Surveys Say Peer AI Success Stories Matter More Than Tool Demos

The person two seats over already ran your pilot.

Image by Tung Nguyen from Pixabay

By Nils Swenson

Twice in the past year, ASA’s Applied AI Task Group surveyed members on where they stand with AI, once last fall and again in April. The surveys covered adoption, tools, challenges, and concerns, but one question turned out to be the most revealing: what support from ASA would help you most?

The top answer, both times, was peer case studies and success stories, ahead of webinars, vendor evaluations, and tool demonstrations. And the share asking to hear from peers grew between the two surveys, from roughly a third to nearly half.

That result says less about ASA’s resource library than it does about what evaluating technology feels like right now.

The two surveys sit six months apart and describe noticeably different rooms. In the first, the largest group of respondents was still exploring and researching AI use cases. By April, the center of gravity had moved to piloting, and a small group described mature adoption across business functions, a category that was nearly empty in the fall. The respondent pools were different, so the comparison is directional rather than precise. But the direction is hard to miss, and it matches what comes up in task group conversations: companies are moving from reading about AI to running it.

The usual way of evaluating technology was not built for this pace. A distributor evaluating an ERP module could take a year over it. Demos, reference calls, a selection committee, a decision. The product at the end of that process was the same product from the beginning of it. AI tools do not hold still that way. Capabilities shift quarter to quarter, a vendor pitch from nine months ago describes a different product, and even the vocabulary keeps moving. Agents went from concept to live demonstration inside a year.

What has not moved is where the friction sits. In both surveys, members named the same top two challenges in the same order: manual and repetitive work, then data quality and integration. The tools are changing fast. The operational problems they are being aimed at are not. Which means the most useful information is not what a tool can theoretically do. It is what happened when a company with your branch count, your ERP, and your data quality pointed that tool at a real workflow.

Guides and checklists still earn their keep, and ASA keeps producing them, from the cybersecurity starter kit to the CRM readiness assessment, because teams need a structured place to start. But a guide lays out the steps. It cannot tell you which step a company like yours underestimated, which workflows the counter staff quietly refused to change, or that the pilot would stall for two months on product data cleanup nobody scoped. That information has exactly one source: an operator who already did it, who can also tell you what they would sequence differently if they were starting over. The guides work better when that conversation comes with them, and the surveys say members know it.

There is a related signal in the same surveys. Asked where they expect to learn AI, the largest group in both waves said self-training. Most people working through this are working through it alone, assembling a picture from vendor demos, articles, and their own experiments. Peer conversation is the correction for that, and right now it mostly happens by accident, a hallway conversation at an event or a phone call to a friend at another supply house.

The Innovation Summit was built to make it happen on purpose. It runs November 9–10 at Caesars Palace in Las Vegas, ahead of NETWORK2026, and the agenda covers the same fast-moving territory: a keynote from Chad Foster on sorting signal from noise in the current modernization cycle, Spencer X. Smith on building customer intelligence as an organizational capability, and Steve Lerch on moving AI from experimentation to execution. The robotics session is a panel of operators who have deployed automation in live distribution environments, focused on what triggered the decision, what broke, and what they underestimated.

But the structure is the point as much as the topics. The Summit closes with facilitated roundtables, peer-driven discussions organized around the questions attendees actually brought with them. The speakers frame the territory. The operators in the room are the part you cannot get from a recording, because the most useful thing at an event like this is not the slide deck. It is the person two seats over who already ran the pilot you are about to propose.

The technology discussed in November will have shifted again by spring. That is not a reason to wait. It is the reason the peer network is the durable part, because the people who worked through this round of tools will be the ones working through the next round too.

And if you are already planning to be at NETWORK2026, the logistics are simple: the Summit runs the day before, at the same property. Arrive one day earlier and you get both. Registration is $500 on its own, or $250 as an add-on to your NETWORK2026 registration. Details and registration are at asa.net/Innovation-Summit.