Guests

Full Transcript

0:00:03 – Paul Risk
What I learned pretty early on is when anyone starts one of these initiatives, it’s sledge from the top down right. That’s CIO’s job, cto’s job, business leader’s job. Your job is to push the company in the right directions. So it’s a pushing motion into the business and the IT. At some point that needs to become a pulling motion. You need to hit an inflection point fairly soon. That’s the key to success and the key to adoption. If at the end of the implementation the leaders are still pushing, you probably sort of miss the mark.

0:00:43 – Fred Cadena
Hello and welcome to Banking on Disruption. I’m Fred Cadena. You found Big Show number four and I’m super excited for the conversation we’re bringing to you today with my friend, Paul Risk. Since the beginning of my Salesforce journey, i’ve been obsessed with driving business outcomes and I found most frequently, the key to realizing the business benefit that’s envisioned at the beginning of a digital transformation initiative is user adoption. Paul is an absolute master of driving user adoption by empowering his business users to learn the Salesforce platform and own the direction of their internal Salesforce roadmap. You’re going to want to grab your notebook for this one. Stay on after our conversation with Paul to hear Dan and I discuss some of the newly announced AI features in Salesforce, the pitfalls of using a lawyer who relies on chat, gpt and discovering the most popular brand in each state. While you’re listening to this podcast, why not take a moment to follow us on LinkedIn at the Banking on Disruption podcast, and on Instagram at at Banking on Disruption? Now sit back and strap in, because our show is coming to you right now.

On this episode, dan and I are excited to welcome Paul Risk, and I’m really excited that we have Paul with us today. His experience includes leading large multinational teams, building practice areas and organizations and creating intellectual capital. Now Paul brings extensive global experience across the UK, japan, argentina, brazil, singapore, mexico, australia and France. When he was a CTO of the warranty group, which is now part of Assurant, he led the implementation of a global Salesforce org to replace 30 legacy systems And now is a single global platform that is used in 20 countries in multiple languages and currencies. That platform, built entirely on Salesforce, pushed the technology boundary right to the edge where they interact with customers, allowing faster response to customer needs and market conditions. This was a complete digital transformation of the technology stack and, along with it, a wholesale change in corporate culture.

Prior to his pioneering work at the warranty group, paul was a CIO of AXA Assistance USA, where he also implemented an innovative Salesforce org. Paul’s also been the founder of Cloudway Partners, a boutique consultancy specializing in Salesforce strategy and transformation for insurance companies. He also has prior experience at Deloitte and Ernst Young. He’s a former professor of information systems at Loyola University, chicago. He’s a co-author of the PMOSIG, the program management office handbook, and frequently speaks on technology topics at major conferences such as Dreamforce, gartner, sim and PMI. Wow, paul, you’re certainly a very busy guy. I really appreciate you coming on the show. One of the reasons Dana and I invited you on today was to focus on some of the pioneering work that you did at the warranty group and some of these other places. We talk a lot on this podcast about transforming business using technology, but that transformation really can’t take place without bringing the business leaders and users along. I’d love to kick it off just by telling a little bit about how did you communicate that vision and the benefits of Salesforce to your stakeholders.

0:04:08 – Paul Risk
Hi, fred Dane, thank you for having me on. It’s a privilege to be here. It’s been quite a journey, a journey with a capital J. The first journey started at AXA, actually, let me back up a little bit here and tell you a little bit about that story. We had an old bespoke system that was written by one of our vendors in CRC++. It cost a fortune every month at licensing fees. I said I think this was back in 2011.

I called a friend of mine at a consulting firm that I’d known for many years and said, hey, i have this problem, i’m walking through it. I said, hey, this Salesforce thing that you’re always talking about, can it solve my problem? He said, yes, it can. I said, great, join a bid on it. He did, and we put it in. It was on time and on budget, which I don’t often say for a project. That’s always good news. But it went really well. The users loved it, the management loved it and I propelled my career in Salesforce and then went on to do great things.

But to answer your question, how did we communicate it? We did a lot of just basic blocking and tackling. We had information sessions, we had town hall meetings. I enlisted the aid of our marketing department, because being an IT professional, marketing and communications wasn’t exactly as strong as mine or my departments. So we said we have a marketing department. Maybe you guys can do us some favors. And they did. They helped us internally figure out how to communicate these things using emails and our internal website and our own communications mechanisms that were already in place. So leverage what we had in there to get that out there And, to be honest, it was challenging at first. It initiatives aren’t exactly the most exciting thing people want to hear about, but we started telling people that there was something in it for them personally and professionally And that really got the ball rolling.

0:05:58 – Fred Cadena
That’s phenomenal. I’m curious You mentioned a few channels. Did you find any channels particularly helpful or any channels that were particularly maybe surprisingly unhelpful in getting that message across?

0:06:13 – Paul Risk
Some of the people just didn’t want to hear the message People who felt their jobs were at risk or people who were resistant to change Other sellers an opportunity to help themselves. So we didn’t bring in any sort of professional change management. In retrospect That was an oversight On our part. We should have probably brought someone in, but we were really not looking at it as a big transformation. I need to just get the AS 400s out of the way, which was either I retire them or they’re going to retire me. That was sort of the imperative right there was, we had a technical problem we needed to solve. Then it morphed into a really larger transformation.

0:06:50 – Dane Grove
Very interesting. What about roles of champions? How did you identify and empower champions within each team to help drive adoption?

0:07:00 – Paul Risk
Yeah, that’s a great question. That was really the key to success. Right there You hit the nail on the head is I did not want this to be perceived as some sort of IT project, an IT experiment or an IT initiative. This was an enterprise initiative. First thing I did was align myself with the head of operations and the head of claims. The first thing I took is we’re going to go to Dreamforce together. guys, i’m going to show you all this stuff I’m talking about because early on, people thought I was absolutely crazy. He’s not going to work at Salesforce, at the sales tool, we have big claims and servicing system. Absolutely no way this is going to work. I said no, no, i did this once before. It’ll work just fine.

I took them to Dreamforce and the comment was from one of the guys is you are feeding a starving man a steak? How can I get two of these? Once they started to see what other companies could do at Salesforce, they believed and they said you know what? this isn’t that different than what we do on a day to day basis. I believe this can help. I had one on my right, one on my left and we never went anywhere alone If we needed to go to finance if we needed to present. Once I got them on board, we did this together. One of them actually joined me when I founded Cloudway Partners. We continued to do this good, innovative work because we could both speak different sides of it. I talked to the technology and the transformative effort and he would talk about how this helped operations in the business in the language that people needed to hear.

