David Dixon, who joins to talk about the world of Salesforce Analytics, is Director of Analytics at Atrium.ai.
David has 20 years of experience with every aspect of data, from database administration to data migration, integration, and analytics including Einstein AI.
Main Points
- What’s new in the analytics and AI space on Salesforce these days?
- How does Einstein Analytics fit together with Tableau?
- What are the basic skills you’ll need to get started with AI and data science?
- What can we expect for the future of analytics on Salesforce?
Links
- David’s LinkedIn Profile
- Atrium.ai
- David’s Dreamforce Session on Tips and Tricks for using Einstein Analytics
Contact Me
If you enjoy this podcast, please reach me out to let me know your thoughts or simply say Hello. My (Xi’s) contact info:
Take care, until next time :)!
Transcript
AI-generated, unedited.
[00:00] Hi. This episode's hosted by Andrew Davis, who was the guest of twenty seventh episode Salesforce DevOps. So Andrew will join me in the future to host some episodes from time to time. I hope you enjoy this episode, and will support him just like how you support me.
[00:36] Hello, everybody, and welcome to the Salesforce Way podcast. This is Andrew Davis, your host, and. I'm very delighted to have with me today, Dave Dixon, who's the director of analytics at Atrium AI. Hey, Andrew! Thanks for having me. Very excited to be here today. So, Dave, you've got a fantastic background. Tell us a little bit about your background, your role at Atrium. Certainly. So, right now, currently, I'm director of analytics at Atrium AI, and we're really focused on delivering the intelligent experience, really combining analytics. Artificial intelligence, machine learning, and surfacing those insights where people work and where they can be most, where those insights can be most beneficial. So, we're a platinum partner of Salesforce's. Been around for a few years, and most of us have sort of deep. Industry and ecosystem experience related to Salesforce. So personally, I've been working with Salesforce for about eleven years now, and working with the Einstein product way back when it used to be called Wave. So back in the early two thousand five. And my background, really, my twenty years of experience, has mostly been focused on. Data, data architecture, business intelligence, and CRM. That's an amazing background. I mean, if you if you look at the Dreamforce keynotes, if you look at what Salesforce is doing with Einstein over the last few years, it's huge. Obviously, it's a massive initiative that they've got, and very exciting. But it's. You know, well outside
[02:06] of the normal daily life for many of us, Salesforce developers and so forth. You're right in the heart of that space. What are some of the things you're most excited about in the analytics and AI space on Salesforce these days? Yeah, so probably a sort of an overarching focus or area that I'm excited about with Salesforce has just been their continued focus on data and analytics and machine learning and artificial intelligence. I mean, you go back and you look at you know the MuleSoft purchase, which in and of itself was huge. Getting that plumbing, being able to bring. Data together. You look at the the Beyond Core purchase, which was the product that then became Wave. So, starting to get into that analytics and BI capability, and then you you look at the you know the predictive capabilities, the DataRAM purchase, and then of course, most recently, you know the fourteen billion dollar purchase of Tableau last year. I just love the the continued investment in sort of the. The foresight and vision that Salesforce has around analytics, machine learning, and data, and how those play to make people more effective in the CRM, marketing, and sort of service space. So. I think that's probably the investment that Salesforce is making is probably what I'm most excited about. So that investment obviously indicates that they see there's a huge potential, and obviously.
