The AI race is starting to expose something uncomfortable. The winners will not be the companies with the most tools. They will be the companies with the fastest adoption. Oren Etzioni, founding CEO of the Allen Institute for AI, just broke down the 2026 Stanford AI Index, and one finding stopped him cold. The United States leads the world in building AI.
The United States does not lead the world in using it.
The supply side is not close. In 2025, U.S. private AI investment reached 285.9 billion dollars, 23 times China’s and more than the rest of the world combined. Many of the leading models are trained in American labs. The infrastructure is here. The access is here.
And yet, when the Index measured adoption as the share of a country’s population using generative AI tools, the United States ranked 24th, at 28.3 percent. The United Arab Emirates topped the list at 64 percent. Singapore came second at 61 percent. Norway, Ireland, and France rounded out the top five.
Etzioni named the pattern directly. He called it the diffusion gap: the country that builds AI is not the country that uses it. Then he closed off the easy excuse. “The gap is not about ease of access,” he wrote. “Americans can use the same tools, on the same day, for the same price (usually zero) as anyone else.”
That should make every agency owner stop for a second. Because the same gap is opening inside companies, and it is opening for the same reason.
The corporate version of the same number
Zoom from the country down to the company and the paradox repeats. Last year McKinsey ran a research, and found that 92 percent of companies want to go hard on AI over the next three years, yet only 1 percent consider themselves mature. That is the diffusion gap with a logo on it. Everyone wants in. Almost no one has actually rebuilt how work moves.
The pattern shows up even in how people perceive their own usage. Gallup and McKinsey found that 99 percent of Americans have used AI-powered products in some form, but only 65 percent realize it.
Presence is not proficiency. Using a tool is not the same as building around it.
The events industry gives you the gap in sharp focus. Across the industry, 91 percent of event professionals use AI in some capacity, but only 15 percent qualify as strategic leaders. The rest are experimenting in isolation, without the system architecture to make their experiments compound. A separate Cvent and Northstar PULSE survey of more than a thousand planners found that 65 percent use generative AI, yet only 16 percent say it has significantly improved their planning and execution.
The tools are there. The workflows are not.
What is actually happening inside companies
The tools are available. The subscriptions are paid for. A few people are experimenting. Someone has a prompt folder. Someone else is using ChatGPT for meeting notes, email drafts, or social posts. But the business itself has not changed.
Sales still lives in one place. Client history lives somewhere else, scattered across inboxes and memory. Project delivery depends on who remembers what happened last time. Job estimates sit in one system and final project costs sit in another. Operations are held together across inboxes, spreadsheets, proposals, contracts, show files, vendor notes, platform exports, and institutional knowledge.
There is a predictable shape to this.
When you add AI to an operation, productivity often drops first, because people are doing their day job and learning a new way of working at the same time. Double the cognitive load, half the speed. You climb out of that dip only when AI gets woven into the daily workflow. Past that point you reach something that compounds. Most companies never climb out, because they never built the workflow underneath. They stayed in the dip and called it a trial.
That is not AI adoption. That is AI dabbling.
The difference is not effort. It is structure. The majority of people using AI right now are using it like a glorified search engine. They ask a quick question, get a quick answer, and move on. That leaves most of the technology’s potential unused. There is a real distance between asking AI to think for you and using AI to expand your capacity to think, decide, and lead.
Where the real opportunity lives
I shared in one of my AI keynotes recently how, the next competitive advantage for agencies will not come from randomly adding another AI tool. It will come from understanding where work gets delayed, repeated, misunderstood, re-entered, manually tracked, or trapped inside one person’s head.
Picture a Monday morning at a mid-sized agency. Leads are in Gmail. Proposals are in PDFs and Google Docs. Project tracking is in Monday or Airtable. Labor is in a payroll tool or someone’s calendar. Invoicing is in QuickBooks. None of these systems talk to each other. None of them know it is Monday. The only thing connecting them is the project manager, and the project manager has not finished her coffee yet.
That is where the real AI opportunity lives. Not in the tool. In the connective tissue the tool is supposed to run on.
This is the work we are building around: helping agencies move from zero to AI-native in a way that is practical, responsible, and tied to how the business actually runs. My partner, Jeff brings the data infrastructure, reporting layers, database design, and automation. I bring the event-industry side, the agency and production workflows, the buyer language, and the operational reality of where work breaks. The first step is not a tool. It is an audit.
The AI workflow audit
The first step is an AI workflow audit. Not a software pitch. Not a prompt training session. Not a pile of abstract recommendations that nobody has time to implement.
We run it as a fixed-scope, two-week diagnostic. It maps where leads and RFPs live, where proposals live, where invoices and estimates live, where job costs live, where labor and freelancer costs live, and which reports leadership rebuilds manually every month or after every event. It surfaces the business questions that cannot be answered quickly today. Then it produces a data map, a reporting gap list, a recommendation for what to build, a sample dashboard wireframe, an implementation estimate, and a set of risk and cleanup notes.
