Today on the Salesforce Admins Podcast, we talk to Tom Leddy, the Product Director of Decision Guides at Salesforce. Join us as we chat about decision making in the age of AI and why cleaning up your data is more important than ever.
You should subscribe for the full episode, but here are a few takeaways from our conversation with Tom Leddy.
Almost exactly a year ago, we had Tom on the pod to talk about the Well-Architected framework. I’ll link the episode below but Tom gives us a quick summary: “It tells you how to build healthy solutions with Salesforce and what a healthy solution should look like,” he says.
Making your org healthy comes down to looking for patterns and anti-patterns. Essentially, you want to do things in a way that sets you up for long-term success.
Tom and his team are hard at work rolling Decision Guides into the Well-Architected framework. These walkthroughs are designed to help you decide which Salesforce tool is right for you when they have overlapping functionality. The answer is going to depend on your specific use case, so looking at a Decision Guide can help you understand the full picture and make the best choice for your business.
Looking forward, Tom sees a lot of potential in combining AI tools like Einstein Copilot with the information in Well-Architected and Decision Guides. There’s a lot of potential to make things more interactive or quicker to digest, but you’ll still need to do some critical thinking and make your own decisions.
In terms of incorporating AI tools into your org, Tom is working on decision guides for that, too. “A lot of the cool AI features are not going to work very well unless you have a good underlying data strategy,” he says. Working through the Well-Architected framework will help you create a solid foundation to get the most out of these new tools now and in the future.
If you’re a frequent listener, you’ll know that we can’t have an episode about AI without mentioning just how important it is to have clean data. As Tom points out, this extends to patterns and anti-patterns as well. It’ll be easier than ever to roll out code to your org and create new customizations, but you need to be sure you’re doing it the right way and not crippling yourself with technical debt.
Luckily, Tom and his team are working on tools to help you make sure your org is, well, Well-Architected. Stay on the lookout for a Data Strategy Decision Guide, coming soon™, and new ways to assess the health of your org with Einstein Copilot. The future is bright, and hopefully a little more organized.
Be sure to listen to the full episode for more tips from Tom, and don’t forget to subscribe for more from the Salesforce Admins Podcast.
Mike:
This week on the Salesforce Admins Podcast, it’s all about decision-making in the age of AI because we need to make decisions too. And Tom is here to help us. In case you’ve been following along, this is a follow-up from when Tom was last on the podcast to talk about Well-Architected. They’ve gone even farther now and they are combining Well-Architected and Decision Guides, and well, that’s all in the podcast. But I also had a lot of questions about the role that AI is going to play in helping admins and architects make decisions. So we’re going to get into that today, which is going to be amazing. Also, you get an update on Tom’s marathoning, but looking ahead at the rest of April, because we are kicking off April now with our Decision Guides, next week we’re going to have Lizz Hellinga on. We’re going to talk about AI and clean data.
You remember she was previously on to help us get geared up for that. I’ve got plans to have Skip Sauls back. He’s going to give us a Data Cloud update. He literally stopped me in the hallway at TDX and said, Mike, I got to get on the podcast and update all the admins about Data Cloud. So that’s going to be amazing. And then, of course, at the end of April, we’re launching a completely new style of episode. Josh Burke is going to be running the last episode of every month. And he’s going to do a deep dive format with different people in the ecosystem and different Salesforce leaders. We’re going to start this one off with Katie Holmes on design and AI, and really how to be a design thinker. It’s going to be amazing. So that’s our April, which I think is pretty phenomenal. But before we get into April, let’s talk about decisioning in the age of AI with Tom. So let’s get Tom on the podcast. So Tom, welcome back to the podcast.
Tom:
Thanks, Mike. It’s great to be back.
Mike:
I totally planned this because you were on a year ago in April, ironically, right after TDX, which is when we had a chance to catch up. And you were talking about Well-Architected automation. I’ll include the link in the show notes below. But for those people that have yet to listen to that episode, let’s talk about what Well-Architected is and maybe what you’ve been working on since then.
Tom:
Sure. So Well-Architected is, it’s a framework that’s available for, it’s on architect.salesforce.com and anybody can access it. And what it really does is it tells you how to build healthy solutions with Salesforce and what a healthy solution should look like. It contains information about guidance. How do you think about things like security and compliance as an architect, but also what are specific patterns and anti-patterns? What are things you should be able to physically see in an org? If this setting is X and this other one is Y, that’s not a well-architected solution. You need to go and change one of those two settings. And one of the things that I’ve specifically been working on this year is we actually have another resource on architect.salesforce.com called Decision Guides. And up until recently, Decision Guides were kind of their own little thing that we’re hanging off on the website by themselves.
