AI agents are quickly reshaping the procurement technology debate, reviving an old question in a new form: if software can increasingly be assembled by vendors or internal teams, why commit to a platform now?
According to Arkestro, that question overlooks the reality of procurement itself. The discipline is not simply a matter of automating tasks or streamlining approvals. It is shaped by messy data, fragmented supplier information, commercial complexity and the need to make de...
Continue Reading This Article
Enjoy this article as well as all of our content, including reports, news, tips and more.
By registering or signing into your SRM Today account, you agree to SRM Today's Terms of Use and consent to the processing of your personal information as described in our Privacy Policy.
fensible decisions in real time. In that environment, the strongest technology will not necessarily be the most eye-catching agent, but the one built on a dependable execution layer.
That is an important distinction as enterprise software vendors race to position AI agents as the next interface for procurement. IBM, for example, is marketing watsonx Orchestrate as a set of pre-built agents for supply chain and procurement designed to improve resilience, supplier management and cost efficiency. Other specialist tools, such as Procbay’s bid analysis agent and Fluenta One’s sourcing agents, are pitching automation across supplier discovery, proposal review and contract validation. Taken together, these offerings show how quickly the market is moving beyond simple chat-based tools towards more task-specific systems.
Yet the central limitation remains the same: agents are only as good as the information and rules they can access. TechRadar has argued that AI agents should not be expected to run supply chains on their own, noting that physical supply networks depend on context, live visibility and human judgement. Its reporting on data quality makes the same broader point in another form: autonomous systems can fail badly when the underlying documents and datasets are incomplete or unreliable.
That caution matters in procurement, where decisions are rarely made on price alone. Supplier performance, contractual terms, governance requirements and operational constraints all shape outcomes. Even where agents can accelerate analysis, they still depend on the quality of the workflow around them. If the data is poor, the logic is weak or the process is badly designed, automation can simply make mistakes faster.
This is why the build-versus-buy debate is likely to remain more complicated than the current AI hype suggests. General-purpose agent frameworks may be useful for narrow tasks, but they do not automatically replace the systems procurement teams rely on to manage sourcing events, evaluate trade-offs and record decisions properly. For that reason, Arkestro argues that the long-term advantage will belong to companies that combine AI with robust data and operational infrastructure, rather than those chasing the most novel agent interface.
The most realistic view is that procurement AI will be additive rather than substitutive. Agents may take on repetitive research, comparison and validation work, but human oversight will remain central wherever commercial judgement, supplier risk or governance is involved. In other words, the future of procurement is likely to be less about replacing platforms than making them smarter, more connected and more dependable.
Source: Noah Wire Services
