Ravi Wesley is a venture capital investor at Marula Square and formerly Industry Strategist at Blue Yonder. He has worked across both enterprise software and early-stage ventures, giving him a distinctive perspective on how supply chains invest in technology and what often gets overlooked. In this conversation with JP Doggett, he discusses why innovation cycles are too often reactive, how financial framing can distort investment decisions and why composability and agentic AI may change the rules.
JP Doggett: You have some interesting takes on how supply chain innovation is funded and framed financially. What do you mean by that?
Ravi Wesley: Many people treat SaaS like a clean OPEX subscription, but in practice it carries CAPEX-like characteristics. You’ve got integration work, onboarding and long-term contractual commitments. The cost profile looks very different once you factor those in.
On top of that, technical debt is real but often invisible until it bites. Legacy systems create drag on innovation, but those costs don’t show up until there’s a failure or an upgrade that takes longer and costs more than planned. If you don’t surface that in the business case, it’s easy to underestimate the real investment required.
JP Doggett: In our exchanges, you’ve also pointed out the problem of reactive innovation cycles.
Ravi Wesley: Yes: too often the trigger for change is a burning platform. Leaders hold off until something forces their hand and when that happens the instinct is to go with the safest, most established vendor. It makes sense politically - no one gets fired for choosing the IBM-equivalent - but it narrows the options.
The irony is that this creates a cycle. Because the decision was made under pressure and often oversold, the outcomes don’t match the expectations. That fuels scepticism and the next time innovation is needed people hesitate even longer.
JP Doggett: You’ve also cautioned about the way vendors talk about composability.
Ravi Wesley: We should be wary of vendors simply rebranding what they already have. Just because a system exposes some APIs or uses microservices doesn’t mean it’s truly composable. If adopting it still locks you into a multi-year implementation cycle, then you haven’t really solved the problem.
True composability should let you try something in a controlled way, swap it out if it doesn’t work and scale it if it does. If that isn’t possible, then it’s just marketing hyperbole.
JP Doggett: You’ve spoken about agentic AI as if you were “hiring an AI agent.” Can you expand on that idea?
Ravi Wesley: The way I think about it is that generative AI is good at ideas and creativity, while RPA is good at structured execution. Agentic AI is essentially the merger of the two... an AI that can both think and act.
But for that to work, the system architecture has to be modular. If everything is locked inside a monolithic platform, the AI can’t actually operate. It’s like bringing in a new employee and then refusing to give them access to the systems they need. That’s why composability matters so much in this conversation.
JP Doggett: A lot of supply chain leaders I speak to don’t use the word “composable.” How do you approach that?
Ravi Wesley: I tend to avoid jargon. If I say “modular” or “flexible,” people get it straight away. What matters isn’t the label but the ability to move quickly without being trapped by legacy decisions.
That’s also why digital fluency is important for supply chain leaders. They need to be able to cut through vendor claims and spot what’s real and what’s just branding. Otherwise you risk buying into a concept rather than a capability.
JP Doggett: How do startups fit into this picture?
Ravi Wesley: The challenge with startups is trust. Leaders worry about scale, security, and whether the company will survive long enough to matter. That’s a rational concern. But if you have a composable architecture, you can lower the stakes. You can test a tool, get value and move on if it doesn’t work.
Some big players like DHL have experimented with structured programmes that bring startups closer to customers. That helps manage the risk while still allowing new ideas to surface.
JP Doggett: Who do you see as leading examples of this shift?
Ravi Wesley: Some of the most interesting stories are where operators became providers. Ocado moved from being a retailer to selling automation technology. Tesco did it with Dunnhumby and now with Tesco Media. Amazon and Jabil are also examples.
It shows that supply chains don’t just have to consume technology. They can create it and, when proprietary data is becoming the critical input to AI, that’s a powerful source of competitive advantage.
Ravi’s perspective highlights a tension supply chain leaders know well: ambition versus credibility. SaaS doesn’t eliminate the need for capital-like commitments. Burning-platform decisions can backfire. And the promise of composability or AI only holds if leaders have the fluency to see past the labels.
As Ravi notes, “we should be cautious of hype but confident in building the capabilities that give us real options.”