How Orium’s AI Playbook Turned Complexity into 5x Growth

If there's one thing that you as an unassailable advantage, that's just gone now. But there's another thing that was never possible that's possible. And your job is to figure that out before your competitors do.

Alright, everybody. Welcome back to agents of scale to show where we talk to people who are actually building the future of work with AI. And today's guest, we've got Jason Cottrell. He's the CEO of Orium and newly elected president of the Mach Alliance. Now Jason has spent over a decade now helping some of the biggest brands, modernize their digital foundations.

Oriam specializes in helping folks launch composable ecommerce platforms. And Jason can correct me on the specifics of that, when we get on here. Jason, welcome to the show.

You know what, Wade? Great to be here. Thank you for having me.

Before we dig in, I I heard from somebody, I think it was you, were joking that you were a walking consumer electronic show. I'm curious, what's your, like, favorite gadget you've got right now?

Sure. You know what? I I'll maybe go in order. I I will say, like, I I've mellowed maybe in the last five to ten years. I would say, my favorite is actually pretty pedestrian.

Like, my Oura Ring, I've had it for years. The the idea of living like, it it is a device. If I if I lost it, I would immediately go ahead and replace it, and it's just so easy to fit into your day to day life. Maybe if I if I went I'm actually I've kinda maybe looking at what's on my desk here. If I were to, go maybe one step further, the one that I find I really like but weird some people out is the the meta Ray Bans.

Mhmm. And I think they're really great. They're proud. If I'm out with, like, kids, I can take quick photos. I can ask quick questions.

Like, I find for me, it's actually, you know, again, pretty easy to incorporate in my day to day life. And I think that's a a great sign in terms of how things have advanced in the last five to ten years. But I find that one, some people when they know it's a connected device, still give you a a second look. But I would the one that gets me, so I'll wear this device. It's an it's an IO, and I think the company's actually I'm not sure they even make them anymore.

But it's basically a blue light band. So when I'm traveling for work, when I'm changing time zones and I've got a flight to Europe, later this week, I'll actually wear it to help me adjust my sleep schedule. And I'll wear it on the plane, and I got this blue light on. And I probably look like an absolute idiot, but it works, so I do it. Yeah.

That's incredible. What do you make of the AI pendants that are, making the rounds? I don't know if anyone's gotten, like, crazy popular yet, but you're starting to see more

of these. You know what? I think it's gonna be one of those things that's strange until it's not. Yep. I think, you know, you look at the original Google Glass and some of the initial instantiations of, say, the concept of smart glasses, and then all of a sudden, you know, it's taken, you know, ten probably almost fifteen years, but the the finally, you get devices that are a couple $100 and actually do some targeted things reasonably well.

And I think I think it'll happen faster than that, but I I think it's gonna be one of those things that's strange until it's not.

Yeah. It's, slowly at first, then everything all at once. Yep. Okay. So tell us about Oriam.

How'd you come up with the idea? What were you what were you doing before? What's the problem you were trying to solve? Give us the origin story.

Yeah. So, I mean, we're a consultancy and system integrator. We're generally working with the largest brands in the world. Companies doing hundreds of millions to billions in online transaction volume. And so, you know, what what we've carved out a role doing, you know, most of these companies have pretty complex systems on how they handle their customer interactions, purchasing, fulfillment, and do that on a global basis.

So we've carved out kind of a strong niche with many of the the fastest growing providers in commerce experience and actually increasingly kind of agentic capabilities, to really deploy that infrastructure. So thinking of us as an adviser, but also a, almost towards a general contractor starting to, you know, steadily bring out the old stuff, bring in the new stuff, but you have to do it without ever kinda missing a beat, because that's a very expensive mistake.

Yeah. And one of the things you all specialize in is helping these, large companies, large ecommerce platforms set up these composable infrastructures. Talk a little bit about what that means and and why you all sort of believe that is the ideal way for these, companies to to build their, systems.

Yeah. And and I would say, you know, it's it's more than in the enterprise at at at scale. It's more than the ideal way. It is the absolute reality. None of these companies run their entire business on one system.

And there was a point maybe fifteen years ago where either it was all custom built or it was all, you know, they could buy everything on one system, and that ran most of the commercial kinda operations for for commerce as it manifested. But now now that things have have scaled, gotten more complex, now that commerce is not just online, it's also store, it's also wholesale and fulfillment and b to b. It's also all these other channels and all these other things. The reality is, you know, most of these companies are running 20 to 40 different systems. And so we try to look for smart ways to first kind of consolidate that a bit, but the reality is, that's what they need to run their business.

