At the tail end of 2025, we embarked on a more formalized customer outreach initiative than we’ve been accustomed to in the past. Sure, our CSMs are in regular communications with our customers, but we were interested in getting some unfiltered, strategic insights about how our customers are using our AI agents: what’s working, what’s not working and where would they like us take our solution in 2026? Super important work; one must always embrace the voice of the customer. Every company worth its salt does it.
Most of the feedback was what we expected, which was reassuring. There was one insight, however, consistent amongst the customers we spoke to, which took us by surprise. In response to the question: “What are Didero’s top three value propositions?” We heard a lot of, and I’m paraphrasing, “Didero was our very first foray into agentic AI. You were a low-risk on-ramp for us.”
In the course of our interviews, we started to unpack those responses. We heard all about external pressures to post wins with AI, from the C-suite, boardroom, and even self-induced. A few of our customers were contemplating hiring expensive consultants to map out a grand master plan for deploying AI across their businesses, most were trying to figure out where to even begin, and a lot felt confused and overwhelmed, understandably so.
And BAM, Didero just happened to be in the right place at the right time. As one customer put it, “Didero isn’t fancy or complicated. It just does the work I need it to do.” In the case of this customer, that work entailed Didero agents autonomously resolving open line items for past due orders. By communicating directly with the customer’s suppliers via email, our AI agents took that number down from 2,300 past due items to 261, in a matter of a few weeks.
“That’s amazing!” our customer declared. Well, we agree…this customer had just posted an agentic AI win! Was it earth shattering for their business yet? Maybe not. But, it was definitely material, and it was measurable.
And so it went, on down the line – we kept hearing about these small wins (and some bigger ones), from other customers using our agents.
As one customer told us, “Initially, we had your AI agents running in the background, making recommendations for how to respond to supplier emails [questions and issues] to our buyers. Because your agents provided us with an audit trail for their recommended actions, we could easily determine if the recommendations were good or not.”
Over time, as their confidence in our agents grew, our customer took the reins off. They let our agents autonomously interact with their suppliers to resolve purchase order issues via email. Our agents also made the corresponding data updates to our customer’s ERP system. Chalk up a win!
My advice to anyone struggling with how to adopt agentic AI is simple: start small and start now. Don’t succumb to analysis paralysis. If you have a high-volume, repetitive, tactical task which people inside your organization are performing today, tackle it with AI agents.
Carve out a small problem set or start with a few vendors. Set up the specialized agents – in our case, those trained to manage purchase orders, sourcing events and AP – and have the agents begin making recommendations in a sandbox environment. Have your subject matter experts vet those responses. Get comfortable with the quality of the output.
Then, slowly deploy the agents into your production environment with human oversight. Eventually (sooner than you think), the vast majority of the work performed by the AI agents will require very little, if any, human oversight.
You’ll start seeing some small agentic AI wins, en route to even bigger wins. You’ll be in the game instead of on the sidelines.
For our part, we’re going to keep talking to our customers. They always seem to be right. Spoiler alert: as one of our core value propositions, you’re going to see “Easy On-Ramp to Agentic AI” on our website very soon!
