What the adoption numbers show
Architects are a common title in computer and engineering circles – systems, computer, network. That is not new information. Lately, however, the term architect has started turning up somewhere I didn't expect, attached to strategist roles at marketing and consulting agencies. The title itself isn't really the point, though. What matters is why the shift is happening. A strategist's job increasingly includes analyzing and defining where AI belongs inside a client's systems architecture, and how it moves the needle for teams like Rev Ops and digital marketing running the work from day to day. This is a standard part of the solution strategy now, not something cemented after the plan is already built. It's a fundamentally different kind of work than picking a tool or running a pilot, and the people best equipped to do it aren't necessarily the ones with the newest tools. More often, they're the ones who've spent the longest time inside how these systems get built.
That kind of judgment – AI architecture vs. standard systems – is harder to find than one might think, and the adoption numbers from McKinsey's State of AI 2026 survey show why.
9 in 10: organizations use AI regularly in at least one business function
37%: attribute any measurable EBIT impact to it
6%: qualify as AI high performers (attributing at least 5% of EBIT to AI)
Adoption looks different when you sort U.S. businesses by size, according to the U.S. Census Bureau's Business Trends and Outlook Survey.
37%: usage among firms with 250+ employees
Under 20%: usage among firms with four or fewer employees
17–20%: national average across all firm sizes
So together, those adoption numbers illustrate a wide spectrum of usage rather than a pattern of single stage adoption for all businesses. Plenty of large organizations are well past questions like "where do we start" They're all too often stuck, and instead asking why they aren't seeing significant impacts in the results. While smaller businesses at various stages mostly, are still asking beginner questions; often citing concerns about tool quality or legal and compliance risk, according to U.S. Chamber of Commerce data.
Research from the JPMorgan Chase Institute shows newer small businesses adopting paid AI tools far faster than businesses founded only a few years earlier.
77 months: to reach 10% paid AI adoption, for small businesses that started in 2019
6 months: to reach 10% paid AI adoption, for small businesses that started in 2025
For consulting agencies, that acceleration is reshaping what a marketing and digital strategist is hired to do.
The AI Architect role is changing
In AI engineering, "architect" already means something specific: the person who owns the system's tradeoffs, not the one implementing its pieces. The role involves consulting steps: what to build, what to buy, how it scales, and how to explain the decision to people who aren't technical.
That same shift now shows up in agency strategy roles. A strategist's job increasingly includes recommending – usually during the solution strategy phase – where AI belongs inside the system being designed for a client. Decisions must be made like what a model needs to know before it can be trusted with anything ("context architecture"). Or how tools, data, and people should hand off to one another, an AI-qualified lead moving cleanly from a paid campaign into a rep's queue instead of stalling in a Rev Ops blind spot. All considerations that allow the pieces to work as one system instead of a stack of point solutions ("systems orchestration"). And a client needs to decide its worth so that a CFO will sign off on the change ("value translation"). Ultimately, none of those three considerations are things a strategist adds after the plan is finished. They are part of the ongoing strategy. So, in this new age of mixed roles, the common denominator among those three skills isn't a tooling problem as much as a judgment problem. It's paramount to know what belongs in the architecture and what doesn't apply in terms of the strategy as well as its impact over the course of the entire client relationship rather than a single campaign or build. Conclusively, this aligns with McKinsey's research: nearly three-quarters of AI high performers report fundamentally redesigning workflows because of their AI use, up from 55 percent the year before, compared with one-quarter of other respondents. In my view, that's architecture work, done at the strategy stage, not tool selection work done after the fact. And it's precisely the adoption vs. results gap that most stalled AI initiatives are stuck in.
Why this favors the agency strategist specifically
An in-house marketer sees one brand, one set of customers, one internal culture, deeply. An agency strategist sees dozens of them, across industries, budgets, and platforms, often at the same time. That cross-client exposure is exactly the raw material an AI architect needs.
