Latest / Talk Commerce / iPaaS Integration and Why AI Alone Cannot Run Your Business with Robert Rand of iPaaS.com
Transcript
- Brent Peterson (00:03.02) Welcome to this episode of Talk Commerce. Today I have the famous Robert Rand from iPass. Robert, go ahead, do an introduction for yourself. Tell us your day-to-day role and one of your passions. Robert Rand (00:15.852) Well. Robert. I head up partnerships and alliances at ipass.com integration platform as a service. And I've always been passionate about commerce. I don't know why it keeps me on my toes, I guess. I've always liked helping, you know, businesses deal with problems that in many cases, they didn't even know that they had because you're just so busy dealing with day to day, right. But no, I I have kids, I don't have many hobbies anymore. Brent Peterson (00:50.222) Yeah, that'll do it. I have kids too, but they're older. anyways. All right. So Robert, before we're going to talk a little bit about the idea of AI and how much it can help an entrepreneur. So, but before we do that, I have to tell you a joke. It's a free joke project. Give me a rating eight through 13. So here we go. I went to the doctor with hearing problems. He said, can you describe the symptoms? Robert Rand (01:15.264) What? Brent Peterson (01:16.526) Let me do this again. Yeah, I screwed it up already. I went to the doctor with hearing problems. He said, can you describe the symptoms? I said, Homer's a fat dude and Marge has blue hair. Yeah, I blew it in the delivery, but it is what it is. Robert Rand (01:40.072) Every joke from you has always been near and dear to my heart. Brent Peterson (01:43.488) All right, good. Yeah, I appreciate that. Let's start off with, do you feel like AI can replace joke tellers? Robert Rand (01:55.414) Honestly, I can write some good jokes. I've seen it. I was at a conference or an event a few months ago, and one of the speakers asked everybody to pull out their phone and open ChatGPT. And of course, this is a room of techies, and everybody did. And he said, OK, ask ChatGPT to roast you based on all of the information that you've shared, all the questions that you've asked to date. Brent Peterson (01:58.572) Really? Robert Rand (02:23.891) And it was hilarious. Yeah, well, it knows you better than you think. So maybe the audience will take that as a challenge. Probably not the roast that you want to share with other people. Brent Peterson (02:26.424) Wow, I'm impressed. Brent Peterson (02:39.188) Right, yeah, my experience so far with JETCBT is it's terrible at telling jokes. But then again, I'm asking for terrible dad jokes. But that does lead us to kind of our topic today. And it's how much a single entrepreneur can take on on their own while employing AI. And we had a great chat in the green room. You know, the idea is that that AI is a great supplement for what people are doing, but can you as a business owner at some point in your life, run your entire business by yourself at scale with a bunch of AI agents? Robert Rand (03:21.407) Yeah. Brent Peterson (03:22.892) That's the question at hand. Robert Rand (03:24.797) It's a big question. And I think, you know, the short answer, whether everyone wants to hear it or not is no. That you're not going to scale very well that way. And I could describe it in a few different ways, which you'll probably have me do, but, you know, what's coming to mind right now, know, because AI... while it brings so much to the table, it doesn't replace the knowledge and experience of real people. And in many ways, it's like the difference between a doctor and somebody that stayed up all night reading WebMD. And we know that those are two different things. And so you can get a lot of information, a lot of things done by sticking AI on something. But You can also wind up with, I think in the green room I used the phrase dumpster fire. Brent Peterson (04:19.362) Yeah. You know, I'm, I'm a big cloud code user now and I've been using it for about 11 months and now have developed quite a bit of training. you know, you have context and runs out of content. So you always, you have to keep it up to date with what you're doing. Right. And I know there's that you can create your own GPTs as well. So I would say that, you know, one of the things that it will do if you're a business owner and you want AI to go off and do all the things that your old team used to do. It may, yes, use your training, but it also may know it just does something because it thinks that it'll be better. in the coding world, it'll have four different solutions for the exact same problem in the exact same day, in the exact same hour. So, you know, it's never consistent, even though you are giving it that training data. In your world then, you know, as you're employing more and more AI to help the middleware that Ipass covers and I know that there's more AI going into that. What are you seeing as risks in getting too much AI in? Robert Rand (05:27.528) So back to my last analogy, I think we are seeing people that are trying to just have AI right integrations carte blanche. And there might be reasonable use cases of that. But by and large, when we're dealing with bigger data sets with multiple systems that need to be integrated, I would describe it as you found somebody on Fiverr to write a bunch of code for you. You don't know if it's good or not. And it might get you through a little bit, but what happens when you need to adjust it? And what happens when there are some data failures or discrepancies and data gets stuck or doesn't go where it's supposed to go? You're kind of up a creek without a paddle. That it's not very sustainable. And I find that that... you know, that's analogous to other areas of utilizing AI, that it's only as good as the information that you're putting in, know, trash in, trash out. And that you have to be strategic about how you're using it to supplement versus trying to, if you allow it, it'll do whatever you want. Won't necessarily do it all that efficiently or effectively for the long run. So in our world, what we've been doing as an integration platform, that sits kind of in the middle of all these different software systems that a business might use between their e-commerce and ERP and point of sale and CRM and product information management and marketing and on and on. It's being able to allow tools in there where we're already no code to low code, but that low code piece, for instance, being able to query and say, hey, look, you I want to