And so the concept behind Composable really was this push in the last five years. It's part of, you know, obviously, why we believe in the concept of the of the mock alliance as well. There are conceptual and technical ways to do this well. And there are vendors that have, embraced, you know, being able to be plugged in and plugged out and working well with other vendors. And and I think most who've been in the enterprise space long enough know that's not always the case.

Some some try to pull walls up. Some are only thinking in their specific domain. So So I think it's we we tend to represent partners that plug in well, play well with others, and technically have the foundation in place to support that well. It also tends to mean that they've been kinda many of the most pioneering partners at in the push to AI and AgenTic and kinda really carving out their role there. They're able to move fast because they they, you know, they can build for very specific domains of expertise.

Can you give an example of something you've done for one of your clients?

Yeah. So so, I mean, I think that, you know, actually, we won an award just last year for one of our customers, and it's a brand called Harry Rosen. And I think it's a really great example because a lot of people think of these use cases as just must be these, you know, big massive enterprises. And and they're a great example where, you know, they do a few 100,000,000 a year, but they were really looking to pioneer. It's it's, a gentleman who is taking over as kind of third generation in the business.

He came in and was on kind of this big all integrated sweep. And every time he wanted to touch something, it was months. And, and really kinda was funny. He's trailing the market. He was worried.

He's saying, like, I I need to make sure that there's a fourth generation that we're thriving and that we're serving our customers and reinventing ourselves, for what's about to come. And so we were actually able to work, over the course of, actually, primarily nine months kinda start to finish to replace the core of, of what they were doing. And and since we've done that, they've, you know, almost 5x'd, their their kinda digital commerce volume and brought it into store, totally overhauled their loyalty, totally changed how they the channels in which they interact with customers serve and sell. All those things, you know, they were able to do in a pretty rapid period of time. And I think part of that is, like, they were pushing boundaries, and then they also, you know, they made a choice to really kind of approach, kind of pulling together the systems that were best for them for their specific use case.

So you mentioned increasingly this is happening with agentic capabilities.

Yeah.

What what is actually working here? Why is this like, why is what you do, this composable approach, such a good way to for for companies to adopt agentic capabilities?

I see two two parts to it. So one, you know, it it there's been all sorts of noise the last few weeks. Right? Obviously, the there's the MIT study, which is how many organizations are struggling to derive kind of the the target ROI from their use cases. There was a great piece from Google that was looking at, obviously, actually, a much higher number that are succeeding.

Succeeding. But in both cases, what are the themes on what is and is not working? And one of the key things that's come out is when it's, you know, coming in and dropping millions, tens of millions solving these broad use cases, those tend to fail. When you're when you're coming in and working with specialized vendors in specific domains, that's where it tends to succeed. And we've seen the same.

Like, when we're when we're building, agents for customers, they're not these one massive agent that do all. And you've probably seen this in your own experience, but, like, it's it's specific points of expertise, specific functions fine tuned, and often kinda combine together in different common that that's the concept of composing. Whether you're doing that kinda within the architecture of an agent or you're doing it across agents, across systems, it it tends not to be just, like, one you know, you sign one thing and you're done. It's being built up in specialty domains often with different partners. And I think what's really exciting there, like, I I like to say it it's it's not going to be one agent that runs your business.

It's going to be many, many agents that run many, many parts of your business. It's almost more like how you think of hiring, managing hiring, training, managing, running, and or and and employing people. I think the second component of comp so the second component of composability that I think is powerful in this domain is this is a Game of Thrones out here. Right? And, you know, one week, one group's in front, and the next another group's in front.

And the reality is it's gonna be the next few years that shake out in terms of, you know, not even just who wins and loses, but who's good at what, who's most cost effective at what. And so I think, what's powerful there is if you spend, you know, all your time, energy, effort locking in on one path, there's a very high odds that you will get outpaced on a one to two year basis. But people who think of they're building systems. Like, if if if we've got a like, we we we run Zapier in our business. Right?

If we've got a Zap that we wanna change kinda one input, one system, one one output, that's that's fine. That's pretty straightforward for us to do. We have that adaptability. And I think the same is true, like, when you're when you're thinking about the agentic infrastructure in your business, how you're approaching the problem. Thinking about composability, that you're gonna wanna fine tune it, curate it, iterate it.