I've been privy to this understanding, and the need for the additional AI conversation within the overall strategy. A strategist running engagements across ecommerce, healthcare, and B2B manufacturing clients has already seen the same pilot succeed for one client and quietly fail for another and usually knows why within the first few conversations. They've built the muscle of walking into an unfamiliar business, figuring out fast what matters to its customers, and translating that into a plan a client's leadership team will trust. That's the same muscle an AI architecture recommendation requires. Understand the business. Structure the system needs to know to align. Be honest about what's worth automating versus what still needs a person in the loop. And finally, let a model handle bid optimization across a digital campaign while keeping a person accountable for which audiences and messages it's allowed to test in the first place.
Consulting agencies are already built for the translator role that value translation requires. Strategists spend their careers explaining strategic and technical decisions to CMOs, CFOs, and boards who weren't in the room when the plan was built. Extending that same explanation to why AI belongs in one part of a solution and not another, and what that protects the client from, is a natural extension of a skill that agencies have been selling for decades. It just needs to be illustrated as part of the collective strategist teams' core competency, and highlighted when attracting new clients.
What years of pattern recognition get you that a tool doesn't
AI behaves like an almost limitless supply of one kind of intelligence: the kind that generates novel options, remixes ideas, and produces something new on demand. It resembles what psychologists call fluid intelligence. What it lacks is the other kind, the crystallized intelligence, and accumulated, hard won pattern recognition that comes from living through a few hundred campaigns, platform migrations, and client fire drills.
A generative model can spin up twenty campaign concepts in the time it takes to read this sentence. What it can't do is tell you which of those twenty concepts is the same idea that quietly failed for a similar client a few years back, or which one is about to collide with a compliance issue nobody flagged. AI on its own misses the nuance. That judgment doesn't come from a larger context window. It comes from having been in the room when it went wrong the first time, on a different account, for a different client. It comes from the strategist.
The most valuable person in an AI adoption conversation is rarely the one with the most tool “fluency.” More often, it's the one who can look at twenty AI generated options and immediately rule out the fifteen that look promising but aren't, because they've seen that pattern before, on someone else's account.
The catch: experience alone isn't the only qualification
It would be easy to overstate the case here, though, and I want to be honest about that. Tenure alone doesn't make someone an AI architect any more than owning a toolbox makes someone a builder. The AI architect skillset, reasoning about how systems fail, evaluating build versus buy tradeoffs, and understanding what governance and data handling require, is not only about knowing enough about the underlying technology, but also about knowing the right skeptical questions to ask. It is the new literacy. It must be learned deliberately. A strategist who spent two decades delivering great campaigns or great platform launches, but never engaging with how these systems work, isn't an AI architect, but rather an experienced person standing near the AI conversation.
The strategists positioned to lead right now are doing both things at once. They're layering real technical and systems literacy on top of decades of cross client pattern recognition, rather than treating one as a replacement for the other. That combination, judgment plus fluency, is rare. Rare on purpose too, because it takes real effort to build on top of an already full plate of client work.
What this means if you're trying to catch up fast
For a business evaluating how to move on AI, the practical implication is straightforward, even if it's not the answer most companies expect. Don't start by shopping for tools. Don't assume the newest hire with the deepest prompt library is the right person to lead the effort, either. Start instead by identifying who, internally or at a consulting partner, has both halves of the equation, someone who has seen enough marketing and digital programs succeed and fail across enough different businesses to know which problems are worth solving, and who has done the work to understand how AI systems are built, governed, and evaluated well enough to design around them responsibly.
That pairing is what turns "we adopted AI" into "we got faster and better without getting reckless." It's also how we think about building strategy teams here at GRAYBOX. Not by chasing every new tool, but by putting experienced, cross industry judgment in the seat that's recommending the solution for a client.
The businesses that catch up fastest won't be the ones that moved first. They'll be the ones that put the right strategist in charge of recommending what to move on.