How do you choose an AI workshop provider?

The three questions to ask any AI workshop provider

The first question — do you have vendor relationships? — reveals whether the workshop content will be shaped by commercial interests. A provider that is a certified partner of Microsoft, Google, or any other AI platform has structural incentives to design workshops that lead towards those platforms. Ask directly: do you have any commercial relationships with AI tool vendors, do you earn fees or commissions from any AI platform, and would any of those relationships affect the recommendations you make in our workshop? The second question — what is your operational experience? — distinguishes providers who have used AI at scale in business contexts from those whose experience is entirely advisory or academic. Workshops grounded in operational experience are more credible and more useful than workshops grounded in theory. The third question — who delivers the workshop? — addresses the common practice of selling workshops on the basis of the principal’s credentials, then delivering them through a junior team.

What to avoid in AI workshop procurement

Several red flags in AI workshop procurement are worth identifying. Generic curricula — workshops that look identical regardless of the client’s sector, size, or objectives — suggest a provider that is not designing for the specific client. Platform bias — workshops that happen to conclude that the provider’s associated platform is the best solution — suggests commercial interests are shaping content. Low pre-work requirements — workshops that gather no information about participants before the session — suggest the workshop is not personalised to the audience. Absence of written outputs — workshops that end with a presentation but no document each participant takes away — suggest the provider’s interest is in the session itself, not in the participant’s outcome. Junior delivery — workshops sold by a senior principal but delivered by an associate — suggest a production model rather than an advisory model.