A good example in our case would be, you know, we've pushed down a path on some fronts that touch into our finance back office. And, obviously, you know, some of those systems, you know, aren't necessarily the best for, plugging into the rest of the business and or kind of automation. Now as we upgrade that we've been able to build to a point now as we upgrade that. The ability to incorporate a more modern system now is gonna come much faster. We've built for kind of that iteration in mind.

I think the last part that's really exciting about composability is the most powerful thing that we found is it's not about just taking an old process and and and snapping it into an automation. It is about fundamentally rethinking the process, and that often is rethinking just what context and information do we have. What's the start with the context and information we have, and what's the outcome we want? And when we kinda back up and approach that, and when you think regardless of systems and when you think regardless of departments, and really think, what's the goal we're going for? And I think it allows us to rethink how systems, people, and processes are gonna come together and how we kinda compose them in the right mix, who we use for what.

And, again, I I think that that businesses that have the ability to do that will unlock some of the highest value out of what Agenta capability is going to offer.

Yeah. I think you're hitting on a couple really important points here. One is most of these companies simply can't they're they're using so many tools. It's not practical to say, okay. We're just gonna do this massive lift and shift to, like, one net new thing.

And but what is more practical is to say, hey. We're gonna take this component, and we're gonna modernize this. And we're gonna swap out this piece for that other piece in more narrower capabilities. And then, you know, over time, you've gotta, you know, kinda ship a Theseus sort of thing where it's like Yeah. It's still both, but it's all new pieces than than before.

I

think I think that's a hun I think it's a 100% correct. And and, you know, also too, like, I think I I I remember, one of the major tech CEOs suggesting essentially it's gonna be AI and data and everything in between is gonna be gone. And and I I don't think that's the case. I I think, you know, we're finding tons there's a lot of things you want done in a deterministic system in a deterministic way. There's a lot of things that you want with kind of heavy controls placed around it and constraints.

There's a lot of things where just plain old automation is is perfect. It's cheap. It's predictable. It's an and, you know, in many cases, we want our systems we want certain systems of record to be predictable and deterministic and to be able to interface with automation, but I want it done the same way every time. And so I think, it's the combination and the ability to to move in slices, I think, that's very powerful.

That that that is gonna be the reality for the next five to ten years at least.

Totally. And the second point you made on, you know, Game of Thrones esque world is, you know, we're in this transformative shift. You know, everybody is predicting where they think it's gonna go. But I think practically speaking, we don't know know where it's going to go. And so in a time like this, it is especially beneficial to have the optionality to say, ah, I'm gonna swap out this, try this new thing.

Oh, here's another new thing. I wanna the ability to try that out versus sort of feeling like you've gotta make these three, five year commitments to a particular way of doing something and then just realize a month later, like, ah, shoot. There was a there was something new that I didn't know was coming. And so it's so powerful to take this approach.

You got it. And and it's also the place to get the fast start too. Right? Like, when when you when you bite off too much, I see it happen over and over where, you know, that same organization is sitting there. You know, they spend the first three months then trying to get everybody in on it, the next three months arguing over who should be doing what, the following three months re reassigning that.

Generally, nothing's been done still at this point in time. Mhmm. And, instead, those who, you know, are starting in targeted domains, thinking about scalability, have a north star. Like, they're the ones starting off making process, building, kinda real points of expertise, momentum in the organization. Those are the ones I'm gonna bet on in terms of who's gonna successfully transform to be an agentic enterprise.

Yeah. Now you mentioned the MIT paper. You mentioned the Google study. You're having a chance to work with some of these very large commerce companies. What are you seeing that is actually working from an agentic point of view?

Like, what are the what are the best doing that is separating them from the rest?

I think, so I'll take the commerce domain to begin with, and and, what I will say is that's actually been particularly transformative because, it's not just internal optimization where sometimes, you know, there can be all sorts of reasons why, you know, market pressures may be delayed. You may not go through kinda some of the optimization that you could as a business. But when your customers are fundamentally finding you a different way, have changed their expectation on how they're gonna interact with you, they're changing the products the criteria and the products which they're gonna buy, based on kind of, now kinda answer engines and kind of more personalized shopping assist, and very, very close to, it's actually one of the I think, been one of the most interesting announcements in our space last week is actually just being able to buy. And there's more and more around payment and fulfillment protocols now that's been defined that companies can invest against. So I would say actually interested, like, how our what we are seeing is, you know, customers who are optimizing for these changes and leading there, are now starting to see, that gain.