0:08:21 – Fred Cadena
I love all of that. I’m curious because I have a lot of the same conversations with clients and getting people to think of Salesforce outside of the sales or traditional CRM box. I love the approach of saying come with me, let’s go to Dreamforce and hear all these other stories. Would love it if you shared some other techniques, maybe that you found useful.

0:08:42 – Paul Risk
What I learned pretty early on is when anyone starts one of these initiatives, it’s sledge from the top down. That’s CIO’s job, cto’s job, business leaders job. Your job is to push the company in the right directions. It’s a pushing motion into the business and into IT. At some point that needs to become a pulling motion. You need to hit an inflection point fairly soon. That’s the key to success and the key to adoption. If at the end of the implementation, the leaders are still pushing, you’ve probably sort of missed the mark, you’ve done something wrong or you’ve missed an opportunity. When people hear about these initiatives, the first thing they say is well, what’s in this for me? Or what’s this going to do to me? Or what’s going to happen to my job? It’s about them. It’s about them being personalized, something you’re doing every day. When you miss the human aspect of it, that’s when you really sort of miss the opportunity to get adoption going and to get that pulling motion. I answered that question proactively What’s in it for you, for IT? I said I’m going to offer you a gift. We will train you in Salesforce 201 or 211, certification to 221. We will offer you all these certifications. We’ll pay for them. We’ll give you time to go study and pass your exams. Go back then. There weren’t thousands of certifications or probably like five or six. This was a gift. I will make sure that you are positioned in your career for the next level.

A lot of IT people said, wow, that’s unbelievable, we’ll take it. We weren’t thinking I’ll pay you 50 bucks or we’ll give you some sort of gamification. I said I’m going to propel your career. This is the trade-off. Stay with me for a couple of years, help me and the company be successful And eventually, when you leave because we all leave you will be able to double or triple your salary at some point. That’s my promise.

A lot of people took us up on that. For them, today’s workforce, sales force. So that shows you. And we’re in a huge IT department, so that shows you the power of that sort of a bargain. A couple people said no, as400 today, tomorrow, forever. They didn’t last long. Long live AS400. Right.

And some people said, hey, i do HTML5 and I like doing it. I said, and that’s okay, i’m not forcing you. Take your time, find your next path. If you want to stay, that’s fine. But HTML5 is not the future of this company. And so they eventually moved, but we offered them something very personal.

For the business people, it was the same thing. You’re not staring at an AS400 screen all day. You’re learning a new skill. That skill is transferable. Your job is going to be easier. All these ideas you have in your head. Not only will you be able to implement them, you’ll be able to implement them on your own schedule. You’ll be able to be a change agent in your own right. So that’s also transforming how the business operated right Running instead of running the IT every time you needed every little thing changed. You could do a lot of these things yourself in the business area. Yes, it still had to do their due diligence and their checking and testing and all that and the deployment, but IT was being pulled out of what is fundamentally a business process in a lot of cases. So for the business people, they were being offered not only new skill sets but a chance to make their own jobs a lot more interesting and a lot better.

0:11:57 – Dane Grove
Very interesting, Did you see? were there certain KPIs that you were measuring through the user adoption? You could see spikes here and there. Was it more just a vibe in the air? Was it a combination of things?

0:12:12 – Paul Risk
It was definitely a combination of things. Traditionally they try and measure things like how much time you spend on a phone and all that. It’s funny my now graduated. Back then, i think seventh grade daughter was going through science fair and she said, daddy, when you measure all these things, you have to control all the variables except the one you’re measuring. So that would stick in my head And I said you can’t control all these variables when you’re measuring things in a call center.

So let’s back up and look at the macro picture. What’s our throughput? How many calls are we answering? What is happening in the contact center? So at the end we look back and say We didn’t increase the number of people, but we increased the number of claims we’re processing by 50%. That’s a good measure. And so we tried to bring it up a couple levels and say, well, what’s going to mean something to the business leaders when they look at this in terms of what’s real and what’s measurable? Did I have to hire 100 temps or did I only have to hire 10? So we tried to look at macro KPIs versus micro KPIs And yes, there were other impacts, but you could sort of pull out which ones were being managed by the technology, versus which ones are being managed by normal factors. You could isolate those numbers.

0:13:19 – Dane Grove
Interesting. Did you go into that change process kind of knowing what KPIs you were going to measure early on for user adoption, or was that like on the fly, Yeah?

0:13:30 – Paul Risk
no, we had no idea. We were just trying to keep that as 400 from blowing up. And you know what we did is we started with the basics. We needed to replace the claims system, so we start with that, eventually moved on to virtually every other part of the organization. Eventually we got better at measuring those KPIs And I can tell you I’ll tell you a couple of success stories.

Way into the project, legal team came to us and said we need to put in a CLM solution, a contract life cycle management solution. By this point the Salesforce thing was pretty much built out And I said, sure, not a problem. So long story short, it costs like 20K and licensing 120K implementation. What you think about it is there’s nothing. You know when you, when you you’re looking at you know tens or hundreds of millions of dollars a company spend on these initiatives. This is nothing, mind. You was sitting on an already built $4 million, $40 million, steady the art Salesforce platform, but we spent 120K plus 40. It was done in eight to 10 weeks, so pretty fast. Those are all good measures, but here’s the key As soon as that system went live, they were able to reduce 20% of their departments So two out of 10 lawyers, savings of what?

six, seven, eight hundred thousand dollars. When you calculate that rate of return, it’s, it’s almost infinite. We say I spent 140 and 10 weeks and then, immediately after, i saved $700,000 in a continuing on $700,000. So that was a great measure. And there were other metrics like that. That sort of started popping up. The initial ones were a little more challenging because there was a lot. There was a lot bigger, it was a little more complicated, but as soon as you start putting things on the platform, the numbers really spoke for themselves.

0:15:08 – Dane Grove
That’s interesting. You know you could see some areas where business impact has to get kind of pulled through, but then in other areas where it clicks right away, there’s just immediate you know, immediate return. Is that a sales thing?

