The Real Question: Why Teams Don’t Change, Even When the Tool Is There
The most common thing founders say when they describe their AI situation is some version of this: “Some people use it, most don’t, and nothing has really changed.” The tools exist. Subscriptions have been paid. A memo may have gone out. And yet the company doesn’t work any differently.
The reason is almost always the same: AI was introduced as a thing to adopt, rather than something embedded in the work people are already doing. That distinction sounds small. It isn’t. Teams change how they work when AI enters the sequence they already follow—step A, then step B, then step C—and helps right there, in the step they’re already taking. They don’t change because they were told to use AI more, any more than they’d change their route to work because someone showed them a map.
This is what makes AI adoption genuinely different from deploying any other business software. With a CRM or an accounting tool, the process changes by definition—you have to log in, fill in fields, run reports. With AI, the tool is often so flexible that it sits beside the existing workflow without touching it. The person writes the email the old way, then uses the AI on the side for something else. Nothing shifts. The adoption number stays low; the operational result stays the same.
The practical implication: add AI between steps, not around them. If a sales rep’s existing sequence is (1) qualify the lead, (2) research the company, (3) draft the outreach, add AI to step two—connected to the tools already open, returning output in the format already used. The rep feels the benefit without being asked to learn a new process. That’s adoption. It follows from benefit felt inside the existing flow, not from a company-wide instruction to use AI more.
The word to avoid is revolution. The moment an AI rollout is framed as a transformation, a rethink, a new way of working, the organization starts generating noise—anxiety about roles, confusion about scope, resistance to the whole initiative. The job is quieter and more precise: map where AI fits and where it doesn’t, without disrupting the rhythm of how the team operates. Empower, add value, elevate. Keep the rest of the flow intact. That approach keeps teams on board and keeps the organization focused on work instead of change management.
What an AI Adoption Framework Actually Is
An AI adoption framework is a repeatable process for deciding which workflows to change, in which order, and how to know whether the change worked. Without one, what happens instead is tool sprawl: scattered individual use, disconnected workflows, and spend that doesn’t change how the company operates. Someone uses one AI for drafting, someone else uses a different one for research, a third person hasn’t started—and the company has no way to tell whether any of it is producing a business outcome. A practical AI adoption framework has four stages, run in sequence.
Assess. Understand how work actually moves through the business today—not how it’s supposed to move, but how it does. Where does time go? Where do errors recur? Where do handoffs break down? And where are staff already using AI on their own? This last point matters more than most founders expect. According to Microsoft’s Work Trend Index, 78% of AI users at work are bringing their own AI tools outside IT approval. Those people are your first adopters and your best signal about where AI is already proving its value. The assessment surfaces that use, makes it visible, and treats it as evidence rather than a problem to contain.
Prioritize. Not every workflow that could involve AI should involve it next. Score each opportunity by four criteria: value to the business if changed, effort to implement, technical feasibility given existing systems, and risk—including the risk of getting it wrong in that particular workflow. This produces a sequenced list, not a wishlist. The first item on the list should be winnable: high enough value to prove the approach, low enough complexity to get done in weeks rather than months.
Implement and integrate. Build, configure, and connect. This is where most frameworks break down, because mapping and prioritizing are easier than actually integrating AI into existing systems and getting people to use the result in a live workflow. Implementation has to be treated as a first-class stage, not a consequence of planning.
Measure and move. Set a specific outcome metric for each workflow change—time per task, error rate, volume handled—before the change goes in. After, measure it. If the number moved in the right direction, the team can see it and trust it. If it didn’t, the framework tells you what to adjust. Then move to the next item on the prioritized list.
The framework is business-first by design. The workflow problem is defined before any tool is selected. That order matters: selecting the tool first creates pressure to find a use for it, which is how companies end up with AI that is technically impressive but doesn’t fit how the team actually works. Without a framework, AI adoption becomes tool sprawl—and tool sprawl doesn’t change how a company operates.
