are using it and the capabilities of the AI. There’s so much to be discovered once you put one of these up. And so really getting started with it is the biggest step. And then from there, usually there’s a lot of different capabilities that get unearthed as you start to explore and discover with the AI as moving forward.

Mike Boyle: Drew Becker is Ad Victoriam’s ‘AI Guy’. What a cool title. I like that a lot. Thanks for, stopping by today, giving us kind of an overview of Salesforce Agentforce. I want to have you back and I think I want to get into a little bit deeper with use cases. So we’ll do that, in the near future. Sound good to you?

Drew Becker: Yeah, that sounds great. And Mike, thank you. I really appreciate you having me on today. This was awesome and I can’t wait to do it again.

Mike Boyle: Yep, looking forward to it. And to the audience. If this happens to be your first time listening to the podcast – we – Drew and I would appreciate you giving us a five-star review. If you’re listening to us on Apple or Spotify, it just helps us push the word out there about the Ad Victoriam Salesforce Simplified podcast. And don’t forget to hit the button to subscribe as well. I’m Mike Boyle from Ad Victoriam Solutions, thanking you for listening to our Salesforce Simplified podcast. As always, our next episode is just around the corner.

Announcer: You’ve been listening to Salesforce Simplified, the podcast from Ad Victoriam Solutions.