0:15:23 – Paul Risk
It’s. I tell you what it’s unique to Salesforce And I’ve worked with Oracle and PeopleSoft and a whole bunch of other Microsoft and a bunch of others. What’s unique about Salesforce makes it special is the business. People can do it without having to run to IT for everything. Right, when you think in Microsoft or Oracle, in a lot of cases somebody has to program it. You need a computer science degree or an engineering degree. You need to have a lot of technical skill And Salesforce you don’t need that. You can, you know. You can look up a couple of YouTube videos and figure out how to create a field or an object, modify them. And it’s super-july complicated. It’s made for the masses And that’s the beauty of Salesforce.

Anybody can pick it up and use it. And that was the sales pitch I was giving to the business leaders. Like you don’t have to have us do it. If you want to change the color on something, you can go do it. If you want to add a field, you don’t need to come to us.

And here’s, you know, dan, here’s a thing I think has really underestimated a lot of people today. That’s pretty good technical argument, you know. I mean kids are, you know, two or three years old, they’re able to pick up an iPhone or an iPad and figure out how to use it. They walk up to the TV and touch it and then they’ll understand why the TV is not a touchscreen. Right, because their phones and their iPads are. And when you think about it, anyone under the age of like 50 grew up with computers right? When I was 15, i taught myself on an Apple II Plus and an Apple II E’s. I was quite a while ago, obviously, but since then everybody’s grown up with computers, So they have a certain level of technical acumen. I think what businesses do, sometimes unintentionally, is they suppress that and say, well, that’s IT’s job. No, everyone in the company has a level of technical skill and I expect them to use. The technology should run through the company’s veins, like you know, like your bloodstream. It’s not something you go behind some locked doors and say, well, you’re a technical person, you know if they can all speak the same language of some technology which everyone is capable of, things work a lot faster, and that was really the key to success here.

We became a Salesforce shop very organically, and I tell you that the day this hit me is when I’m sitting in my office and I’m having a conversation with the attorney and she’s talking about well, if we do this with a license, then we don’t have to worry about this much objects and fields. And it’s like I’m having a conversation with someone on IT, but she’s a lawyer. I said, well, you’ve become jokingly said, you become one of us. But really what has happened is we’ve all adopted the same language. Right, it has this language. That’s sort of mysterious And we talk about things that nobody understands And sometimes that makes them feel dumb. But when you all speak the same language, people come together and Salesforce has that language that we can all speak And people in business aren’t intimidated, people in IT don’t make them feel stupid And we can all get our point across really fast. That’s the magic of Salesforce.

0:18:01 – Dane Grove
You know it’s interesting, recently had to pay a bill at Broward Health actually, and logged into the platform, went through the process I could tell it was something that was constructed more by, like purely by, it. Marketing was not involved. User experience, you know, like behaviors were not taken into account. So definitely I’m picking up what you’re putting down.

0:18:32 – Paul Risk
So let me share a story here that crystallizes really the whole adoption journey for a lot of the individuals here. And when we first started with our sessions, we’d have Salesforce come in for a week and do a private session for 18 or 20 people And I made it available to everybody. The IT people had to go, told them you got to go, you got to take the class, you got to pass the exam. Business people you’re invited. It’s up to your boss to decide how they want to manage that. So the admin assistant in claims came to me and said do you mind if I take this class? I’m sorry, speak at one of the Salesforce events. This looks really interesting. And I said it’s fine with me if it’s fine with your boss. And she came and took the course and passed her certification test.

Sometime later we had a bank that needed us to update a Salesforce community, or experienced clouds, as called today. So she took this on. Now, mind you, she’d been a secretary right Or drop the skip coffee and answer the phones and all that. But she took the class and she passed it And she decided to build a community using Community Builder And over the course of two months at work at night on weekends. She built this amazing claims site for this bank where you had 20 benefits on the back of your car rental, car insurance or travel assistance or all that. She built that site by herself, for free, right, because that was her job. Now, if you’d had to go to IT, i’m spending like four, five $600,000 to hire a system integrator. Come and do this for me And then maybe that’s a low price. But she did this. It looked beautiful, the client loved it And then the business took it and sort of white labeled it so they could pitch it to other banks And this just really gave her an enormous sense of confidence.

Salesforce did a blog about her and one of the other admins who did something similar And other people were saying, wow, look what she did. I mean, why can’t I do that? And I said you can and I will pay for it. And then that really sort of blossomed People all the admins were in the business areas. We didn’t have any admins in IT because we had people like her Her name is Michelle who took this upon themselves And now people said, well, i can bolster my career, so you only needed like one or two success stories like that.

Eventually. I had a dozen of them who were that of that caliber. But people would see that success and then they believe that, wow, the company invested in her And she’s doubles her salary. You know, this is very, very powerful to say, well, every company says, well, our people are our greatest resources. Fine, i put my money where our mouth was and I said I’m going to prove it that people are our resource, i’m going to invest in them. And HR was like well, aren’t you afraid they’ll leave? They leave, they leave, you know, but I bet you they won’t want to leave. And in fact, one left and a year later she came right back and I said I understand why you left, i understand why you went there and I understand why you want to come back. I’m going to give you a more interesting role. I’m going to let you flex your intellectual muscles a little bit more. As an investment, it really, really worked out.

0:21:30 – Fred Cadena
I love that story and I love the approach. It reminds me a lot. You know, similar to you, i came out of the industry. I’d leveraged Salesforce at options express.

The people that listened to the show have kind of heard the story in the past. I won’t go into like the details, but the way that we really succeeded with the platform at the company was to democratize it, and I don’t have as many success stories as you do of like people deciding to then transition from being a hardcore options person or a hardcore marketing person into being a Salesforce practitioner. But I definitely have a lot of stories about people coming to me and say you know what we’re doing this and we’re using Excel or we’re using access or we’re paying, you know, $10,000 a year for this point solution license for X. Can’t we do this on Salesforce? and oh, by the way, now it’s part of the 360 view, it’s not living in a silo on somebody’s desk anymore, it’s front and center with everything else. And I think that’s really what made me fall in love with the platform is how it really democratizes access to it in a way that you know other platforms before you know have been able to.