We always knew that they were valuable because what they do is they’ll give you a trade-off analysis between different Salesforce tools that sometimes we have different tools that kind of overlap with each other and look like they might do the same thing, and you need to know when to use which tool in a certain scenario. That’s been there for a while. But last year when my colleagues and I were going around and having discussions with our ecosystem about Well-Architected, we would get two pieces of feedback is number one, the people that already knew about Decision Guides, we just kind of assume that they were part of Well-Architected, and they would talk about how great they were. And then we were sitting around, “Okay, cool. This isn’t what we came here to talk about, but thanks for the great feedback.”
Or then people who didn’t know the Decision Guides existed would actually ask us questions about, “Oh, why doesn’t Well-Architected have information about how to select the right products?” And then we would bring up, “Well, actually we do, but it’s a different tool.” So one of the things that we’ve decided to do this year is bring Decision Guides under the overall Well-Architected umbrella. And then we’re going to start enabling deeper integration between that whole product selection, trade-off analysis, and the rest of the Well- Architected framework. So that’s going to be evolving over the next few months, and it’s pretty exciting. And specifically, Decision Guides are my area that I’m going to be focusing on.
Mike:
I mean, I like that. I like that for many reasons. But the biggest reason is critical thinking. I think critical thinking and walking through, and not only just giving people the answer, obviously that’s not critical thinking, but showing them possible paths, and helping them understand pros and cons to choosing different tools is just so fundamental to where tech is. Tech, 10, 15 years ago when I was doing things in Salesforce, you wanted to automate something that was a workflow, that was it. The Decision Guide was to automate or not.
Tom:
Right.
Mike:
And I’m using that as an example because since then, we’ve had a few different automation tools and some which are being moved forward. But I think the ability to just sit back and problem- solve like that, because the solution to a corporate issue isn’t as black and white as sometimes we make it out to be in terms of how you can configure that technology solution.
Tom:
I think that’s a good point too, is if you look at the overall Salesforce product catalog, you’ll probably notice that there are a lot of things, products and tools that they look like they kind of do the same thing. And maybe some of them even have overlapping functionality. And you might be thinking, “Well, why do you need all of these things?” And the reason is because when you think about it from a business perspective, maybe you’re solving similar problems, but there’s always going to be little nuances where if you have this specific set of requirements, tool A is actually the best tool to use to solve that problem. But if even one of those requirements changes, now you might want to look at a different tool because it actually works differently and it’s going to be a better fit, and you’re not going to run into as many issues down the road if you use that tool instead.
Mike:
Right. So now let’s talk about the Einstein in the room because AI is everywhere.
Tom:
Right.
Mike:
And I feel you watch the TikToks on things and there’s a guy on TikTok that talks about how to use different AI tools and write blog posts or create videos. Do we need Decision Guides in era of AI?
Tom:
I think we do for a couple reasons. One of the things is at some point it might be nice to maybe… AI might be able to give you a summary of a Decision Guide and say, “Hey, here’s what I’m trying to do,” and it’ll give you the options. But at the end of the day, as an architect or an admin or really any technical professional working with Salesforce, you have to make a decision yourself of, I need to go in this direction because maybe the choice that you need to take involves purchasing a tool, and you need to contact your account executive, or we already have this tool, it works. We have the right people to support it. But really, only you know things like, do you have a staff of developers? If there’s a tool that requires coding skills, do you even have people that can support that? Right?
And AI isn’t necessarily going to tell you that, but it probably will give you a nice summary of here are the pros and cons of these tools, and then you still make the final selection. And one of the other things that we actually have coming out, this is Safe Harbor, but this is one of the things that I’m working on for later this year, is we’re going to be adding a new Decision Guide, and it’s specifically going to be about org strategy and data strategy overall. And one of the reasons that this is important is because a lot of the really cool AI features that are available and that we keep getting rolled out almost on a daily basis now, they’re not going to really work very well unless you have a good underlying data strategy, and your data is high quality, and the right places that your data’s coming from, and you’re able to analyze it properly. So being able to do that correctly to even enable AI to work in the first place is going to require a series of decisions that we’re hoping to encompass in that guide.
Mike:
I mean, AI works because of data. Right? So it goes back to the very second thing you learn as an admin besides setting up a profile, which is whatever data you put in the system is how it’s going to be reported back out.
Tom:
Yeah, exactly. And it’s funny because I think that message gets lost sometimes is that if you think about it’s how it’s going to be reported back out in the form of a report that you would run, that’s true, but that’s also how it’s going to come back out. And if you’re asking an LLM for a response based on whatever data you’re feeding it, it needs good data or you’re going to get, it’s the same old garbage in, garbage out, but now you just happen to have AI in front of it.
Mike:
Yeah. Looking at… I mean, we’re fresh off of TrailblazerDX, and you probably saw Prompt Builder and Einstein Copilot, and those just got launched. Thinking ahead for architects and admins and people that just build software solutions in Salesforce, how do you see those tools augmenting what they do with a Decision Guide?