And and, you know, customers of ours, brands that we're watching who are thinking about assets that will kinda transition from what helps human buyers today gives them some gain, but also gonna help them as agents play more and more of a role in discovery, in choice, and eventually in purchasing, they're starting to win. And I think that gap is gonna widen, but they're building the capabilities towards that, investing towards that. And even though there's still some uncertainty, they're taking steps. And and that iterative, composable approach helps them do so. They're not, you know, making one year big bets.

They're they're moving fast and adapting. And in many cases, it means that they're not just, trying to meet, you know, optimize content for, think about the buying journey, be able to take purchasing and orders. They're not just doing that. They're thinking about then how that's going to how is all that going to be reorganized so that it supports internal operations? How can we deliver more to point of sale with clienteling to their contact center?

Inline agents that can eventually negotiate on the brand's behalf with customers. They're also starting to think rethink. I think that it it snowballs quickly to rethinking their internal operations as well. And I think that's a powerful thing.

Yeah. And, of course, you're talking about here, you know, Chattypati announced that they're launching apps, which allows you to show up directly inside the experience. They also announced partnerships with, I think, Stripe and Shopify so you can check out directly inside of ChatJPT. And so you're seeing just a lot more shopping happen inside of that surface area in

the past. That's exact that's exactly so we we were we were one of the few SI launch partners for the ACP announcement. So the Mhmm. The Stripe, ChatGPT, Shopify kinda combo. And and there's there's a number of our other commerce partners like Commerce Tools, Commerce, and others who are kind of adopted in that standard too.

And and I think what it is is it has the critical mass, that, you know, may this be the defined standard five years from now? No. But it's going to be a primary one. There's there's there's probably no way to not have some support for it in terms of how you'll be discovered as a brand that does commerce and and how you'll be able to take transactions for at least, certain types of purchases. So it's got enough critical mass there that that people can now brands can now invest against it.

Mhmm. And I think most, I mean, most consumers have no idea this is coming. But the fact that it's announced that you can do it with Etsy now that, like, companies are clamoring now to get get this capability added and get approved for it. I I think that that's amazing. I'm really excited to see what that does even to a degree during this holiday season, but especially next year.

Yeah. It's you know, it's been interesting to watch my own personal shopping patterns change with these tools. Like, I I'm finding that I that I buy more stuff. And it and it's I buy Otter things too because it it's like these niche problems that I have, where in the past, I'd just be like, like, it's it's not that big of a deal. The amount of research I'm gonna have to do to actually go figure out how to solve this thing is just like not practically worth it.

And so I just kind of like live with this annoyance in my life. Whereas with Chatty Bitty, you can literally just be like, hey, I have this annoyance. What should I do about it? And it, like, gives you a whole plan. And, you know, some of them, it's just like, hey, make these lifestyle changes or these behavioral changes or whatever.

But some of the times, it's like, here's a thing you can buy for it. And I'm just like, oh, if I can just buy that thing and it cost me $20 or whatever, I'd be like, great. You know, solve my problem. And so there's these, like there's this category of, like, minor nuisances that just would have been ignored in the past that now I'm finding, hey, I'm not I'm not ignoring them because I have a tool that can help me do it. And I have to imagine that's gonna show up in a bunch of, like, commerce players, conversion rates, sales, etcetera.

I I think that's correct. Because also, you know, now often you're not just going in to find and buy a thing. Right? That's certainly a category of how people are using these tools. But I'll, you know, I'll give a good example would be, I had to re I I wanted to I decided I'd be smart, I'd I'd take on kinda repainting, or kind of refreshing a porch at at one of our rental properties, a concrete porch.

I had never done that before. But actually working through ChatGPT with with video, let me show it to you. How would I approach this? What are different ways I could solve this? What are the pros and cons?

And then working through to, okay. Here's my approach. What product should I buy? What can I get near me? Like, being able to walk through all of that all the way through to, I think at the end of it, I was literally on video, like, am I applying this?

How am I applying this? And it was almost like a back and forth discussion being able to see it, have a convo. It's almost like like, I don't like like a like a tradesman is there with me or my dad is sitting over my shoulder offering advice kinda thing. So I think, you're it fits into these journeys. It fits into these overall optimizations you're trying to do in your life, and then the product you're selecting is just a part of that.