0:22:40 – Paul Risk
Yeah, that’s a great point. It’s for the masses, or what Salesforce calls citizen developers, but it really empowers people in the business areas. They have a lot of great ideas. It’s just that sometimes they’re not able to implement them. either People aren’t listening or you have to go to it, and the first thing I T says as well, you know where’s your budget and I need a project manager and some requirements will put on the calendar and we’ll get back to you in six months. Here they can vet them themselves. So that round trip from the customer request back to the customer can be measured in days or hours, not months or years.

0:23:12 – Fred Cadena
In that situation, do you see it role being enabler? So you know, making sure that there’s the proper infrastructure and governance in place, you know, whether that be through a DevOps process or a sandbox strategy, where you can give people the environment where they can develop, they can work on things themselves, but then it could be integrated back into kind of the greater whole in a way that’s not disruptive and in a way that has controls around it, or how do you approach that generally?

0:23:43 – Paul Risk
So obviously we had a lot of sandboxes that were reasonably good copies of the production environment and we would encourage the business try new things. Here are all these sandboxes, these environments. You can’t break anything. Go try it And if you get stuck, you know there’s people on it and there’s people in the business who can help you. One example was one of the people in operation said we need to create a page layout. Now all the fields already exist, but they’re designed more of like a point of sale. We need something for high speed data entry because people will still fax or mail us forms. We got to type it in. This is something to make sure I understand. All the fields you need are there. You just need to create the page line on the right order. She said, yes, As I recall, you took the 211 training and you passed it, Yeah, but you know I’m afraid It’s okay, Go get them shot, Go get it. You know she tried it and it was pretty awful. So you know we cleaned it up, we fixed it, but in production, while later she tried again, It was a little less awful And we cleaned it up, fixed it, put in production. The third time she needs something like this, she’d more or less gotten the hang of it. And then, you know, by now, by sometime later, she was maybe not at Michelle’s level of qualification, which was pretty good, And she had the confidence to do these things. So, once again, an IT process was business process, was pulled out of IT and handed back to the business. Now IT still has to go through, make sure it follows standards and procedures and you know coding, guidelines, naming conventions, all that stuff, And we would deploy it and IT was very much responsible for it.

I will tell you one little funny story. We had four or five groups around the world that were putting things into production And one day operations comes to me and says the apostrophes aren’t working. So what do you mean? apostrophes? Any dealership name with, like an O’Brien, an O’Hare, an O’Malley, anything with an apostrophe isn’t working. I was like, okay, I’m not going to ask how you figured that particular niche out, but okay. So I asked my team what did we deploy last night? Nothing Call the UK, Nothing Call Mexico. What did you guys deploy last night? Oh, we just had a little thing. Okay, How little. And he says well, it turns out that till the over the end, like in Pina or Nenia was causing issues with some of the dealership names, So we fixed it. I said, ah, Nenia didn’t get along with the apostrophe, So all right, So Nenia and O’Brien aren’t seeing each other. I died, We all laughed about it. They fixed it in five minutes and redeployed it.

The interesting thing here is that it allowed us to correct something in literally minutes instead of the hours or days would often take IT. So it made the businesses life a lot easier. They could put these things in production and it made IT’s life a lot easier and that we could make corrections really fast. It was to your point. It was very democratic, It was. It was very much a sort of like a almost like a family sandbox. Everyone could go in there and do their things And I would always remind you, there’s hundreds of people into the company who are not in IT, who are more than happy to help you.

Right, That’s CLM solution I talked about earlier. That was deployed by the admin and marketing because legal didn’t have anyone. But you know he had a lot of experience working with us. He was in marketing and he has 211 and he worked well with everybody. He said I’ll go around the world and deploy it for you. So business people were able to rely on each other instead of having to go to IT. So that was another really nice thing is that you saw a lot more cross coordination between departments and people who never normally never have a need to interact.

0:27:01 – Dane Grove
Well, very interesting. How about sustaining that user adoption over time, making sure it’s that habitual tool? Did you find that that just happened sort of organically, or were you guys putting programs in place et cetera to kind of fuel up?

0:27:18 – Paul Risk
There are a couple of different angles we took. First of all, we flat out said yes, ron was going to be unplugged. Actually, i’m personally going to go into the data center and plug the thing myself. But you know, one of the pros we’d had and I think a lot of companies have this is they go down one of these paths with older systems, old legacy technology. They never quite finished it. What we’ll get? that in phase two or phase three, and then you know, phase two sort of rolls around and they just build something else.

So we had multiple copies of this program and then its successors and the sons of its successors and the grandsons of its successors And you had all these systems which more or less are supposed to do the same thing, but because no one ever finished it. So he said we’re going to stop that. Everything goes into Salesforce. You’re not going to make copies of Salesforce, but we made it a point where, on a technical level, we’re going to get rid of these things. All those functions are going to go in.

One day we’re going to unplug it. If nobody screams, we’re done. If they complain, we’ll go back and figure out what functionality is missing. From a people perspective, nobody wanted to go back to the little systems because now their careers are personally very much tied to the Salesforce system. As we had invested in it very much had a vested personal professional interest to make sure this was successful. What I had done is I had aligned their professional goals with the company’s corporate goals And that was the key here. They wanted all of us to succeed because their own personal success was dependent on it.

0:28:40 – Fred Cadena
I love that, aligning personal goals with company goals and bringing everybody along for the ride You mentioned at the beginning. one thing that you potentially wish you had done differently is having more formal change management. Any other lessons learned? anything in hindsight? what would you do differently today?

0:28:57 – Paul Risk
You know, the technology is a lot better today than it was back then. In retrospect, i think the approach worked pretty well, maybe a little more structured. We had a roadmap. The roadmap was a little bumpy sometimes And they go out and spend more time getting the roadmap a little bit more refined. We have a business model underneath that helped quite a bit.

I think a better business model is always a good thing, and by business model I mean if I train this person here, how long before I can get some sort of return and what’s that How? what’s that measurable? But one thing we did do that was very successful is we took our roadmap, which was like a 15 by three foot sheet, and hung it outside of IT hung one in the boardroom, hung one in the business. And everyone would look at it and say this is our plan. And although they may not have understood the plan, they said clearly there’s a lot of thought put into this thing And we are clearly headed in a specific direction. And sales people would show it to our customers and we’d show it to our board members and say we are, we are going here and we’re going to get here, and that was a rule key to success.