The questions it forces are simple and uncomfortable.
- Where do your leads, proposals, invoices, labor costs, and final project costs live today?
- How long after an event do you actually know whether it was profitable?
- Can you see your likely revenue and capacity for the next 90 days?
- What report does someone on your team rebuild by hand, over and over, because the system cannot produce it?
A good audit looks at how the agency currently sells, plans, staffs, delivers, reports, follows up, and makes decisions. Where is time being wasted. Where is client information scattered. Where are teams repeating the same work. Where are proposals, contracts, emails, budgets, show documents, and event data disconnected from each other. Where could AI assist with research, summaries, routing, analysis, content, client communication, reporting, or operational decision-making. And where does the business need better structure before AI can even be useful.
That last question is the one many companies want to skip. They should not.
AI does not fix the mess. It exposes it faster.
AI does not fix a messy workflow by magic. It usually exposes the mess faster. AI walks into an organization like a mirror with a motor attached. It reflects what is already there, then speeds it up.
If your sales process is inconsistent, AI will scale the inconsistency. If your data is scattered, AI will surface fragments. If your team has no shared process, AI becomes one more thing people use differently depending on their comfort level. The industry is already building agents that send RFPs and draft proposals. The harder truth sits underneath all of it: if the dataset is incomplete, stale, or too shallow, the agent only makes a flawed process faster.
This is why tool-first adoption is such a weak strategy. The implementation gap is not an AI readiness problem. It is a systems design problem. The technology exists. The data architecture and workflow design to use it well often does not. A login is not an adoption strategy. A training session is not a culture shift. A policy document is not a workflow.
But when the workflow is mapped properly, AI becomes a serious business layer. It can help an agency respond faster, document better, reduce repetitive admin, improve handoffs, build stronger client intelligence, create better reporting, preserve institutional knowledge, and make decisions with more context. That is the difference between using AI and becoming AI-native.
Why this matters now
For agencies, the pressure is already here. Clients expect faster answers. Margins are tighter. Teams are stretched. Senior people are carrying too much undocumented knowledge, the kind that vanishes the moment they leave. Newer team members need ramp-up support. And leadership needs clearer visibility into what is happening across sales, delivery, finance, operations, and client experience.
There is a continuity risk hiding in here that owners underrate. If your new hires depend on hallway context, scattered Slack threads, undocumented decisions, and “ask Sarah because she knows how we do it,” your AI tools will hit the same wall those new hires hit. Only faster. AI can help with all of this, but only if it is applied to the right workflows in the right order.
The audit is how you find that order. It gives leadership a grounded picture of where AI can create value now, what needs to be cleaned up first, what should be automated, what should stay human-led, and what should never be handed over without oversight.
Where AI belongs and where humans stay in control
Becoming AI-native does not mean handing over judgment. It means deciding, on purpose, what gets delegated and what does not. A simple way to sort the work is in three steps.
- The first step is what AI can handle safely: drafting, summarizing, formatting, the mechanical reading.
- The second step requires your review: anything that goes to a client, a sponsor, or leadership.
- The third step is never delegated: final decisions, sensitive personnel matters, stakeholder positioning, the judgment calls.
AI accelerates execution. Judgment still belongs to the human layer. The reading is what gets automated. The decision stays yours.
The companies that win this next chapter will not be the loudest about AI. They will be the ones that quietly rebuild how work moves through the business. They will know where AI belongs, where humans need to stay in control, and where better intelligence can turn scattered work into better decisions. The future of AI adoption is not just prompt literacy. It is operational literacy.
If your agency is still in the “we should probably be doing more with AI” stage, that is exactly where an AI workflow audit starts. We help agencies see the business clearly, identify the highest-value AI opportunities, and build the first practical steps toward becoming AI-native.
Etzioni found a country that pioneered the technology and then underused it. The same story is playing out one company at a time.
Access is no longer the advantage. Adoption is. And adoption starts with understanding the work.
About the Arthur
Anca Platon Trifan, CMP, WMEP is an AI strategist, keynote speaker, and CEO of Tree-Fan Events Productions LLC. With more than 20 years of experience in event technology and AV production, she helps organizations design smarter workflows, strengthen operational resilience, and integrate AI in ways that support people, not just processes. She is also the host of Events: Demystified, where she explores the intersection of AI, leadership, and the future of business events.
If you want a clearer view of where AI belongs inside your organization, book a call with Anca to explore your workflows and opportunities for AI integration here.
If your agency is using AI in pieces, but the business still runs across inboxes, spreadsheets, proposals, contracts, finance tools, project files, and people’s memory, this is the place to start. The AI Workflow Audit is a fixed-scope, two-week diagnostic for agencies that want to understand where work is getting delayed, repeated, manually tracked, or disconnected before they make another tool decision.
We map where your leads, proposals, estimates, invoices, labor costs, project data, reporting, and client history actually live today. Then we identify where AI can create practical value now, what needs to be cleaned up first, and what should remain human-led.