And I think it it, you know, companies that understand that rebuild content for that, think about how they fit in that journey. I think there's a lot of opportunities for, you know, actually, very pretty specific niche products to to do very well with that because ChatGPT can kinda help help unlock that that power, or it and Gemini and others can unlock that path.

Yeah. It's the it's like it's bringing like, in the case of that porch, like, you're buying things that in the past, the tradesmen would have bought. But now it's like opening up a new market for that product to somebody who's like, well, why can't I do this? Like, my

I can do this. Yeah.

This is, you know, which may be famous last words, but, you know, a lot of these things aren't aren't that complicated.

Yeah. Yeah. Exactly. It can give you the confidence to do it. Yeah.

For sure.

Totally. So, hey, you just got elected as the president of the Mock Alliance. For those listening, what what is that, and, what what role does this, consortium play?

Yeah. So so the Mock Alliance, I think, is a really incredible organization. So it's over a 122 different software vendors. Most of the global systems integrators, so the Accentures and Deloits of the world. Two of the major hyperscalers, so AWS, GCP, and then, you know, a number of the major kind of payment providers and other infrastructure providers for kind of modern cloud in content customer data, and digital experience.

So I I think I mentioned that cohort before of companies who really play well in specific domains and play well together, play well with your existing systems, and have been built around that ethos and with the appropriate technical underpinnings. And so, you know, the alliance has done a lot of work the last few years really pushing legacy vendors both to kind of up their game, and to do it years faster and to a higher standard, but also developing a mark of trust that, you know, that the tens of thousands of enterprise customers served by this cohort kind of know that these the vendors in this cohort will will kind of interoperate and work pretty well together. And really where where the alliance is is going and what why I've taken on the president position, what was interesting to me, is we're pushing to be one of the first true kind of agentic ecosystems. So partnering with many of the standards bodies and trying to kinda bring that into practice amongst our members. Our members already are building kind of over a 100 different kind of point agentic capabilities often tied into their products.

What we're pushing is how do we make sure that we start to open up more agent to agent use cases, agent to system use cases to do it sooner, and to do it in a way once again that kinda within twelve months or so, we've got kind of these real productized available use cases for companies to deploy for their customers. And I I think it's it's it's a really interesting challenge. Like, there's there's lots of agent building going on out there. I think you would probably see some of it too. A lot of it is just, like, it's a point agent.

It's not designed to do anything else on its own, or it's separate systems. And then maybe, you know, say Zapier, you're interacting with the systems, but the agents live off somewhere else. We're really trying to push to look at the innovation that's going on in terms of building agent capability, what jobs that solves, all the innovation our our our members are already doing, but pushing to start to link that together in our domain of expertise, our ecosystem sooner rather than later, and that's what we're pushing on.

So as someone I I I definitely love this approach. As someone who's benefited a ton from things like REST APIs and OAuth to be able to integrate 8,000 different tools together, Like, interoperability does create such a valuable set of things for customers because now they're able to get everything that they bought to work well together. In this agentic world, where do you see we have interoperability issues? What are the places where we're gonna need standards, we're gonna need some of these capabilities to emerge?

It you know, it's interesting because I I think we're we're watching very quickly as some of these standards emerge as they evolve. What I would say is so we're we're watching, obviously, like, elements like MCP a two a start to define kind of how some of these points of interface might work. Standards like ACP, AP two starting to come out that start to define, say, how payments, transactions, trust might work. What's really interesting to us and is the unsolved domain right now is is actually more how will composability of agents like, what what how will the kind of ecosystems of agents come together? The organization that we've looked to and partnered very closely with, that I I think is solving a lot of that is a group called, Agency, AGNTCY.

And it's a it's a consortium over of over a 100 firms. Or Oriam also as a supporter of that. And, you know, they're really looking more at, like, how will discovery work? How can identity and trust work? How can agent agent communication work?

And how will observability and governance work? And I think, you know, we're finding there there's elements of that that many enterprises can pull out and make sense of what works for them. And for us as an ecosystem, the same, is what are the parts of this that can make sense, as we look to kinda, I would say right now in the next six months, my and to twelve months, my priority is just challenging our members to build together sooner and and think of use cases across each other sooner. But to to start to then bring kind of, you know, see what what's working, what are the approaches that kinda can become a more defined approach for our ecosystem in this commerce customer data and, and experience space, and, you know, start to bring those to the table. But I I think what agency solving is one of the more interesting ones for me that that's getting less coverage, but really will be what's required to get proper agent ecosystems working.