0:29:53 – Fred Cadena
One more thing. This is kind of pivoting away from the conversation, but you did mention that the platform and the technology is a lot different today. We got to hang out a couple of weeks ago at World Tour. Obviously a lot of you know cool announcements And in general like what, what are you most excited for right now around Salesforce.

0:30:12 – Paul Risk
Well, i know, i know everyone talks about the AI thing And, as a little sidebar, i actually did my thesis in AI my master’s thesis 30 years ago. Yes, ai existed 30 years ago And the technology has improved dramatically. The mathematics behind it isn’t actually all that different, but the sensor technology and input technology is dramatically improved, so the mathematics has a lot more to chew on. I think that’s going to make a difference. I think it’s going to be a few years before you really start seeing it. I have customers who come to me and say well, i want to put in AI. And I said you know, let’s work on your NI first natural intelligence And then we’ll get to the artificial kind. Because you know you have to sort of stand, crawl, walk, run before you can do this. And companies say I want to go straight to AI. And I’m like you know, you’re still working on that little big wheel in your driveway And now you want to jump into the Ferrari. And I get it And I think the Ferrari will be wonderful for you.

But first of all, let’s clean up your business processes. Let’s make sure the people are well versed in what they’re trying to do. Let’s make sure they understand the business goals, right? I would ask my IT guys, what’s your job here? And I get a lot of answers. I said, no, we work for an insurance company. Your job is to sell insurance. Do not forget that. Yes, we do things that enable technology wise others to do this, but our job is to sell insurance, or whatever we sell.

So before you get into some of the really cool stuff like AI, which I think is going to really be a game changer over time, that has to go through maturation cycle too. Right, there’s going to be a little ecology. Right now, everyone has their own approach And, aside from writing kids college essays or high school essays, you need to make sure it’s going to come up with the results that you expected to, because, you know, is it going to make a medical decision? is going to make a sales decision? they’re going to do something that people one didn’t expect and didn’t want or didn’t expect and are pleasantly surprised with. So, as people move through their Salesforce maturation process, the AI will become a natural part of that. But let’s suffer, get fundamental blocking and tackling. Get people to buy into what the project’s doing, the transformation is doing and how that’s going to help your job.

One thing I think that is not talked about, or is talked about but only in the negative light, is it has got to take a lot of people’s jobs away. Well, you know how many blacksmiths do you know? Two horses you know? 100 years ago, the car took away those jobs, but other jobs sprang up in their place. You know of all the jobs that exist today. You know the top 10 didn’t exist 30 years ago And I think you’re going to start seeing more and more of that’s a constant evolution. You’re going to see more and more of that happening. Anyway, i think that’s the. That’s some of the more more exciting things. Salesforce is continuing to verticalize itself as they mature as an organization, and you’re going to start seeing more and more intelligent, robust solutions in particular industries as well, some AI based, some not AI based.

0:33:09 – Fred Cadena
I love that And I’ve been having a lot of very similar conversations with my clients that I’ve been learning more about. You know Salesforce’s approach to generative AI and LLMs in particular, And I do think they’ve got the right approach in this bring your own model approach, rather than building yet another LLM, letting their customers, you know, plug in some commercial LLMs through an API or something that they’ve developed internally. I think directionally, that’s that’s the way things are going to go more smaller, specialized model that are built to purpose, And I think that’ll help eliminate some of that kind of black box and uncertainty stuff you were talking about, Paul.

0:33:51 – Paul Risk
Yeah, and that’s that’s a good point. Each of those specialization my specialization is insurance and Salesforce that intersection. A lot of people are going to be experts in e-commerce or something else. So the AI needs to be tuned and it needs to learn right.

My thesis when I wrote it many years ago I said it was it was reactive, adaptive and then predictive, which meant that the system would first of all react what’s the right answer mathematically, then it would adapt, it would add more data to its database so it can make better decisions and that would start predicting based on those decisions.

Today I would add one more to that. I would say new systems that are actually very perceptive, because the amount of data and the amount of sensor technology allows the system to really make a decision similar to what a human would make. You can, you can put in emotional inputs, you can make feelings, and then the system can decide on upon those parameters. So I think you’re going to see a little bit more of that kind of AI, but it needs to be very tuned to a particular industry or task. You can’t just say my AI will do everything and answer every question on the on the other planet won’t? they’ll have one that’s tuned for it Insurance and particular kinds of insurance. They’ll have another one that’s tuned for healthcare and very specific kinds of healthcare. They’ll have another one tuned for the automotive industry and maybe only a niche to that. So you’re going to see solutions that are really tuned to make decisions, or help humans make decisions, for very specific industry and subsectors.

0:35:12 – Dane Grove
You know I’m curious do you feel that AI can bring some stability to economies, to life as we know it and you know, maybe we’re all a little less frantic and kind of not as stressed out as a result Like big picture, is something like that possible in your mind?

0:35:32 – Paul Risk
I’m an optimist and I love Star Trek, so I think it’s a step in the right direction, i think, until we invent like replicators and some of the future technology that literally eliminates the need for work. That’s got a bit of a journey still, but I think what it will do is allow people to focus on bigger picture problems.

0:35:50 – Dane Grove
For sure.

0:35:51 – Fred Cadena
If you’re a Star Trek fan, this is a renaissance. I’m looking forward to weeks till the next season of Strange New Worlds drops. Well, thank you, paul. Really appreciate the time before we let you go. If our listeners want to connect with you, what’s the best way to get you?

0:36:05 – Paul Risk
Thanks for it. They can find me on LinkedIn, Paul Risk. I’m easy to find Type in my name. I’ll probably pop up. Send me a message. Happy to connect.

0:36:14 – Fred Cadena
Fantastic. Thanks again. And now we’re back with Quick Takes. Dane, what do you have for us first this week?

0:36:28 – Dane Grove
Recently announced a series of AI driven tools, for example, ai generated product description. Salesforce is releasing a new feature that uses generative AI to automatically suggest SEO optimized product descriptions for your e-commerce storefront, or you know, and or using commerce cloud, all to ease the process of managing and building product catalogs. Ai powered features and service cloud, including agent assist, case routing automation, chat bots. And then, finally, ai image generation tool for marketers. This is a tool that can generate on brand stock photos, social media graphics and other visuals.