[03:36] You know, we're starting to see that, but I suspect we're probably just in the early days of seeing what Salesforce is going to roll out in this space. You mentioned Einstein Analytics, Wave. You mentioned Tableau. Probably, I'm not alone in being a little confused about how they fit together. What do you think is the vision for how those two tools complement each other, fit together, bring benefit? Yeah, I mean, obviously, there is a fair bit of overlap when you look at sort of the descriptive analytics capabilities of what Einstein provides, as well as what Tableau does. Traditionally, Tableau has played heavier in the on-prem space, and you know, playing data where it lays in sort of whether that's an operational store or your finance systems or SAP. And really strong capabilities around connecting to that data and pulling that data together and making it easy to visualize and explore those data sets and sort of your your core systems. Whereas with with Einstein, well, obviously you can bring data in from other systems. Once when you bring that data in, you are putting it in an Einstein data store, right? So you are making a copy of that data. The pros of that are you get some really quick performance in terms of being able to slice and dice, drill in, ask questions, you know, and answer the next question, right? The cons is you have you have a little bit of data latency, but first and foremost, where
[05:07] Einstein has played has been in the core Salesforce space, so service and sales and service analytics, right? So sales related use cases, so. They've played in a little bit different areas traditionally. In the near term, you're going to have both products because architecturally there's some some pretty significant differences. And there's some complementary pieces as well. So if you look at Einstein and that Einstein Analytics Plus SKU and Einstein Discovery, you have the ability to make predictions. So do linear and logistic regression. Make predictions around the probability of something occurring, or what we think the value of something will be at some future point in time. And Tableau natively doesn't have that prediction capability. And just for the listeners to clarify, descriptive analytics being looking at the present and past historical trends, as opposed to projecting out into the future, basically making reports and dashboards based on existing data. Oh, that's great! Yeah, I'm sorry. Yeah, it's playing buzzword bingo. That's absolutely right. Like, yeah, so we say descriptive analytics. It usually basically relates to looking at things that have happened, and then we'll say, you know, predictive analytics for things that we think are going to happen. So I think I think there's some benefit to be had in terms of having Einstein being able to make those predictions, and then potentially surfacing those. Up through Tableau for users that wouldn't maybe normally be using Salesforce. So I think you'll see some some overlap and some expansion of user
[06:37] base. Into both of those offerings as time moves on. Well, I was just thinking it's a cool thing about Einstein Analytics supplementing Tableau those capabilities, the predictive capabilities. You know, Tableau is the bigger purchase, and we tend to think that you know Tableau is going to be supplementing what Salesforce already has. But it's nice to see it the other way around as well that Einstein already has these predictive capabilities and a lot of other research that's gone in. It, you know, and it's a it's a great point. I think you are going to see this will potentially open up. For Salesforce, maybe some user communities and some groups that they've not traditionally done as much with, right? In sort of the mid and back office, where Tableau has played a lot heavier, and. You know, the flip side of that is is that you know Salesforce is clearly the the number one CRM in the world, and I think that this will open up some opportunities for Tableau basically to to get some more exposure to that user community as well. So. I think it'll be benefit to both organizations as they come together. Super helpful to see that that context. I, for one, have spent almost no time with Einstein Analytics. You know, looking at that hands on, whereas that's been your life for years. Well, it's been part of your life for years, and it's just got so much depth. And my guess is that a large portion of the Salesforce community hasn't also hasn't spent a whole lot of time hands-on with Einstein Analytics or with the AI capabilities. Maybe outside of a couple of Trailhead modules,
[08:07] what are some things that you wish everybody? If you could just tell everybody in the Salesforce world some of the capabilities about what these things, what these tools can do, what would you let us know? Yeah. So I think you know. From the descriptive analytics perspective, so more of your traditional business intelligence and analytics perspective. If you have questions related to, you know, sort of sales forecasting or marketing or service, you know, analytics, there are some fantastic out of the box templates you can get started very quickly. It's a very fast time to value. In the sales, service, and marketing space with analytics, if you're already a Salesforce user, as well as from an experience perspective, surfacing insights and analytics where your Salesforce users already are. It Einstein plugs in very very well in that space. Tablo can do it as well. Einstein does it a little bit more eloquently. And then probably what most people get excited about is just the predictive or the machine learning capabilities. So it's one thing to make a prediction. There are tons of tools out there that can can help you do that. But what's really great about Einstein is that not only do you make a prediction. But you also get descriptive information around why that prediction is scored the way it is, and depending on how you configure it, you can get information on what actions you should take to
[09:37] improve the scoring. So, a suggested course of action. So. I think when you think of Einstein, I think key things you'd want to take away from are that really seamless embedded experience for your Salesforce users. Really quick time to value in terms of getting insights, and also predictive AI capability, right? That is designed to be set up and delivered with considerably less effort and with much more velocity than traditional AI machine learning offerings. So, those are probably the three key things I think you'd want people to take away from. And then probably the last piece, I guess, would be just because it is part of that Salesforce platform, it plugs seamlessly into the sort of the action framework. Within Salesforce, it plays well with all the other Salesforce components, and you can go from insights or making a discovery to action, like doing something about it in Salesforce. Very quickly, and it's it's a light lift to set that up. So those are those are sort of the key components of Einstein. I think it just prompts me to think you've you've used the word seamlessly and quick time to value and so forth. So you know a lot of the things in Salesforce are quite easy for people to do hands on. But if they brought a partner in like Atrium, somebody like Atrium to help, what are some of the common? Scenarios that you help with, yeah. So, I mean, as a Platinum Partner, we we kind of do the full range of service offerings as you'd expect, right? So,
