Transcript 0:00 In late June 2026, the last of the top eight source-to-pay suites in the Gartner Magic Quadrant quietly shipped their AI agent studio. And with that, every serious procurement platform on the market now has one. 0:14 So this is a good time to ask, what is an AI agent studio? Why did all of them build one at basically the same time? And what does it mean for you, the procurement pro who has to make this thing work on Monday morning? 0:27 Let's get into it. [upbeat music] Welcome to ProcureTech Unpacked. I'm Joel Collin-Deverse. 0:43 This is the show for procurement professionals who wanna understand the technology shaping our function. Okay, so let's define this thing because the term is slippery, and honestly, that's half the challenge. 0:55 An AI agent studio is a no-code or low-code environment built right into your enterprise platform where your team designs and wires up LLM-supported business processes. 1:07 So instead of hard coding one fixed word flow the way we've done it for years, you drop LLM-powered steps into a process and you chain them together towards an outcome that you're looking for. 1:19 The studio part is what matters. It's not just an AI feature sprinkled into the product. 1:25 It's a place where you can build these processes and where governance controls decides what the model gets to act on, where a plain boring deterministic algorithm does the job instead, and where a human has to stay in the loop. 1:38 And here's the additional challenge. Nobody agrees on what to call it. What one vendor calls an AI agent studio, the next calls an AI agent builder or an orchestration studio or an agentic platform. 1:52 Same capability, different sticker. So the skill worth building is recognizing the thing no matter what the vendor stamps on the box. If you take one thing away from this episode, take this. 2:03 There's a real difference between a studio and embedded AI, and most vendors are going to blur that line in the demo. Nearly every platform now ships with LLM features baked straight into their workflows. 2:16 Those features are useful, but they're the vendor's processes running the vendor's logic, not yours. A studio is different. A studio hands you the keys to the building. 2:27 Embedded AI means you consume someone else's automation. A studio means you author your own. That shift from renting a fixed LLM-supported process to building your own is the real story underneath all the naming noise. 2:41 So here's how you pressure test it. In a demo, don't let vendors walk you through the pre-canned agents. Ask them to build a brand new agent live in front of you. Not run one, build one. 2:54 What you're watching for is simple. Could your team do that after go live? Or are you going to be filing tickets with a vendor every single time you wanna change a step? 3:04 One more thing so you don't get tunnel vision with those big logos. This is not a Magic Quadrant only club. 3:11 Plenty of vendors outside the Magic Quadrant have shipped this capability to ConversionIS, Levelpath, Oro Labs, Tonkean. They all come to mind and I'm sure I'm forgetting a few. 3:21 And in a lot of cases, they've got their months, sometimes years ahead of the suites. So the studio conversation isn't just for the eight names that everybody already knows. 3:32 Now why did everybody build one roughly at the same time? Well, because practitioners finally clocked the thing that early AI hype skipped right over. 3:41 If you want consistent performance out of a business process that leans on an LLM, you can't just prompt the model and cross your fingers. 3:50 The more you ask of a single LLM call, the more the odds of hallucinations, uh, go up. So getting reliable results at scale means being deliberate. 4:02 Deliberate about how, when, and why an LLM enters your process at all and pairing it with good old-fashioned deterministic automation and algorithms along a workflow that you actually design. 4:15 And when I say this out loud, it probably sounds familiar because workflow builders have been around for decades. Honestly, an AI agent studio to me is just the next iteration of that same idea. 4:27 Design the process first, then automate it. The only thing that's new is the tool in the box, an LLM that can finally handle the fuzzy steps that a rule engine never could. 4:39 That's a big deal, but it doesn't change the fundamentals required for success. Quick aside, and this is the part that lives better on a page than in your ears. 4:49 I won't read out a wall of links here, but if you want to get to practical starting points, I use something called a process scope diagram, also known as an IGOE diagram, I-G-O-E, uh, to actually design these processes and AI agents. 5:04 I've also written before about how to manage your procurement business processes end to end. Both of those pieces are linked in the show notes if you wanna learn more. All right, so what does this actually mean for you? 5:15 Two things. First, thinking in business processes is still a skill that pays. Before you can hand a process to an AI agent, you have to understand the process. The studio doesn't save you from that work. 5:27 It assumes that you've already done it. Second, once you can do that, you'll notice very quickly that not all of these studios are created equal. 5:37 Some let you bring their own LLM, others lock you into a frontier model, take it or leave it. Some ship with prepackaged procurement agents, others hand you an empty toolbox and a good luck. 5:49 Some let your team build and deploy freely, others make the vendor vet and approve every agent before you can do that. 5:56 Some are true low code that a business user can drive, while others still need a technical admin sitting in the loop before you can start using the agents. 6:05 None of these characteristics are automatically right or wrong, but every one of them shapes your operating model and your future flexibility, and most of them never come up unless you bring them up in a discussion. 6:18 So here's the bottom line. If you're serious about building an agentic procurement organization, an AI agent studio isn't optional anymore. You need one. But landing a decent flexible studio is only the starting line. 6:32 The real work, the part that separates the team getting value from the ones getting stuck in automating their own confusion in a permanent proof of concept 6:42 is building agentic procurement process management as a capability that sits inside your procurement team. A good tool helps, but it doesn't do it for you. Your people do. The studio just gives you somewhere to do it. 6:56 That's it for this one. If it was useful, the deeper version, the one with the IGOE diagram and the business process how-to articles lives in the Pure Procurement Newsletter. You can subscribe over at pureprocurement.ca. 7:09 I'll see you next week. [upbeat music] That's a wrap on this episode of ProcureTech Unpacked. 7:15 If this one resonated, subscribe to wherever you get your podcasts and sign up for the Pure Procurement Newsletter for weekly insights between episodes. 7:23 If something we covered sparked a question or an idea, we'd love to hear from you. All the links for the reports we discussed are in the show notes. We'll see you next time. [upbeat music] Resonate.