0:37:25 – Fred Cadena
What’s your take? My first blush is that one of these things is not like the other. Kind of go back to the old Sesame Street song, right, When I look at the announcement around commerce cloud and the marketing image generation tool. Both of those are generative. I know there’s some generative things coming in service cloud, but the announcement highlighted agent assist using NLP and case routing using machine learning, neither of which are generative. So I think that’s that’s interesting. There’s a recognition that a generative AI still needs a little bit of a guardrail before it can just go straight out into the public. You know the commerce cloud feature and the image generation feature. Both talk about marketers taking the output and reviewing and editing it before it goes live. I think there might be some concerns in the contact center, based on, you know, the seniority of the employees and the volume of customer inquiries, that that same vetting wouldn’t necessarily happen. But what do you think?

0:38:37 – Dane Grove
Well, i love the fact that Salesforce continues to push for betters of this experiences. I can’t tell you how often, across a broad number of products in my in my day to day, that I have such inconsistent experiences from a service standpoint. You know, we have, like, intelligent appliances in our kitchen. We’re running a pretty sophisticated stack inside work stream. There’s, you know, apple based, windows based technology in the house. There’s game systems, and then sometimes I get on the phone with someone from service and I’m like okay, you are an outlier, like you’re either amazing outlier or you are horrible outlier, and so I love the fact that they are continuing in that direction. The piece that’s also interesting is the generative AI component. With that, like, how is that going to work for? like, how are we going to, how, how can we start generating some of that, that content, that perspective for service agents to leverage and and feel safe and secure that they’re giving accurate information, good information. When does that happen?

0:39:54 – Fred Cadena
I think the capabilities exist to do that. Today. The concern is a couplefold. One is the generation of service responses is very specific to a customer situation, specific to the policies of that company. I can ask for AI image generation and that that image may not differ significantly between two competitors. You know, if I’m, if I’m Walmart and I’m Amazon and I need an image for a product description to go in an email, you know that image doesn’t necessarily need to differ broadly. If I’m producing an answer to a customer service request, it has to be right, it has to be something that the company can stand behind.

And I think that’s a little bit more difficult. And where you get there to kind of get back to your point is it’s in classifying what good data looks like. So it’s not just collecting responses from agents, it’s classifying those into which are the, which are the best ones, which are the gold standards. So to your point earlier, when you have these service discussions and it’s either an outlier on the really good side or the really not good side, you want to make sure that you’re feeding those outliers on the really good side into the model and somehow excluding the not really good ones.

I think the same thing I see quite frequently on the knowledge base side, and knowledge bases are difficult to maintain. You know that’s always been. The most difficult thing is to keep those articles updated. Once again, i think it’s a story that rhymes with so many other stories in this AI world. But getting the data identifying what is the authoritative source, what you want those suggestions to be based on is the bridge that we need to get there. I think the technology to generate the responses is there. It’s just the pool of data that it’s going to pull from is usually not available.

0:41:58 – Dane Grove
Really great points like switching back to the AI generated product descriptions.

I think that’s actually a pretty cool feature.

I’m assuming that what they’re talking about is using business manager and managing that product catalog, and I’ve had some exposure to supporting some of that talent demand And that’s an interesting blitz, because what organizations are often looking for is someone that has marketing skills, understands the target audience and how to engage that target audience from a content standpoint and not just written content, but also as it relates to images and then has the ability to you know those web development skills, i would imagine, to you know to kind of position all of that well within an e-commerce storefront, like use of white space, maybe wear some of those clickable links, call them, call the actions or buy now kind of stuff happens, just being able to come up with the content itself, like be a good writer, right, and maybe some of that goes back to marketing, but I think that’s actually a super helpful feature, one of those things that probably just helps to kind of solve that blank page problem, like I think.

You know, i think about my own experiences with AI and where I really appreciate it. It’s not producing a final product, it’s just helping me solve that blank page problem, and I think this is a tool that kind of falls into that category.

0:43:39 – Fred Cadena
The product itself. I think is really cool in listening to you talking And you know I’m not, you know, a commerce cloud expert by far, but when I hear you talk about all of the steps that are involved, i think it can be a real time saver. So there’ve been a couple of times when I’ve helped out on setting up e-commerce businesses and I just get forward. You know, as a user on the e-commerce side, i don’t give a lot of thought to. You know really how many things are in an average product catalog. But when you’re sitting there trying to put an e-commerce store together and you’re looking at you know 50 products, 100 products, a thousand products, you know, and certainly many multiples of that. If you’re a big, you know e-commerce brand. And then you have to come up with two or three interesting, vibrant, engaging paragraphs that are going to make people want to click And oh, by the way, they also have to be SEO optimized to even drive people to the page. It’s no easy lift And I think anything that can help shortcut that process build in those best practices.

I imagine that part of this is some kind of a uniqueness engine or it does some kind of analysis. I know this is probably my biggest problem in writing is if I’ve got to write a very similar thing 20 times. It ends up sounding very much the same 20 times, right, and hopefully the machine can just help with some of that word choice and making each of the product descriptions sound sufficiently unique. So I think it’s really cool. You know the way out of my depth from a usage perspective, and the last point I’ll make on it is it also reinforces.

To go back to the conversation we had earlier with Paul, and we’ve mentioned a couple of times this idea of a much smaller, more specialized generative AI model. Right, this is a model that was purpose built to generate product descriptions. I am sure it was trained on a data set of really great product descriptions, right, Like here’s an image, here’s stats about the product, here’s a description, and it’s a model that is not going to go right a Seinfeld episode in the cell of William Shakespeare but it is going to generate really good product descriptions over and over again. So that’s really cool too.

0:46:06 – Dane Grove
Yeah, yeah, that’s a really good point. It is clearly a specialized model, the image generation tool for marketers. have you played with many of those types of? you know AI, like generative AI tools?

0:46:21 – Fred Cadena
I have tried a couple. The first thing I learned is that I am not nearly skilled enough in prompt generation for these image generators. And there’s a person I follow I’ll drop it in the show notes for those of you that are interested but she does a daily post on LinkedIn where she posts the image that is generated by an image generator and then she posts the prompt that she used to create it And it’s always, you know, basically a paragraph long prompt And then she kind of critiques the output And if I’ve come to look forward to reading it on a daily basis, you know some of the images are better than others. I definitely see the images getting better over time. But the part that really floors me is what goes into that description. You know I’ll go to the image generator and I’m saying you know, give me, you know, a house in a field and you know it’ll generate a house in a field. But when you really want, you know to generate that image and you have that kind of artistic eye and that artistic vision for what you want to see, you know those prompts can get really, really detailed. So that’s the first thing that comes to mind.