[11:07] all the way from strategy and visioning to help. You know we we run workshops and can do sessions and strategy engagements around how to achieve the intelligent experience. Salesforce talks about that a lot. We'll talk about that a lot, but that really means that intelligent experience means. Bringing insights, whether it's through artificial intelligence or machine learning or analytics, bringing those insights into the point of action and within the flow of work where users already are, so that you can surface those insights within the Salesforce instance, where people can then action on them and drive business outcomes. Right. So it's all about. Driving business outcomes through intelligence. So we we offer those strategy services and vision mapping and sort of journey mapping, and then obviously we do the implementation services as well. So we can you know stand up and build dashboards, deploy Einstein, help with model development and deployment, and then you know the integration piece of it and configuration piece of it. So we can configure Salesforce apps. We do a lot of integration work, and both for analytics and machine learning, a common a common element, the really the fuel for all of it is data. And so, having your data in the right shape, in the right form, so that you can use it as inputs to models, or you know, use it to surface insights through analytics, that's actually a big part of a lot of the work we'll do is data shaping
[12:37] and data architecture. So, sort of all those things together are kind of where we focus our service offering. With the idea that the core idea of what we're trying to achieve is to, you know, drive better outcomes through intelligence. You know, it reminds me. I know you've done a ton of work with integration, database administration, data migration, all those kinds of things. So, I'm sure all of those things kind of come together. And that's that's part of your work and part of the part of the thing that you coach. I would imagine other folks at Atrium to be able to do handle those large migrations, the data shaping. Is that all? One skill kind of builds on on the next. Yeah, you know, it's sort of interesting. Like we've talked about this before with some of our internal team members. Because you need a little bit of a lot of different skills, right? You need to you need to think about you know you sort of need to think data. You need to think about data in a structured way. But the flip side of that is, you need to be able to communicate and talk business in terms of outcomes and what people are trying to measure and to achieve, right? From an analytics and an AI perspective. So it's really the blending of those two things, right? The The sort of harder technical skills around integration, data, data shaping, with maybe the the softer skills around business analysis and visualization and and sort of personas, journeys, and understanding sort
[14:07] of what businesses are trying to achieve. It's I think it's a combination of all of that. That really makes you effective, and obviously, we have people that specialize in one area more than other. There's still specialization, but what we do as a practice focuses on sort of that whole spectrum. So, it's a it's a fun space to work. It's a great space to work. So, very fortunate to do that. Nice. I can see how the the combination of those two, you know, using the technology knowledge to really solve specific business needs and make everything as simple and easy to understand as possible, would be. Would be compelling, yeah. It's you know, it's like most consultants, right? The the joy comes from taking people's business problems, things that they're trying to achieve, and figuring out technical solutions to make it so that their lives are better, they can work more effectively, produce more, do more with less. People can you know achieve better outcomes. So, I mean, the solution building is where where a lot of the joy comes in, but it is it is a fun place to work for sure, a fun space. So, are there industry trends where you see Salesforce either leading the way or following in the area of AI or analytics? Just kind of comparing what's going on with Salesforce with a broader industry. Yeah, you know something. I'll tell you that I'm excited about from just sort of my past experiences and some of some of what I've seen, both with Tableau's offering and the the core Salesforce offerings
[15:38] around natural language processing and being able to ask questions of your data. And have your data learn. So there's some great capabilities in both Tableau and Einstein product around that. And and I really think that over time, that's what's going to truly democratize analytics. Like we've we've been talking about this idea of analytics self serve for like two decades. Like it's been forever that we've been like, oh yeah, self serve analytics and democratization of analytics, and it's been a buzzword for decades now. But the reality is, is to get the right answers out, you have to know how to structure the question, and you have to have understanding of your data, right? So with traditional analytics, no matter how intuitive it is. There's room for error, and there's some challenges because some people they they simply they don't think in terms of data, they don't think in sort of an analytics mindset. But if you look at this natural language processing developments with the Ask Data capability and with the new Einstein capabilities. And how there is reinforced learning in terms of if you ask question and was this answer helpful, and it can learn over time. I really think that's going to drive some real democratization of analytics, some real self serve analytics capability, and I am excited to see where that goes. Thought you were going to talk about natural language processing about interpreting the the text in your documents, but you are actually saying ask the questions
[17:08] and just you know. That's right. What's the you know? What are the fastest growing segments in my business? Yeah, you say things like, "Hey, so show me my performance by region, month over month." And then, okay, looks like looks like we're down a little bit in Argentina. So, in Argentina, show me, you know, my breakout by product in terms of month over month performance. And you know, and just being able to ask those questions like that, just like you are talking to somebody, right? And I think it will even get further. I think you'll even be able to ask more abstract questions at future dates. But that sort of development, I think, is really powerful. The sentiment, sort of the sentiment analysis capability, again, which Einstein has now baked in as well. I don't know if you knew that, but as of the most recent release, I think it was the most recent release, and maybe beta, but I think is the most recent release. There is the now the capability to basically do some sentiment analysis even within your data flows for analytics. Take things like notes from your activities and basically say is there positive sentiment or negative sentiment in that, and then use that as an input into your model, your AI, or just use it to surface sort of insights through you know descriptive analytics. There's some cool capability around that as well. But I, yeah, I think I think what I'm really getting excited about is is some of this capability in natural language processing, because it starts to speak
[18:38] to experience and user interface, and what that will look like. I think it's really going to shape. And change the way people interact with their systems and what the interfaces look like in the future, and how people go about thinking about answering questions with analytics. I think I think there's a lot of really powerful positive change that can come out of that.
