Zack. Eric doing great, man. How are you? And it’s Friday. Feeling great. Ready to talk about Agentforce. Ready to talk about Salesforce. I am excited to be doing this with you. You know what better topic for a Friday morning than I am an Agentforce. I’m two cups of coffee and I’m ready to get this. I’m about halfway through my first one, and I think I might have to get up mid-conversation to get another one.

It’s going to be great. There we go. There we go. Hey. Well, you know, for you and I, we live in this world, right? We think about this day in and day out. But for most people out there, Agentforce maybe a foreign concept. So let’s just start there on what is age for some, why should business owners care?

Yeah, I think the easiest way to describe it is it is how to get LEM capabilities within your CRM. So that is just the basic way to think about it. Right. And non marketing speak Agentforce is Salesforce is solution to integrate generative AI directly on the Salesforce platform. So it allows you to build, customize, deploy autonomous AI agents to help with tasks in your CRM.

So simple use cases like connecting your knowledge base to a service agent where somebody can go and just get immediate answers to anything from the knowledge base, as opposed to having to scan the website or call and or email and, and get a response that way. Right. There’s also internal use cases. So if you want to chat with an AI copilot, that can help you summarize records in the CRM, or you want to ask questions about data, and it can just go and search for that data for you instead of having to click through a bunch of different screens.

So there’s a lot of use cases for it, but ultimately it really just boils down to Agentforce is how you can leverage generative AI internally for your Salesforce or as well as externally for users to be able to interact with data that’s stored on a platform or even connected to anything through data, cloud. So a lot of people watching probably heard of ChatGPT, right?

ChatGPT is a large language model. That’s the one that’s in the news. You get a new one of these launch and every other week everyone’s that’s more advanced than the one before. The real benefit to Agentforce is it allows those sort of large language models to leverage the data internal to a company. So the data that’s housed in that CRM.

So previously, a customer may have gone on to a help site and typed in a question and maybe had to sort through several knowledge based articles. Now the agent will kind of they can they can speak directly with the agent. The agent then can go access that data and answer the question for that’s fair enough. Yeah. If we think about kind of what generative AI is, it’s it’s grounded on some source of data.

Right? So ChatGPT is grounded on just a monumental selection of data across the internet, across books, really just everything. Right? It’s it’s being fed a ton of data and information for general purpose so that you can ask questions, ask it to generate things for you, things like that. Well, Agentforce is going to be much more constrained. And that’s a good thing for your business, because it’s not going to be answering questions based on data that isn’t related to your business.

So we call that grounding, which is which is the source of the data that the AI is allowed to access and then generate responses from. And because Agentforce is grounded on your business data, it’s more trusted, it’s more reliable. You have a much higher confidence that what it’s going to be generating is going to be based off of information specific to your business.

And the way that we do that is by connecting it to the CRM, by connecting it, connecting it to other business sources through data cloud. If you have other systems where you’re storing customer data that the agent should be able to interact with, and it’s also highly trusted through a product, it’s called the Einstein Trust layer. And it’s doing things like masking personally identifiable information before it gets sent out to the LM.

There’s a zero retention policy with OpenAI in terms of the data that’s sent from Salesforce to the LM gateway with an AI. So there’s a lot of sort of safety protocols that come in place as well, which would you don’t necessarily get if you were going to use a more general purpose? LM is that break it down for me.

Like if I’m a business owner, why do I care? Yeah, yeah, I think there’s a few reasons that you would care from just a simple what’s the the lowest hanging fruit. Why would you want to get started with AI? I think really it comes down to know that this technology isn’t going anywhere. Right. We’ve kind of gotten to the point that we understand that there’s going to be business utility well into the future.

I think there’s a lot of conjecture around it. Some people are saying this could fundamentally change everything about our lives. Maybe that’s true. I don’t know, right? We’re going to have to see where this goes. But I think what we can agree on is this is, at the very least, pretty helpful tool that businesses can take advantage of.

So I think if you want to get started with AI and you recognize the importance of forward thinking and understanding that this is going to affect us in some way into the future, it’s a smart idea to start dipping your toes into it, right? Like maybe you want to get started with AI, but you’re not sure where to start.

Well, if you run your business on Salesforce or you’re thinking about running it on Salesforce, that is sort of the easiest way for you to get started and actually getting results with it, because you already have the data, you’re already running your business from a trusted source, and you’ll be able to easily create some sort of an AI tool that can leverage that data and provide value.

So now I’m making a lot of assumptions there, which is you trust the data that’s in your CRM. And if that’s not the case, yeah, the AI is going to produce bad results. Right. So you do have to have some foundational trust in the data that you have. Having clean data, accurate data is important. And the other piece is understanding your business process.