Yeah. MCP is one that we've built a lot on, model context protocol. You mentioned a two a on these other sort of, protocols that help agents talk to agents. Where are you seeing like, what are the use cases you're finding adoption of this to be quite effective?

You know, I think I would say some of it is obviously kinda like honestly, some of it's just from a tool chain perspective. Mhmm. You know, we're finding, I would say, you know, those who are rethinking actually their day to day work are tending to be the fastest adopters right now. Like, we we've got staff who, you know, they've used Zapier to put together I think you do, like, the custom custom MCP where they can stitch together, like, an optimized and controlled kinda endpoint they can work against. And then they're they're running that against their code generator, or they're running it against kind of, a CRM or sales agent that our teams built.

Like, some of those use cases, I think just people can move fast, within constraints. You know, they say the guardrails of product like yours can put around it today. I see those team members, those teams really pulling apart. From a broader kinda company perspective, what I would say is, many are working with MCP. I worry that in many cases, they're just essentially plugging up against it, but the system underneath doesn't have the appropriate guardrails, observability controls, like access control and and permission constraints, honestly, to make it work properly.

And so that for us is a big focus is either we've gotta start putting some of the, kinda agent frameworks, and we've gotta start wrapping more around it. Or that's where I think, like, products like yours really come to the table where they can offer, a layer of control and governance that that can allow you to move faster without having to overhaul absolutely everything you're doing. But I would say right now, think that's actually been installed. Like, there's lot of promise around MCP, but many using I worry about how many are using it. This there's not enough the systems underneath it weren't built for it, and and they're lacking key components of being able to work with this kinda safely and scalably.

Yeah. We're definitely finding that our customers are benefiting a lot from our MCP server being able to inherit our access controls. And so you can, you know, do granular approvals at a, like, per per scope level, basically, at the end of the day Yep. Is is pretty powerful. What you know, you mentioned before, like, hey, there's a lot of use cases where I love a plain old deterministic workflow.

How are you how are you thinking about educating your clients on, like, when to use some of these pieces? Like, when do you use an API? When do you use an MCP server? When are using an agent? When are you just using a workflow?

Like like, how how do you explain that concept to others?

Yeah. I think I mean, what's interesting about it is, look, it it starts with let's keep it as simple as possible. So let's start with just straightforward automations as much as possible and go as far as we can with that. I think there's the second component, though, of just, like you know, when you start to hit the limits of that, then rethinking, you know, what where where is a person picking up on this? What are they actually doing?

How are they thinking about the problem? I think that's where you can find interesting examples on where, LLMs can be inserted to kind of how how and where they would find their place. Like, if I look at, like, we're you know, for contract generation in our case, I don't generally want an LLM, dude. I'd like, we like, a a straightforward mail merge or similar document merge or document population on on parts of a template is all we we don't wanna be checking everything. We don't wanna be rewriting it.

But you know what? For parsing all the calls we do with a prospective customer, using that to define specifically what needs to be done, validating that with or alongside a person. You know what? We can now do that better and generally, sorry, faster and generally better than if we were having a human do that alone. So an LLM's great at that, and then we can plug verify that bit, and then, yeah, plug that in the contract, which I don't want to change.

I want I want that exactly how it is. So I I think we can start to break out parts of the workflow, but I think the key is push as far as you can with deterministic or kind of more straightforward automation methods, and then think about where you plug in kind of some of the nondeterministic methods. I think the key that's changed, though, is you can take it so much further. So the the reward to go down the path, on on many more processes, is higher. So it it you you find the mix.

It might be 80% in one camp, 20% in another. But the end reward to get to kind of to actually be able to rethink that process, is powerful. I laugh because for us internally, we're really at a point for many processes now that, like, it's frustrating for us if it's not automated. And you you probably are maybe I take this for granted. You're probably, you know, ahead of us just given the ethos of your firm too.

But now it's like, why is this not automate? It it went from why should we automate this to why is this not, like, this is not automated. Mhmm. We've had this concept of, the three times rule in our firm. I think it it's one of the things that I've observed with those doing kind of, like, AI native startups from from the ground up, and we've really, really pushed to bring Asnithos to ourselves and to our, and to our customers.

And when you find those, that that's a huge unlock. And I think one of the big differences I've seen for, like, AI native firms versus those who are not.