Tom:
So long term, this might be a… Well, we’ll see what the timeline on this looks like, but if I had my ideal what the future would look like, it would be somebody is doing configuration in a system, or maybe they’re a developer and they’re writing code, but alongside them, they’ve got their Copilot window that is giving them recommendations of you realize you’re about to create an Apex class when a flow might actually be better to do this. Or you’re about to wire this up to build an integration to this system, have you thought about maybe using a middleware tool? Because just doing a direct call out is probably not going to be your best option in this scenario, and it’s kind of powered by what the content that’s in the Decision Guides.
Mike:
Yeah, I mean, there’s so much possibility there with both, knowing, I always come back, and you’ve probably heard this, well, you probably have him, but the listeners have of AI feels like the first time in math class that you got to use a calculator. And I remember the teacher, I was like, “Yes, totally going to get all the answers right,” because I’m horrible at math, horrible. But then she dropped this just mega real-life bomb on me, and it was, “But you need to know that whatever you put in the calculator, it will give you that answer, but if you’re not putting the right data in and you need to know what to expect.” And I was like, “Oh, I really just wanted the calculator to figure this thing out for me. Don’t you know what I’m trying to do?”
Tom:
Yeah. It’s funny that… Because I remember that too when first got to use calculators in math class, and it wasn’t the simple arithmetic at this point. It’s not where you’re just doing four plus four on a calculator. It’s usually this complex algebra or calculus or something like that where you have to go through a series of steps to even figure out what you’re putting in the calculator in the first place, and then assuming you’ve performed those steps, right? The calculator will give you the right answer exactly. But it’s kind of the same way.
Mike:
In thinking through, because a lot of what admins do or tasks that architects do, we sit down. I’ve seen examples of having Einstein start to create a flow for us. Looking at that and looking at some of the architect, the Well-Architected stuff, I’m feeling like they’re kind of complimentary, right? It’s almost kind of if you go to Utah, they have this fry sauce that supposedly makes french fries taste better, but french fries taste good all on their own.
Tom:
Totally.
Mike:
So should we be seeing a, let’s use Einstein copilot maybe in a way to give us ideas we didn’t think of?
Tom:
Yes. And I would say, like you said, those two things are definitely complimentary in the way that if you look through any of our Decision Guides, they’re actually packed with a lot of good information, but they’re also 45-minute reads. And if you want to read the whole thing, I mean, that’s great, but if you’re able to get a quick summary from Copilot or from our Decision Guides or anything else on Well-Architected really, where it might give you something that either you didn’t have time to read through the entire guide, or maybe you read it but you missed it because it’s a lot of content, whatever. But yeah, you’re right. It might be something that you could see in Copilot and say, “Oh, I didn’t realize this was even an option. But yeah, this looks great.”
Mike:
I noticed at TDX, one of the really cool things that you and your team do in the architect area is you have these kind of see-through boards, but essentially whiteboard or diagram stuff. What were some of the things maybe brought up at this TDX that you saw different from last year’s TDX? Because a year in technology can be like 10 years in the rest of the world.
Tom:
Yes. So a lot of the things that we covered in our workshops this year were a lot of data strategy and a lot of just general automation strategy that some of the underlying foundational pieces that you need to have in place to get ready for things like AI, like we talked about earlier. So last year we were solving problems that were more like general architecture problems that are still definitely important. You need to make sure that you have the right integration strategy if you want to connect multiple systems to Salesforce. But this year, a lot of companies want to roll out AI functionality, and it’s huge. It’s going to be a point of discussion for a long time. So we want to make sure that you also have all of the foundational pieces in place to be successful when you’re doing that.
Mike:
Yeah, I would agree. Thinking through AI, what have you heard from architects in terms of tools and AI to help them learn and maybe accelerate faster in understanding software architecture?
Tom:
So there’s a few. So there are tools like Copilot, for example. There are LLMs in general. There’s a variety of third-party tools that we’re seeing that are coming out. But really what I’m more excited about are some of the official Salesforce tools that I’ve seen on our roadmap that people can use to help make decisions. There’s AI tools, like you said, that’ll help you start creating a flow or create an Apex class.
And then maybe you’re going to, obviously you’re going to want to review that ahead of time. You’re not going to want to just generate a bunch of code with AI and throw it in your production environment. But it is something that will, I think, will speed up your time significantly as a developer. And then from an architecture perspective, it’s more going to be a matter of making sure that you’re taking a step back and looking at everything you’re creating because you’re going to be creating it so much more rapidly, making sure you really have a handle on what is being rolled out and what all the downstream implications are so that you don’t push something in, like I said, too fast into production and end up with a domino effect where it causes some other issues down the road.
Mike:
Yeah. Or increased reporting complexity.