The second thing, and kind of more of a business application. Ironically, i was having a conversation a couple of weeks ago with a friend of mine. She actually works for Adobe and she was at New York World Tour And Adobe’s got a similar product that they’ve put out And when I was talking to her the first thing that kind of connected in my mind was this would be great combined with a data cloud where you can get that hyper personalized view of your customer And then you use those insights when you’re generating an image for an email to generate an image that is hyper personalized to that individual. So I think we’re going to see a lot move in that hyper personalization not just the right image, not just the right moment, but also the right image that’s going to evoke the emotions, to get somebody to take action. So that I find really exciting.

0:48:29 – Dane Grove
It’s one of the things that I think I’m really enjoying about working with these different tools is you know how do you, how do you get the right information out of them? What are those prompts look like? What does that feedback look like? How do you ask the questions? It’s actually fun. It does require patience and lots of experimentation, but it brings up some really great points.

0:48:52 – Fred Cadena
Yeah, it is definitely fun. I enjoy it as well. I have gotten much more skilled in leveraging, you know, tools like Claude or chat, gpt and prompt engineering around that. My challenge so far on some of the image generation side is just a lack of that, you know, artistic vision and really coming up with those rich descriptions of what I want to see, but definitely something I’m going to play around with a little bit more. I wanted to bring something back up to you, going back to some of what we were talking about earlier making sure that what comes out of generative AI has been reviewed for factual correctness before putting it out into the wild.

I came across an article this week about an attorney that was representing a man who was suing an airline for some injury that happened during a flight, who used chat GPT to prepare his court filing.

The airline had filed a motion to dismiss the suit and the person’s attorney, you know, filed a response and that response was a 10 page brief and it cited half a dozen other court cases that you know spoke to why the case should be allowed to continue.

The problem was that after the brief was filed, neither the judge nor the lawyers for the airline could find any of these references, and the reason they couldn’t find them is none of the references were real. You know, chat GPT had basically completely invented these court cases, and so now the lawyer is in hot water. He’s facing a disciplinary hearing, basically for filing a false brief, and now obviously, this case is at risk, and I’m not a lawyer, i don’t really pretend to be one, but I wouldn’t be surprised if the person that was suing, that relied on this lawyer’s work, might have a case from al-practice because of the way this filing was put together. So I think, just again, as we talk about how AI is really having a lot of positive impact, this still goes back to, you know, some of the limitations of how large language models work, how generative AI works, and just making sure that we’re all cautious before we take something that comes out of one of these models and just put it out there.

0:51:13 – Dane Grove
As you know, i’m really fond of Claude, so that’s for those of you who have not heard of Claude that’s a model that’s produced by Anthropic I think is the correct way to pronounce that name. They are backed by Salesforce Ventures and they have a different approach to delivering as an LLM, receiving feedback, et cetera, and I asked Claude today for some perspective on something that actually refused to provide. I think this is because Claude has learned that I am always looking for information that includes like citations and references, and it gave reasons for not providing the information I was asking for, for example, high potential for inaccuracy and wild guesses, risk of implicit biases, unsupported predictions that could mislead or confuse people. I was really impressed with the platform’s position on refusing to provide the information I was asking for and then given some really good reasons as to why. But we didn’t, you know, when I first started using Claude, i, you know, and I don’t know if this is the platform getting better or if it’s just my interactions and constantly asking for citations, references et cetera But I was actually really impressed with that Boy.

I’d love to have been a fly on the wall in that courtroom When that old thing unfolded. That also must have been just a really, really embarrassing moment for that attorney And my heart goes out for that, you know, to that person. But and big, big mistake.

0:53:05 – Fred Cadena
Yeah, totally. I mean I’m not trying to have any fun at his expense. I would be mortified. I’m sure anybody really would be.

We talk a lot about our kind of fun use cases, but when the stakes are really high, you know, making sure you’ve got that second level of verification I’m not saying there wouldn’t be a use for generative AI in something like this, but making sure that you have some kind of process in place where you know, maybe a paralegal reviews everything before you just go and file it with the court would make some sense.

I will say, dan, i’ve learned a lot from you in how you’ve trained Claude and how you’ve started going through your prompt engineering. One of the things that’s really powerful about Claude, or at least how it’s implemented, is the fact that it is able to go out and get fresh data, and I think that you know between that, between asking it to either provide the source citation or just asking it to limit the search to. You know, only verifiable data, i think really helps increase the veracity of the responses. So I think it’s a little bit of an umbrella in galoshes kind of a situation. Right, you want to probably do a little bit of both. You know both make sure the prompt engineering is better as well, as you know, just using verification, especially when it’s a super critical response like a court filing.

0:54:36 – Dane Grove
So, Fred, I, you know, I appreciate that feedback and, hey, I don’t think any of us are experts right now. I’m just trying to be sort of open minded and experimental with what you’re describing as prompt engineering or just trying to figure it out how to get the right information, the information I’m looking for and accurate information from these different tools. But you’ve also been really helpful in my journey, reminding me that these tools are not to, you know, produce a finished product, but simply to help address that blank page problem and get us to a better starting point, a more mature starting point, if that, you know, if that kind of sounds like an oxymoron to say mature and starting at the same time. But I think you know what I’m getting at And I, you know, I’m just appreciating that feedback and that guidance from you. that’s been super helpful.

0:55:30 – Fred Cadena
I appreciate it. It’s definitely a learning journey And the last thing I might leave, you know, our listeners with is like, as you’re exploring AI, find you know a trusted person, group of professionals that you can bounce ideas off of. I mean, it’s been invaluable as I’ve been dipping my toe more and more into the huge gains that have been made in gender of AI in the last few months. I wanted to ask one more thing before we wrap up for the episode Dane, and that is a little bit more fun topic. I did notice that on deck released its annual study of each state’s most loved local brand And I thought it was interesting. I know I said the map to you. I don’t have a chance to look at it, but just an interesting collection of brands across the country. I think it tells a few different stories, but I wanted to get your take first.

