Market Dojo CTO: procurement is getting agentic AI wrong

Market Dojo, a Bristol-based provider of AI sourcing software, has published a pointed challenge to the dominant narrative around agentic AI in procurement, arguing that the industry is chasing full automation when it should be building human-AI partnership. The company's CTO, Nic Martin, contends that the central question driving procurement strategy is wrong, and that misframing it carries real organisational cost.

The argument lands against a backdrop of strikingly low executive confidence. Gartner data cited in the Market Dojo report shows that only 36% of chief procurement officers feel very confident in their ability to redesign procurement workflows around AI, while a mere 14% believe they have the talent in place to meet future demands. That gap between ambition and readiness, Martin says, is being papered over by a fixation on automation rather than augmentation.

Autonomy is not the destination

Market Dojo's position is that agentic AI, autonomous software capable of multi-step reasoning and task execution without constant human instruction, should function as what the report calls a "procurement expertise engine": surfacing category knowledge, best-practice guidance, and sourcing intelligence at the moment a human practitioner needs it, rather than removing that practitioner from the loop entirely.

"Current conversations focus on what AI can automate," Martin writes in the report. "We believe procurement leaders should be asking a different question: how can AI help people achieve more than ever before?"

The company frames this under a "Sustainable AI Philosophy" built around three pillars it labels People, Pricing and Planet. The people strand promotes augmentation over replacement. The pricing strand promises AI value without inflated licence premiums. The planet strand emphasises lean deployment, a nod to the growing attention paid to the energy footprint of large-scale AI inference workloads, a cost that procurement functions running high-frequency sourcing events at scale would eventually encounter on their own balance sheets.

The convergence angle: agentic AI meets enterprise spend

The debate Market Dojo is entering is considerably larger than procurement software. Agentic AI is currently the most actively contested design philosophy across enterprise technology, spanning legal workflow automation, financial planning, customer operations, and supply-chain management. In each vertical, the same fault line appears: vendors promoting full-autonomy pipelines against practitioners insisting on human-in-the-loop governance.

For the cross-sector investor or platform strategist, the procurement domain is a useful leading indicator. Procurement sits at the intersection of operational data, supplier relationship networks, and contractual compliance, making it one of the richer test beds for agentic systems. How the category resolves the autonomy question, whether regulators or corporate risk functions ultimately demand explainability gates on autonomous sourcing decisions, will have direct read-across to agentic deployments in financial services, healthcare procurement, and public-sector contracting.

The talent dimension is equally telling. The 14% figure on workforce readiness echoes findings from wider enterprise AI surveys and underlines a structural constraint that capital alone cannot quickly fix. Organisations that invest in AI tooling without investing in the change-management and upskilling layer are, on Market Dojo's reading, likely to see adoption stall regardless of the sophistication of the underlying models.

Market Dojo's report is, in form, a piece of thought-leadership marketing designed to differentiate its own augmentation-oriented product positioning. That caveat noted, the underlying data points and strategic framing reflect a genuine inflection in how enterprise buyers are approaching agentic AI purchasing decisions. Whether the "people-first" framing becomes the category consensus or a niche selling point will depend substantially on how early high-profile autonomous-procurement deployments perform, and whether any of them produce the kind of visible failure that accelerates regulatory scrutiny of AI decision-making in supply-chain contexts.