Tom:
Yes. That’s another one.
Mike:
One thing that looking at what we talked about last year that I want to touch on, ironically, it’s towards the end of the podcast, but it’s still important, remediating technical debt. And I think, too often, our discussion around both Decision Guides and AI are around net new. Always, what can we create net new? But really, I’d be interested, what’s your take on Decision Guides and using some of our Einstein Copilot and Prompt Builder to remediate technical debt? Should we think about it differently?
Tom:
Yeah, I think the biggest difference is if you’re creating that new, it’s nice because you have this greenfield environment, you can go and look and say, “What is the best way to do that?” But the reality is we all know that this far along, the chances of going someplace where you have a brand new org with nothing in it are smaller and smaller. You’re going to, more than likely, if you’re starting at a new organization or if you’re a consultant and you’re working for a company, the first thing you’re going to do is you’re going to have to untangle a bunch of things that were maybe created that might’ve even made sense at the time when they were created a few years ago, but the business changed or whatever.
And now it doesn’t make sense to have an org set up the way that it is, and being able to use AI to identify some of the anti-patterns that exist in your org, and then figure out the steps to remediate those, and then maybe even prioritize them. So it’s not just saying, “Hey, there’s 15 things that you have to fix, which you can see in a health report today, but here’s the ones with the biggest impact and here’s what you should be doing to fix them,” I think will be super powerful, and it’ll make it, going forward, a lot easier to remediate technical debt, for sure.
Mike:
Yeah, I’d love to see Copilot really dive into Health Check and optimize, especially Optimizer.
Tom:
Yeah, for sure.
Mike:
That would be to prioritize, and also to think about linking together changes, that would be a really cool feature to see.
Tom:
Yeah, definitely.
Mike:
Yeah. Well, I appreciate you jumping on. I know we touched on it last podcast, but how is the marathoning going?
Tom:
Great. So I just, about maybe 12 days ago, ran the LA Marathon in Los Angeles, and it was a great race. It was the first full marathon I had done in a while. I had been doing halfs more recently, but this one, the weather was great. The course was a little bit hilly, but overall, I mean, I got to run through Hollywood and Beverly Hills and-
Mike:
Oh, wow.
Tom:
Got to see a bunch of friends in Los Angeles, and it was a really good time. I had one friend in particular that I ran it with, and we had a lot of fun during the race.
Mike:
And to recap, I think, I don’t want to speak out of term, but you’ve ran a marathon in every state, is that right?
Tom:
Yeah, every state and every continent.
Mike:
Man. Okay. Wow. Holy cow. That’s really cool. I appreciate you sharing that. Well, thanks for coming on and getting us updated about Decision Guides. I think one thing that I’ll definitely be looking for is the data strategy Decision Guide that you mentioned. We might have to have you back to talk through that, because I’ve done a few episodes on the importance of clean data and making sure that you think through your data strategy, especially with AI now. So I’ll look forward to that.
Tom:
Yeah, I am as well, and I’m working with the product teams on that and getting a lot of great input on it already. So yeah, I would love to come back and talk about it once it’s live.
Mike:
Will do.
That was another great discussion with Tom. I appreciate him coming back, and I’m still kind of a little amazed that he’s ran a marathon in every single state and on every single continent. I don’t know that I’ve done anything in any single state and every single continent, which is pretty cool. I’d love to know, have you done anything like that? I’ve also never done a marathon, so that’s pretty neat. But I want to do one thing. Make sure that you are following the Salesforce Admins Podcast and sharing it out. It’s really easy to share. All you have to do, if you’re in iTunes, you tap the three dots and you click the share episode. Then you can post it social or text it to a friend. Maybe they’re working on building out some decisioning and want to know how to well-architect things. And of course, there’s a ton of resources on everything, including the links in the show notes that I mentioned during the call and a transcript.
It’s all at admin.salesforce.com, so be sure to join there. But with that, until next week, I’ll see you in the cloud.
Today on the Salesforce Admins Podcast, Admin Evangelist Josh Birk sits down with Kat Holmes, Chief Design Officer and EVP at Salesforce. Join us as we chat about diversity, accessibility, and her book, Mismatch: How Inclusion Shapes Design. You should subscribe for the full episode, but here are a few takeaways from our conversation with […]
Today on the Salesforce Admins Podcast, we talk to Lizz Hellinga, Consultant and Salesforce MVP. Join us as we chat about why product management principles in Salesforce are crucial if you want to take advantage of new AI tools. You should subscribe for the full episode, but here are a few takeaways from our conversation […]
Today on the Salesforce Admins Podcast, we talk to Lisa Tulchin, Senior Curriculum Developer at Salesforce. Join us as we chat about choosing the learning path that fits your learning style and strategies for training your users. You should subscribe for the full episode, but here are a few takeaways from our conversation with Lisa […]