If you can’t articulate a process, even in just a conversation like this between you and me, there’s no way we’re going to be able to articulate that to an AI, right? So you do have to do some foundational, just fundamental business work to where you can articulate a business process and then rope in an AI agent to help with that process.

And then that could lead to greater productivity, that could lead to the agent being able to handle some of the processes that maybe take a lot of time, that are low value and allow your actual human employees to focus on higher value conversations and higher value tasks. So that we’ve been involved in Salesforce AI for several years, and we’re at the forefront.

When age, of course, was announced, we’ve actually built several of these agents for some of our clients. Talk about some of those use cases we’ve seen and some of the problems we saw. Yeah, yeah. And to your point, Einstein is not new. AI is not new. This whole generative concept is, you know, what’s really been around for two years in terms of the cultural zeitgeist, right?

But predictive AI and other forms of AI have been around for a long time, and we’ve been working with things like predictive scoring inside of Salesforce for a while. So seeing this expand to the generative side is exciting because it allows us to do just a lot more sort of tangible things with AI instead of spitting out predictive scores, which is not minimizing the power of predictive analytics, but being able to combine it with something actionable is really cool.

Basically saying, hey, you can now do things for us, as opposed to just giving us information that we can take action on. You can now take action. I think that’s what’s really exciting about it is for some use cases, we we’ve had the opportunity to work with several customers. We’ve had a DNA genetic testing company where their goal was to reduce case handling time by 30%, and we help them implement an agent to basically make work towards achieving that goal.

They have a knowledge base in Salesforce. We connect an agent to that knowledge base, and that is now deployed on their website, and they’re able to have the agent handle sort of that first line of defense on all of the questions that may come in about, hey, what test do you offer? What are these results mean? What does it mean if this is showing up on my test?

And so at least anecdotally, we’ve seen that they’re achieving an effective 40% selection rate, which means essentially 40% of all the inquiries coming into that agent do not escalate to a human being. So that’s a massive time savings. And really that took that was a 2 to 4 week process to implement that and get that up and running.

So, depending on the use case that you have, it may not be terribly complicated to get it built, and you can do it in a relatively short time span. Another business we worked with is an insurance brokerage for employers. So they provide health plans, medical insurance plans for employees at different businesses. And they get a lot of questions about the policies that they offer and what are the deductibles here and copays, when’s my renewal date and things like that.

Right. So that is typically a phone call or an email and somebody internally having to go and pull that policy information, find the details and then answer those questions. Right. Well, now with the agent that we built for them, the customer is able to just ask the agent. The agent will identify who they are, verify their email address by sending them a code and making sure that they actually are who they say they are, and then it can go and pull down their related policy information just by knowing who that person is.

And that’s all through the power of their CRM. So the CRM stores the information about that person and their business and the policies related to their business, and then it can pull those answers from that data, give them tailored answers specific to the policies that they have. So really simple use cases. But I think really powerful and saves a lot of time.

And of course that’s really just the foundation, right. Like we expect that a lot of businesses will take that and then continue to grow, either find more opportunities to enhance those agents to be able to have more capabilities or creating other agents to handle, you know, other specific sort of business processes or problems. Yeah, I love that. I’m just thinking, you know, there’s two businesses you mentioned that’s a better customer experience.

We’re getting better answers more quickly. It’s saving the businesses staffing expense and things like that. It’s making sure that they’re getting consistent quality of that, that knowledge answering those customer inquiries. So across the board is phenomenal stuff. What other agents have you seen from Salesforce? I know they’re they’re constantly rolling out new agents. There’s a bunch on the roadmap.

So talk to me about some of the other use cases Salesforce has in mind that some of the other agents are they’re visioning. Salesforce. Just put out an interesting report that analyzes some of the data that they’ve had from the past six months. And there is an 86% split. So 86% of all the agents deployed across everything in Salesforce have been service agents.

So part of that is it’s been around the longest. Another part of that is I think those are the most tangible, easy to understand use cases. So service is by far the largest use case so far. Now as as we get more types of agents and Salesforce comes out with a build your own agent. But we may see that change and shift.

But right now predominantly services the biggest use case, there’s an SDR agent, which allows you to automate outreach to leads in your CRM. So essentially a new lead comes in and and secret agent can identify that lead craft an email to go out to that lead and then nurture them through the top of that funnel. So for example, if that lead comes back with questions about your business and your products, or the agent is grounded on data about your business and your products, and it can respond to those questions with the goal of booking a meeting with the sales rep.