transform the data in this way before it gets from system A to system B. Can you write me something for that and be able to speed up that process so that... Robert Rand (07:27.881) You're not trying to write code that you're less equipped to do or things like that. Or for those that are equipped to do it, just speeding up the process. Less thought in it and they can review and see, yeah, yeah, that looks good. And so finding areas to further lessen the burden on the people involved and just let the machines do more of the work. Brent Peterson (07:51.886) Yeah, I think one thing that we had talked about is there's a degree of expert knowledge that a user has, and you typically go with a partner to help you with that, right? Like, let's just say you're launching a new store and you need to integrate your ERP with your e-commerce platform. You, as a business owner, could just go to chat.gbt and say, hey, how do I do these things? And I don't need a partner, right? But there's a... A solution partner who's going to help you with that has that built-in knowledge, plus they have a network of knowledge they can bring in that is not just about context that maybe AI could answer, but it's also about real-world problem solving that a contextual model can't quite understand. And maybe in, I was going to say maybe in a year, but maybe in six months. Who knows, right? It's going so fast. Robert Rand (08:50.081) Right. Brent Peterson (08:51.244) I think there's still a degree of humanness that has to be in there because you can't just put it on autopilot and go. There are certain tasks that AI is good for, but there's also certain tasks that AI tends to really screw up. Robert Rand (09:07.53) And when you're thinking about data, for instance, you when we're talking about the accounting data from your e-commerce or other sales channels into your ERP, that data is sacrosanct, right? You've got to pay the right taxes on it. You've got to, there are things happening that can't, can't be left to. to AI hallucinations or any other form of discrepancy there that really need to be right. And so that's where systems like ours are purpose-built to make sure that there's accountability for the data, that everything is correct, that if something fails, it's going to automatically retry on a schedule, that you're going to get notified and have a bit of a root cause analysis at your fingertips, and that if you need to make adjustments, that you can do that in a logical way where you can see what's happening. and you can pick up where the last person left off. And it's not just a bunch of random code. And you can turn on new integrations to the hub and turn off old ones and not go back to the drawing board on all of Brent, what you were describing, this internal knowledge that's often lost in the mix because a lot of these projects, it's not just about... getting data, you know, from a point to another point, it will often have to do with diving deep into the business operations and rules, how does this business decide which warehouse is gonna fulfill or which retail store is gonna fulfill this order? And what kind of reserves of inventory do they need to have because they sell in multiple channels and a million other things. And that's where, so we have a network of about 80 partners that are agencies and SIs and VARs and MSPs and consultancies Robert Rand (11:00.356) and other service providers, and that network keeps growing. And often what we'll see is a bit of a hybrid, right? Where you'll tap somebody like that that can bring some project management in and make sure the right questions are being asked and that everybody knows what success looks like at the end of the project. And then the merchant or other business that's implementing the integrations may help with some of the manual effort that's gonna go into the testing and different things that are happening. And so there's a breakdown. It's not an all or nothing, but you want those experts to come in and make sure it's successful because there's nothing worse than launching things and then having to put out. left right and center because you know you again you had AI you know do some things and nobody really knows up down left or right about what's happening. Brent Peterson (11:54.706) Yeah, and think one thing that too, that's, you you mentioned knowledge and documentation. It is an opportunity that merchants and SIs have to just dump documentation into a place where then you can have AI read it. Kind of like Google Docs is now, but now a little bit smarter. You put everything into Google Docs and nobody understands where anything is or even SharePoint, right? Like I've heard SharePoint is a dump certifier, but AI is helping us with that. So I think employing AI where it's really a benefit to us, and then employing humans to be a management console on top of that, right? I think humans are the important part of how we want it, because really, even if you have an agent that's running some of those features, it doesn't necessarily understand what that new client may need or that client that's approaching your ICP that... would need something that's a little bit different than what the core is. It may suggest something crazy, kind of like what you were talking about when the VTECs people first came in. There might be some things that you don't want it to tell people. So I think that's where the human factor comes into the whole equation. From iPaaS and from what you're doing in iPaaS, how are you seeing automation and how is it growing in that sort of connecting the dots space. Robert Rand (13:22.21) Well, I mean, for us, we're a more modern platform than most people have experienced for data integration in the first place. And so we've got for the e-coms and all these other softwares, we've got templates already to get you a lot of the way there. And so I think that that... onto itself as a big differentiator versus a lot of the legacy systems that gave you a way of integrating, but really said, now go write it all. And so we've got the API calls already set up and we've got the common fields of data set up and most of the implementation comes down to what's unique about this business and their business rules and their. unique fields of data, like there are unique product attributes based on what they sell and making sure that all that is mapped and successful. And where we're adding AI is for the pieces. I was mentioning with you earlier about, for instance, if you need to go beyond the no