[19:04] It's um it's powerful for me to think about the the sort of contrast between your background where you've got a really solid experience as a database administrator migration integration like I mentioned analytics and on lots of different platforms you know really in the expert category compared to the you know. Just asking questions of your data and the democratization. So there's a sense in which I'm guessing that you would need both, right? You need some solid, skilled people at the core to make sure that the data is structured properly, and you've got all the integrations set up, and that Salesforce or whatever your tooling is provides you those mechanisms to ask the questions. But then it's But then the real power becomes when you're enabling everybody in your organization to ask those kinds of questions. Oh yeah, I think you're 100 right. Like, so I used to say all the time, you know, information is data organized and made useful, right? And so you're still going to have that need to organize and to shape and to transform data. That's there's going to be a need there, right? I just think the interface after you've done that, in terms of how you get the value and how you extract the insights out of that data after it's been organized and made useful. I think we're going to see huge advancements in that, and I think natural language processing is part of it. Okay, so rewinding back to the present by a few years, people are excited about artificial intelligence, machine learning, those kinds of things. Are there, let's say, somebody wanted to get started? Let's
[20:34] say somebody wanted to get the sort of basic knowledge, you know, start playing around, get some kind of better sense of how how this all fits together, and how they might use this for their for their organizations. What would you recommend? Areas to focus on first. Yeah, I mean, if you're starting from scratch, you're just coming out of you're just coming out of school, or you're you're brand new to the space. I would start with the descriptive analytics because sort of the discipline around how you think about data, how you think about asking questions, I think lays a foundation. For sort of the more advanced predictive and AI based capability, this is actually an area where I think Trailhead's way or Salesforce with their Trailhead. Is is way more advanced than most vendors, and the fact that you can access it for free, I love. Right, so there are some fantastic trailheads out there. If you just Google. You know Einstein Trailhead. It'll probably take you to the to the right link or pretty close to it. But there's some fantastic lessons where you set up demo environments. You learn by doing hands-on demos within the Salesforce Trail Trailhead org. So that's that's where I would encourage people to to really sort of. Sort of start, and then you know if you want to start getting in. Once you go through that and you do that, maybe build some dashboards. Get a little bit more familiar with with handling data.
[22:04] I think you know if you wanted to get into some more advanced topics, you know there's there's the potential to use things like Khan Academy or Udemy if you want to pay for the courses that will basically walk you through some basic statistical, machine learning, and AI concepts. Really, I've heard people say that really AI is just math at scale, right? Like machine learning is just math at scale. And so that the sort of foundational concepts around the statistics and the science of that that background, there's some good free courses out there as well. But to start the journey, I would say check out Trailhead because that's there's some great resources there, and they're well done, and they're not just lecture; they're they're hands-on exercises that get scored. So they've done a great job with that. Fantastic, so Dave. Huge, huge thanks! Really, a great pleasure to be here with Dave Dixon from Atrium AI, Platinum Partner for Salesforce Consulting. Thank you so much for being on Salesforce Way. Absolutely happy to do it, and Andrew, thanks for your time and thanks for the opportunity, man. Have a hope you have a great rest of the day and
[23:13] good weekend.