So everything that that agent is going to do is essentially going to funnel down to a call to action, to book with the sales rep and actually get that lead into the sales funnel right. And then once it gets into the sales funnel, that’s a human right. A human sales rep is taking over from there. And and actually having that high touch conversation, moving through the negotiation process, hopefully closing that opportunity.

Right. Another one is a sales coach. And this is more for internal use in Salesforce. It’s it’s actually pretty cool. I’ve used this myself. You can embed a window in your opportunity in Salesforce. And as a salesperson you can record yourself talking about different parts of the business. So let’s say you’re in the negotiation phase of an opportunity.

Well, that agent can understand what your business process is, how your sales motion operates, because this is all provided to it with unstructured data like PDFs. So you can take something that you’re probably already providing to new sales reps when you’re onboarding them. Hey, this is how we sell. This is what we sell. This is how to talk about this.

This is how to handle objections. Right? You give that to the agent so it understands that context. And then I can sit here just like we’re having a conversation and let’s say practice a pitch or actually go through a negotiation. And the AI will pick up on details about the opportunity and it will throw objections at me, and then it will record how I handle those objections.

And at the end of it, it’s about a five minute sort of conversation back and forth. You will analyze that conversation and come back with recommendations around, hey, I noticed that you did this really well. Good job. Hey, you could have handled this objection better. And here’s why. And that’s just it’s pretty cool, right? Because I think in a vacuum it’s not that cool, right?

Like any tool can kind of do that. But because it is trained on your data and your process and it’s pulling information about your sales motion and combining that with the specific opportunity in order to give you that feedback could be a really good opportunity. Just even if you’re a very confident salesperson, you know, any deal, you’re working where you want a little bit of extra practice and you have the time to do it.

This is a great opportunity to get some real time feedback without having to, you know, rope in a friend to do that for you. So those are kind of the immediate agents that are available today. There’s a Agentforce for set up agent that is available now. And it’s I would say in its earliest stages, and it’s growing in capabilities into the future.

But the idea is if you’re a Salesforce admin working in the back end of setup, you can have an agent now where you can start asking questions about set up, so you can ask questions about access to different objects and fields, compare permissions. Users ask it to tell you where something is located, what objects do we have in the system.

And so that’s that’s growing and capabilities over time. At TD, I talked about giving the agent the ability to literally compare two users side by side. So if you know, you and I are in the CRM and for some reason I can’t see something, but I know you can see it well, I can ask that agent, why is Eric able to see this?

But I can’t and that agent will compare the permissions and just spit the answer back out to you, which is a massive timesaver, especially for those of us that have been in the Salesforce ecosystem for so long. That would be like finding a needle in a haystack. You’d have to go through so many different screens in the back end to figure out where all the different permissions are for these two different users, and profiles and permission sets, and to be able to have an agent kind of do that for you much faster, by the way, and for you instead of you having to do it yourself.

I think that shows a lot of promise. So again, that’s not available today, but that is coming into the future. That’s sort of the long term roadmap of Agent Awesome set up. So those are really the primary agents that we have today. And then the long term goal here Salesforce is working on a build your own agent. So if you didn’t want to start with a template of well I want this for service, I want this for sales, I want this for internal coaching, I want this for set up.

You’ll have the ability at some point to actually create your own agents or to start from scratch. And that kind of exists today. You can take a service agent and sort of repurpose it, right? Not use the standard topics, create your own sort of purpose behind it. But there will at some point be like a tailor made purpose built to build your own agent sort of interface.

That’s going to come super exciting. You mentioned earlier, in the conversation, you know, Salesforce’s annual conference. What else? You were there. A bunch team members went, what else did you all see regarding Agentforce in terms of what what we can expect to see pretty soon? Yeah, I think the biggest takeaway is this is Salesforce’s focus.

Agentforce is sort of their primary focus. Everything is going to touch Agentforce in some way. So not to take away from any of the other products and services that Salesforce has to offer. Those are still super important, but Agentforce is going to be plugging into all of those in some ways, it’s a big topic of conversation.

It’s not going anywhere. Salesforce is doubling down on it. So I think any investment into Agentforce, so we can safely say Salesforce is going to continue to put a lot of time and effort behind them, so that that’s one main takeaway. I would say there’s several other things coming, but one is, there’s a lot more learning pathways available.

If you’re trying to get started with Agentforce as a configurator or a consultant or an admin. So, trailhead now has what what they call Agent Blazer status. So everything’s a blazer with Salesforce. May you’re a sales blazer. You’re a trailblazer. You know, your service blazer, data blazer, agent blazer. So the blazer part is your trail blazer.