How to Drive AI Adoption Across the Company
The gap between having AI available and using it in a way that changes business outcomes is real and well-documented. Gallup’s May 2026 data shows that 47% of employees said their organization has integrated AI tools—but only 30% used AI a few times a week or more. Gallup calls this the integration-adoption lag, noting that implementation does not guarantee use. And even among organizations that have rolled out tools, only 25% of employees say their organization has communicated a clear plan for using AI.
The gap between available and used is not primarily a technology problem. It’s an ownership and sequence problem. Three things close it.
A workflow map. Document how work actually moves through the business today—not how it’s supposed to move, but how it does. Where does time go? Where do people do the same thing twice? Where are staff already using AI individually and quietly? The workflow map is the foundation: without it, any prioritization is a guess, and any implementation is likely to land beside the existing process rather than inside it.
Prioritized opportunities. Once the map exists, rank the opportunities it reveals. Microsoft’s Work Trend Index 2026 found that organizational factors explain 67% of reported AI impact, against 32% for individual factors. This is the evidence for starting at the workflow level, not the individual level. Early wins—visible, measurable changes in one workflow—build the credibility and the confidence that makes later changes easier to adopt across the company.
Someone who owns the agenda. This is the part most founders underestimate. Without a person accountable for the AI agenda month to month—tracking what changed, deciding what comes next, pushing the team through the friction of a new tool embedded in a live process—the rollout stalls after the first tool is installed. Someone has to hold the thread. In founder-led and operations-heavy businesses, the people who could do this are already running the company. The capacity to manage a sustained AI rollout simply doesn’t exist alongside everything else, and that’s the most honest explanation for why most AI in growing businesses stays individual, ad hoc, and invisible.
Getting Field Staff and Office Staff to Use AI
The same framework applies to both groups, but the conditions for adoption differ in ways that matter for how the rollout is configured. Field staff adopt AI when it removes friction from a task they already find annoying, in a tool they’re already holding. Dispatch scheduling, job-site paperwork, end-of-day reporting, call summaries after a client visit—these are the tasks that eat time and generate errors, and they’re the right candidates. What doesn’t work is presenting AI as a company-wide initiative to a crew that is already managing a full day on-site. The benefit has to be obvious and immediate in the specific task they’re doing, not abstract and future.
The configuration requirement is higher for field staff than for office staff. AI that requires a laptop, a login, and three steps to open will not get used in the field. AI connected to the scheduling software, the job management system, or the phone they’re already using—and that returns something useful in the format they already work in—stands a real chance of becoming part of the daily flow.
Office staff need AI embedded in tools they already open every day. A separate AI tool that requires a context switch—closing what they’re working in, opening something new, pasting text, copying output back—will be ignored, not because people are resistant, but because the friction exceeds the perceived benefit. When AI is built into the email client, the CRM, the document system, or the communication platform, using it becomes the path of least resistance rather than an additional step.
Both groups share one concern that should be addressed directly rather than avoided: the question of what AI means for their role. Fear of becoming obsolete is a real driver of AI resistance among workers. Telling the team which tasks AI is handling, and why—and being clear that those are the tasks that consumed time without adding judgment—is not just good communication. It’s how you keep the people you need on board with the change.
Why Most AI Rollouts Fail Before They Change Anything
The 2026 survey findings from Writer and Workplace Intelligence reveal that 79% of organizations face challenges in adopting AI—a double-digit increase from 2025. The number is striking, but the pattern beneath it is consistent: the same few failure modes appear again and again, regardless of company size or industry.
The tool is bought before the workflow is understood. This is the most common failure mode and the one with the most downstream consequences. Once a tool is selected, the tendency is to introduce it and expect the workflow to adjust around it. It doesn’t. Staff continue the old process; the tool sits alongside it unused. With 48% calling adoption a “massive disappointment,” the gap between strategy documents and business outcomes has never been wider.