You said two things that resonated quite a bit with me there. The idea that you know, if you it's like every like, if you do something three times, you gotta go automate it. Like, that to me is such a powerful behavioral unlock. I talk to a lot of folks that wanna do automation, but they don't necessarily know how to recognize when they should do it. And it's been interesting to watch this agentic wave come upon us because it's opened up everybody's creativity again.

It felt like it's given them permission to be like, oh, why shouldn't I automate this? And I'm finding that in talking to a lot of these folks, you know, nine times out of 10, they didn't actually need AI. Like, this was perfectly solvable with, you know, plain old fashioned deterministic automation workflow software for the last ten plus years. But just the fact that we have this, like, magic AI thing out there, now people are starting to think about it as possible again and starting to recognize those things. So I do think teaching people how to recognize that stuff, is is pretty powerful.

The second thing that we've noticed is, hey, you talked about how that like three recognize it three times as a powerful cultural tenet as well too. We've noticed this too where, you know, for the longest time, we've had this value. Don't don't be a robot. Build the robot. And, you know, right now, you see a lot of employee bases, you know, being fearful of, like, what what might AI do to my job?

And I think that is a fairly human reaction, an understandable one. But if your culture is one in which that says, I'm curious about what is possible with these tools, you're constantly going to be able to evolve and adapt to the times, and your skill sets will be much more future proof if you're leaning in versus if you're scared of where these things are going. So I think you hit, yeah, just two critical, bits there.

There there's an interesting kinda it's actually, a framework that, I think we took from a post that that Zapier had done a few months ago, and we've been using it for most of this year, with our own team. And so one of the things I found really interesting there is people don't jump to magically just rethinking their whole profession. And most start with some degree of hesitancy. It's natural. And I think it was kinda activated, adaptive, transformative.

I think we're kind of the general premise premise, and and we've we kinda subset and define that a bit more for our use case. But what what's interesting is no one jumps to the fourth spot. They everybody moves in in those stages. Like, almost almost they they maybe they take a week. Maybe they take a year.

But everybody moves through that. So we've kinda taken the challenges recognized and then just try to move it get you get you through each next one. Don't jump to the end. Get you through the next tier, as fast as possible. And so, I think one of the things that as we get to the later stages of or so at at the early stages, what we found is so powerful is peer to peer sharing.

Because honestly, me standing up as a CEO and saying, this is gonna be wonderful. This is great. I think people kinda they go in the side eye. And sure. Versus when it's when it's peers and their professions are like, this is how it's working for me.

This I can't like, I never have to do this again. Like like, this this gave me this unlocked this whole new capability for me. When people are coming to the table with I think combined with how we're using some of these tools made in our day to day use, just personally, those two things seem to get people started in my experience. I think our role as leaders, as we get to the later quadrants, is really helping people rethink, like, their, how functions are going to evolve in this new era. Because I think one of the most powerful things is, yes, some parts will get faster, some parts will get more automated.

I mean, look. We've had DevOps for quite a while now. We write more software than ever, have more engineers than ever, and some of that pattern's accelerating. But overall, I think, like, it means we're gonna write more software first, but it also rethinks what's possible. Peak people can play across different domains that they that that were not possible tomorrow or sorry, sorry, previously.

But I look at, say, our, you know, QAs, and people are saying, well, we're not gonna need any QAs anymore. So, a, yes, we're writing more software. And, actually, the infrastructure to support all of that, means, actually, you need automated QA, if anything, more than you needed before. So there's a step one. There there's kind of a a a you know, some role drops off, but there's something needed.

But we're finding an actual agent building, like the eval's skill set, is a very viable transition for some types of QA, especially those who may be pushed more, like, have a bit more dev background or self talk and or have been down to, like, have that testing mindset. And also, it's not 10% of a team anymore. It's, you know, 30 or 40%. And so so I think there's all in nearly every function, I think there's an interesting rethink that you're about to go through if you look at, like, truly what are you motivated by, what's the mindset and value says and what values do you bring to the table? And and I think for for most, there is, an opportunity.

So the more that we as leaders can paint what that is for, our team, I think then people really in those last few steps then can start to engage in in that and feel the permission to start to push, like, bringing those new capabilities kinda to their teams, to the company.

Mhmm. You know, I think the sort of comment you made on learning these capabilities in stages resonates so much. And I see this is where a lot of folks fall fail is that they try and jump to the finish line. They're like, I want this, like, magical experiments where I, you know, automated all of customer support or, like, engineers are never writing another line of code again, etcetera, which sounds, you know, incredible to sort of be able to do that sort of stuff. But the reality is it's sort of like trying to go to calculus and you don't know what fractions are or you don't know what multiplication is.