0:56:23 – Dane Grove
First of all, it’s a fun graphic. It’s a fun infographic to Rick U And I had to read a couple times and really make sure I was clearly understanding the information. So they analyzed two million tweets about brands from each state using an AI sentiment analysis tool, and so I guess that I take it to not just point to modern brands that are most popular today, but maybe also reminiscing on historic brands, prior brands that we miss. And yeah, it was pretty fun. For example, colorado Backflip Studios is their most popular brand. And then how about South Carolina with Duke’s mayonnaise, you know? so what were your takeaways from that infographic?

0:57:28 – Fred Cadena
I think along the same lines. I mean, i know a lot of Southerners that are very particular about their mayonnaise bread and Duke’s. Definitely, as a share of fans, i’ll say this I like how they spelled out the methodology. First, i have a lot of kind of questions about whether or not that’s right or what bias that might introduce. So, like one thing that I was thinking about Illinois, where I spent 16 years in Chicago, that is Grubhub And I’m sure you know, in Chicago especially, you know pandemic. Till now there’ve been a lot of Grubhub users but probably not a whole lot more than some of the other really populists and urban environments like New York City or the LA area et cetera. So why would Grubhub bubble to the top? And, as you were kind of going through giving your reflections, one of the things I thought about, especially when you mentioned the more nostalgia is the other potential bias is you know employees of these companies. So there’s still a lot of employees of Grubhub in Chicago, a lot of people that probably know employees of Grubhub in Chicago. So you know, how has the methodology of you know looking at those tweets, looking at things that have the official hashtag, how have they introduced a little bit of bias into those results.

I recently, as you know, moved to Omaha in Sofran, nebraska, it’s lit up Omaha Steaks And I can tell you and I’m not trying to throw any shade on any brand but nobody here really talks much about Omaha Steaks. Like in the year that I’ve been here I don’t hear a whole lot of conversation about Omaha Steaks. But I’m sure there’s a lot of marketers that live in Nebraska that are probably tweeting a lot of things about Omaha Steaks. So that was kind of one of my first reactions as well.

0:59:30 – Dane Grove
Yeah, that’s right, Definitely, Let’s see. There’s Newman’s own, there’s Big Y LL Bean. How about Cheer Wine in North Carolina? Have not heard of that. And here’s a fun one.

0:59:43 – Fred Cadena
I’ve not heard of that either, but I like Cheer and I like Wine, so I’m not gonna get that, yeah.

0:59:49 – Dane Grove
and then you’ve got Mississippi Mossy Oak, which I take it to be like an outdoor brand. I’m actually familiar with Mossy Oak. I think they make some great stuff. And then another one of my favorite brands is very popular in New York and that is Calvin Klein. So that’s kind of interesting, right Calvin Klein New York and then Mossy Oak Mississippi.

1:00:13 – Fred Cadena
It checks out. I definitely would associate Calvin Klein much more with New York and Mossy Oak much more with Mississippi. I think it’s interesting. I love the diversity of brands that I see here. I noticed there were two guitar companies on the map, quite a few clothing brands, some stuff that is very regional, like Wynco Foods up in Idaho, vaidmar in Iowa again, a lot of regional differences and regional brands, which I also thought was interesting.

1:00:48 – Dane Grove
Yeah, and in Florida. Chewy is headquartered here And we love our dogs in Florida. They’re everywhere, They go to restaurants and grocery stores with us, et cetera. But I have a feeling that some of that is also a result of just Chewy’s marketing, that’s. And then there’s Pizza Inn in Texas Yeah, That would surprise me.

1:01:17 – Fred Cadena
I grew up in Texas and I remembered the existence of Pizza Inn. In fact I remember as a child probably fifth or sixth grade getting a phone not a mobile phone, but one of the things you used to plug into the wall. That was a molded plastic Pizza Inn man with the pizza up in the air. But I don’t remember a lot of Pizza Inn kind of outside of that. I was surprised to see that that brand is apparently much bigger in Texas than I thought it was.

1:01:51 – Dane Grove
Yeah, yeah. And then just above it in Oklahoma is Orangely Frozen Yogurt. It’s super interesting infographic. Thanks for finding it, fred. It’s been fun to kind of poke through that and look at some of the different brands.

1:02:05 – Fred Cadena
I’ll put you on the spot a little bit. Looking at just this image of 50 brands. Do you have a particular favorite in the collection?

1:02:14 – Dane Grove
That is a good question, i will say before answering. One of the ones I’m curious to learn more about is the Midnight Sun Brewing Company out of Alaska. I’m curious about that brand. I’d like to know more about the beers that they’re brewing. Which is a favorite on my wife’s behalf Calvin Klein. I’m always looking for Calvin Klein stuff from my wife. I know her sizes and I keep an eye out for sales Gosh my own. I struggle to answer that question. I’m not sure There’s several of these brands I like, yeah.

1:02:57 – Fred Cadena
I think I’m in the same boat. There’s definitely a lot of brands on here I like and I use, but none of these would hit my top like my personal top 50 list of brands, and I think that’s interesting as well. I don’t know if that says more about me being an outlier from the public or more about how this study was put together.

1:03:17 – Dane Grove
Yeah, exactly. Yeah, i think I feel pretty similar. Well, maybe we should drop this link into this episode, where people can check out the most loved consumer brands by every state and fund a check out.

1:03:37 – Fred Cadena
Absolutely, and I will definitely make sure this gets included in the show notes. Appreciate your time and opinions, as always, dane, and we’ll talk to you again in about a week.

1:03:47 – Dane Grove
Ditto Fred, Have an amazing day.

1:03:50 – Fred Cadena
You too. Well, everyone, we hope you enjoyed episode four of Banking on Disruption. I can’t thank Paul enough for sharing his experiences and driving user adoption through empowerment. We have a lot of exciting stuff planned for upcoming episodes, but, most importantly, we want to hear from you, dane, and I would love to hear your thoughts, feedback and ideas for new episodes. Why not drop us a line? New episodes drop every other Thursday, but in the meantime, you can visit our website at bankingondisruptioncom for show notes, including a full transcript of today’s show. Also, if you liked what you hear today, please subscribe to the podcast and leave us a review. And finally, we would love it if you followed us on LinkedIn and Instagram at at bankingondisruption. Until next time, this is Fred Kadena, wishing you success in your digital pursuits.