code and into the low code, how can we make that a faster experience for you to be able to transform and translate data in more unique ways for your very specific use case? And again, sometimes data needs to do more than get from field X in one system to field Y in another. Something has to happen in between and making sure that that process is easier and faster to set up. After that, we've got things in our roadmap, like every software company, we are an actual data hub. You can see all the data in our system and we become an amalgamated source of truth of that data. so right now, we see users trying to pull all that data toward analytics and BI and other systems, getting it from all their sales channels, all their finance and operational and other systems to some end point. But also, Robert Rand (15:17.448) thinking about just querying it on our side and having certain opportunities there for those that want it and for those that don't, we treat data... Again, as Sacra sang, that's first party data of our customers and our subscribers own that data. And so very particular. I think that's another important factor that comes into all this, of course, is in some cases you want to be unique. In other cases, you want to be able to do things the way that others are doing and capitalize on available systems. But you also want to make sure that you're protecting your data and your core intelligence. We're all data companies now. And so, you know, these things also fall into line when it comes to data privacy and security of that data. And again, you know, when something is written out, you know, more at random that isn't going through the kind of processes that we do for SOC 2 compliance and other compliances, you know, you've got a, you've got, you don't always realize it upfront, but you're taking on more in the long run that, sometimes you think you're saving money, but then you see that your AWS or whatever hosting bill to manage these data flows is skyrocketing every time you update your product. or something's going on and with a system like ours you've got a flat predictable monthly fee you're not subject to those kind of you know roller coaster spikes and so I think that you know that that's where Robert Rand (17:00.202) you know, there's that long term outlook on these things that when you try to get back to what I think a few years ago, we saw a lot of merchants and a lot of businesses focusing on that, that total cost of ownership and that long term, not just how am I going to get this working today, but what's this going to look like months and years down the line? It all, you know, starts to really add up to where where you should just get something quickly set up through automation or where you should be using a more defined system or a more defined team to manage something. Brent Peterson (17:38.764) Yeah, you know, maybe we could kind of end our conversation today on the difference between the self-serve small connector platforms like Zapier to the more robust ones like iPass. Robert Rand (17:53.57) Zappier is a great product. I'm glad it's out there. They are really good when you need to get a little bit of data from A to B. They're not as, so they're not hub and spoke like us. They're not focused on, for instance, we'll get an e-commerce order and we'll send it to the ERP for accounting, the, you know, some kind of shipping and fulfillment system, warehouse management or point of sale or what have you for fulfillment. We'll send it to the CRM so the sales team has visibility, et cetera. et cetera, and you can turn on and off different spokes to the hub, and you can manage all the errors and exceptions in one nice, neat place with proactive notification, et cetera. So the ability to deal with all of that with these pre-built integrations and to kind of unify all of your data is much more unique to us, where we're great with, as your data sets are growing, that, again, predictable expense, where Zapier the pricing gets, you can really get out of hand as you scale with them. We're also more mission critical if data fails to flow with us, like when that AWS outage that we saw a couple of weeks ago, where maybe some e-comm or other systems went down for a bit, our system will keep automatically retrying persistently. It'll retry frequently at first and then less and less frequently over time. Last I knew was Zapier, they'll try a couple of times and then you're out of luck. and those orders aren't getting where they're going or. that accounting is going to be off or whatever's happening. And so there's just a lot of layers like that where we see them as a great tool for getting some more basic data going from A to B. But when you're thinking about more complex data flows of, instance, in order fulfillment, you know, order is going to come in and then it's got to get shipped and maybe there are partial shipments and those have to come back to, let's say the e-comm to report on that and send tracking information to the shopper. Robert Rand (19:55.812) and have the order information correct, and then more data back and forth and maybe there's a return or a cancel, you're getting into much more involved data flows. I would see that as being a lot harder to manage through a Zapier by far. And so there's economies of scale into what people are trying to achieve and what system is the right system for it. And what system you can grow successfully with in your Brent Peterson (20:28.3) Yeah, great point. So Robert, we have a few minutes left here and as we close out the podcast, I give everybody a chance to do a shameless plug about anything they'd like to plug. What would you like to plug today? Robert Rand (20:40.193) boy, well, I think we've done a pretty good job of talking about ipass.com. heading into the winter here. I've been to all the trade shows. I, the last few months, I think I've been to a dozen trade shows in the last three months. So I don't have any to tell you to have people meet me at upcoming. I know Dave from our team is gonna be at Meet Magento Canada in the weeks ahead. That's coming up here in November. But no, I'm. I don't have a lot more to tell people about for today, just always excited to join you for the podcast and looking forward to the next time. Brent Peterson (21:26.124) Awesome, thanks for that Robert. Robert Rand is with ipass.com, a middle layer for your business. Robert, it's been such a great pleasure. Thank you so much for being here today. Robert Rand (21:39.298) Pleasure's all mine.