And so a lot of this is more progressive where it's okay. Like, how do I how do I get a taste? How do I get my toes wet? Okay. I I know how to do that piece.

Okay. Great. Now how do I start to compose these things together and chain these things together? And all of a sudden, that's how you build up to some of these more sophisticated systems, and it takes time. And I do think it's important that, like, leaders are able to step in to guide the company through that transition, not just simply go, we're gonna go do AI.

And then to your point, people are kinda like, oh, okay. Yeah. Sure. Sure, Wade. Sure, Jason.

Like, I I guess we are at the end of the day. One thing I was curious to talk to you about is, you know, you've talked about one of your favorite quotes is, hey, if you've if your business had one advantage in the last decade, it's now gone. And I'm curious here, how do you think about uncovering new advantages, discovering new ways of competing in the age of AI?

Yeah. You know, and and I do love that quote and it it's Ethan Moloch and and, you know, I I love kinda obviously, kind of perspective he puts out in this space. Yeah. Like, if there's one thing that you as an unassailable advantage, that's just gone now. But there's another thing that was never possible that's possible.

And your job is to figure that out before your competitors do. And I think, you know, if if I look at our own business, right, like, in consulting advisory, like, that concept of billable hour of utilization, like, that's just been the thing. And, that's what you optimize for. That's what customers are used to buying. And I think, you know, steadily over the coming years, that that is and should go away as a concept of procurement and measure.

And on one hand, I think that's terrifying for some. On the other hand, like, what we're finding is, like, customers want outcomes, and we can optimize for specific types of expertise. And I can fix price things and say, will get you this defined change from a to b and put a price around that that the customer can count on despite variability and uncertainty that I never could have accommodated before, and do it and do it comfortably. And the customer gets exactly what they want. I get exact like, I I, you know, we define kind of a higher value place for us to play and can start to optimize for how well we do it, how fast we do it, not just how many people are parked in the seat doing it.

I think it I think that's win win, but it's a shift. It's a shift for firms like ours. And I think it's just one example of how that's changing just kind of across the board, across the board, for, for firms. And, you know, I admire, like, the like, obviously, you've been pushing as well to really put kinda Zapier at the forefront of what's going on too. And and you've been doing it before.

It was kind of it's part like, it's clear at the pace you're rolling it out. You've been thinking about this for a while, and really pushing to where you you have kinda new advantage as well.

Yeah. Another unique thing is you established Oriam as a b corporation, which

Yeah.

Means in addition to shareholders, you have other constituencies you care about, employees, customers, society at large. I'm curious, where do you think AI helps you benefit those other stakeholders and the other constituencies?

You know, I think I mean, the reality is this is going to be disruptive. The just as the Internet was disruptive, just as, like like like this but and this is, I think, gonna be as or more substantial, and it's gonna happen faster. And and there's no way around that. The what I'm optimistic about is, you know, we have ways now to model, circular economies, ecological outcomes, to invent and bring to market new drugs, new therapies, to be able to do personalized medicine, to be able to manufacturing, things that were were, that I think hold us back as a society right now, that we we have new solves for. It's it's almost like a larger version of what we we were talking about earlier, like, say, more functionally or within our firms.

And so I think I think that, you know, a key is, you know, trying to make choices, at least for us as a firm along the way, in terms of who we partner with. Like, how are these workloads powered? Who are we partnering with? You know, who and where are we partnering with for that? How can we bring as much of our team as possible and and optimize kind where they are at the end of this, and that they're benefiting from being on the forefront of this?

And and I think that, you know, trying to keep in mind too, like, there's these broader rethinks, that are going to be possible and trying to at least be be some small part of that where we can see those opportunities over the coming years. But, you know, I I think it's it's it's never been it is as an entrepreneur, and I I I love this stuff, This you know, I I've never been more, if you like, I've never seen so so much disruption and so much opportunity all at the same time. But overall, I think, you know, that that will work out net benefit for for the world.

Yeah. I mean, it's hard to think of a technology that has as much transformative capabilities as one like this. Certainly, haven't seen it in my lifetime. Yeah. Jason, thanks for joining us.

This is our latest episode of agents in scale, and, I hope you all will follow along for our next episodes. Cheerio, everyone.