And you’re specifically a trailblazer using Agentforce Agent Blazer. So the Agent Blazer path now has several different training modules for a force. And they’ve done something unique with this where they split it into three different levels. So there’s there’s like a level one, level two, level three levels one and two are available today. And level three is coming into the future.

And this is a way to sort of start with baby steps, go into sort of, you know, the middle ground with that level two. And level three is really focused on pushing you towards doing a lot with an agent, getting into practice orgs and building them out and then actually getting your agent for specialist certification. So that’s an exciting update.

You know, we’ve got 66% of that released. And the last sort of third of that is expected to come in the near future. And another big announcement that I think anyone that works on the development, admin, consulting, any side of Salesforce where you configure rating, if configuring any side of Salesforce, where you’re configuring, there’s now an Agentforce Developer Edition org.

So for a very long time, ever since AI, an Agentforce and Data Cloud has come out, you could not get a developer edition that included these features. So the only way to do it would have been to be a partner and to go in and get an SDO, which is a simple demo org, which is really meant for for demo purposes.

Not really configuring and building things that you could deploy later or package up. Well, now you can do that. There is a dedicated Agentforce and data cloud developer Edition org so that you can build and test and develop in your own practice environment and beyond. Just practice. You can even build solutions and package them and deploy them.

So that’s a really exciting update. And then beyond that, there’s of course a bunch of roadmap updates that you can check out from the Roadmap Exchange at Salesforce. There’s, a lot of focus on making Agentforce more friendly to developers. So integrating it into the Salesforce CLI as an example and making it easier to package and play things.

There are a lot of other sort of updates coming. There’s a really cool feature called an AI workbench, and this is not available yet, but it’s essentially a really powerful testing suite that’s going to come to Agentforce, that’s going to replace the current testing center, and it’s just more robust. It allows for a lot more testing scenarios and feedback.

And then there’s going to be a wonderful thing called Interaction Explorer, which for businesses that rely on being able to analyze every conversation that the agent is having to understand patterns, interaction Explorer is going to allow you to do that. So it’s going to essentially allow you to look into every conversation that that agent is having. Pull out some analytics and data about those conversations in aggregate help you determine every time this topic comes up.

This agent is answering it in a way that I need to change, right? Maybe it is not answering this particular topic in the way we expect, and unless you you didn’t have those interactions, you wouldn’t be able to identify that. So now that we’ll have the ability to do that, we’ll have the opportunity to surface some of those, patterns and issues with the agent and refine those instructions to help it respond in a better way.

So a lot of updates. I’d encourage anyone listening to go in and just check out the Roadmap Exchange at Salesforce, check out some of the sessions from TD, Salesforce plus. A lot of that is just available for free that you can go check out. And then of course, a lot of folks like Salesforce, Ben, have some great recaps of everything that’s been going on.

So, just check that out and see if anything kind of piqued your interest. Zack. It’s been a great conversation today. So the takeaways, I think, you know, at a high level get started somewhere, right. If you if you’ve been thinking about it, start with your data, start with your business process and start playing with the tools. I heard the other day that the AI we’re using now is going to be the worst AI we ever use, and it’s right.

It’s the worst it’s ever going to be today. Right? And so the two messages there is you need to get started now. You need to start understanding these tools. And then the flip side of that keep humans involved I think, you know, for for a little bit, just until we have a little bit more faith, always keep a human in the loop.

And one thing I’ll say, you know, before we wrap is I think, agent for an AI, if you’re not involved in this world and you’re trying to or maybe you’re just on the edge and you’re not sure what to do, it can be intimidating. This is a wild leap in technology. And the actual technology infrastructure itself is pretty complex, right?

But I think a tool like Agentforce allows you to get started, and it’s not that difficult, is relatively plug and play. Obviously every business is unique and there’s, you know, that’s where we can come in and help figure out your business process and apply that to the technology. But it’s not like you’re going to have to start from scratch and build your own lab and build all the logic and code behind that.

The whole benefit of this is you can leverage generative AI specific to your business in a relatively short timeframe. So by no means what I minimize the amount of effort that needs to go into planning and configuration, and any of that. Right. But what I do want to get across is I realize it’s intimidating. Before I got involved in it, I was intimidated by it.

I’m still intimidated by it in some ways, but it’s not as hard to get started as I think it feels like it is. Before you make that jump. This is a great way to really get started with it without having to commit to transforming your entire business. You can just take it one step at a time, and if you like it, then you can start expanding and building upon it, which I think is a really great way to get started with a new technology.

Awesome.