The AI Adoption Path
The short answer
Plan AI adoption in five dimensions, strategy, people, data, technology, and governance. Score yourself, pick one workflow that can pay back in 90 days, run a small instrumented pilot, train the humans involved, and then scale. The five-minute quiz below gives you a baseline across all five dimensions with no email required.
We built the AI Adoption Path for the people who decide what their organisation does about AI, this quarter, with the budget they have.
The framework
This is the same path we walk clients through, distilled. Each step maps to one or more of the five dimensions the quiz scores you on.
Before you buy tools or run pilots, map where your organisation actually stands across strategy, people, data, technology, and governance.
Avoid the moonshot. Choose a workflow with clear inputs, repeatable outputs, an owner who wants it, and a measurable result. Marketing ops, sales research, and internal knowledge search are common winners.
Time-box to 6 to 8 weeks. Define the before metric, the after metric, and the kill criteria up front. Ship to real users, not a demo audience.
Train the people who will use it, name an internal champion, and write the acceptable-use rules before you scale. Adoption fails on culture far more often than on technology.
Most stalled AI projects are data projects in disguise. Inventory your sources, fix permissions, and decide what is safe to send to a model and what is not.
A one-page AI policy covering data handling, vendor approval, disclosure, and review beats a 40-page document nobody reads. Update it quarterly.
Promote the pilot to a funded programme with an owner and a roadmap. Be willing to switch off the experiments that did not move a metric. This is the step most teams skip.
Most AI assessments measure technology only. We measure the five dimensions that actually decide whether adoption sticks.
Direct answers to the questions we get most often from leaders planning AI adoption.
Start with a five-dimension baseline across strategy, people, data, technology, and governance, then pick one workflow that can pay back inside 90 days. Run a small pilot with clear before and after metrics, train the team that will use it, write a one-page AI policy, and only then scale. The Artificial Breakdown AI Adoption Path quiz gives you the baseline in five minutes with no email required.
Most organisations move through four stages: experimenting (individual employees using consumer tools), piloting (one or two sanctioned projects with named owners), operationalising (AI embedded in a core workflow with metrics and policy), and scaling (a cross-functional programme with budget, roadmap, and governance). Skipping straight from experimenting to scaling is the most common failure mode.
A useful first pilot should ship in 6 to 8 weeks. Reaching the operationalising stage in one core workflow typically takes 3 to 6 months. Becoming a scaled AI organisation, with multiple workflows, governance, and a funded roadmap, usually takes 12 to 24 months. Timelines stretch when data and policy work are deferred.
Sponsorship sits with the CEO or a C-suite peer. Day-to-day ownership works best when split between a business owner who cares about the outcome and a technical lead who can ship. A dedicated AI council with marketing, operations, IT, and legal representation prevents the technology from becoming an IT-only initiative.
You need a data inventory, not a data strategy. Know which sources exist, who owns them, what is sensitive, and what is safe to send to a third-party model. A full data strategy can follow once you have one workflow proving value. Waiting for a perfect data foundation is a common reason adoption stalls.
Written by
Chief Innovation Officer at ZGM Modern Marketing Partners
Pete has spent the last two decades helping marketing and technology leaders adopt new platforms without breaking what already works. Read more on the about page, or listen to the podcast.
Last updated May 26, 2026