No one owns the agenda. A rollout requires someone whose job is to hold the agenda—not just at launch, but every month after. Without that person, momentum dies after the first tool is installed. There’s no one to measure what changed, no one to push through the friction of the second workflow, and no one to course-correct when the initial configuration doesn’t fit how the team actually works.
The tool is introduced before the team understands why. When staff don’t know which tasks AI is handling and why, the natural response is to route around it—using the old process because it’s known and trusted, or in some cases actively disengaging from the rollout altogether. Research found that 29% of workers admit to sabotaging their company’s AI strategy, whether by using unauthorized tools, inputting sensitive data into public systems, or refusing to engage. This isn’t malice—it’s a rational response to a rollout that didn’t explain what was changing, or one that sent mixed signals about what AI meant for their roles.
The rollout starts with the tool instead of the problem. Picking AI tools first and then finding workflows to apply them to is the opposite of a business-first approach. It produces AI that is technically functional but mismatched to the specific way the team operates—impressive in a demo, unused in practice. In founder-led and operations-heavy businesses, there’s an additional structural reason rollouts stall: the people who could own adoption are already running the company, and the capacity to manage a sustained rollout simply doesn’t exist alongside everything else.
The Honest Alternatives — and What Each One Costs You
When a founder or CEO decides to do something serious about AI adoption in their company, there are roughly four options on the table. Each has genuine strengths and real costs.
Large consultancies offer deep resources, structured methodologies, and senior relationships. They are also built for enterprise programmes: large teams, long engagements, and deliverables measured in strategy decks. For a startup or a founder-led services business, the typical output is a report that describes what should happen, not an implementation that changes how work gets done. The cost of that mismatch is both financial and temporal.
AI freelancers are fast to engage and genuinely skilled within their domain. The limitation is scope: a freelancer handles one tool or one workflow well, but rarely owns the cross-workflow agenda or drives team adoption beyond the narrow deliverable they were hired for. When the deliverable is done, the agenda stops.
DIY AI tools are the default choice—and they’re better than they’ve ever been. The honest constraint is not the tools themselves; it’s the internal time and skill required to run a rollout at the level it actually requires. In most growing businesses, neither is available at that level. The result is the same scattered, individual use that already exists, with the addition of a few more subscriptions.
Hiring a full-time head of AI is the right answer for a company at a certain scale and pace. It’s also a large, slow bet: recruiting, onboarding, and ramp-up take months before anything changes operationally. For a business that needs things to change in the next quarter, the timeline doesn’t match the need.
None of these is a wrong answer in every situation. The question is whether the option fits the actual size, pace, and constraints of the business in front of you.
How No Black Swan Approaches AI Adoption
No Black Swan is a small AI integration agency for startups and growing businesses, run personally by founder Dori Fussmann. Every engagement is hands-on—there is no junior team the work is delegated to.
The engagement has two parts. The first is a paid setup and assessment (the first month): understand the business, map current workflows and constraints, surface how the team already uses AI, find the real opportunities, prioritize them by value, effort, feasibility, and risk, and agree a practical plan. The second is a flat monthly engagement, scaled to the project’s size and the client’s pace: build, integrate, and refine—configure and connect the chosen AI into existing systems, drive team adoption, measure what changed, then move to the next opportunity. Setup starts from $250 for the first month, and the monthly engagement typically from $2,000, scaled to scope and pace; the exact scope and fee are agreed after a free 30-minute consultation, because every business differs.
In practice, No Black Swan acts as the company’s outside AI owner. It holds the agenda month to month, tracks what changed, and decides what comes next. That role—owning the AI agenda rather than delivering a one-time project—is what separates sustained operational change from a single-tool install that gets abandoned when the initial excitement fades.
The edge is first-hand. Dori’s background spans investment banking at Citi, Morgan Stanley, and Ion Pacific; founding and operating The Vets, a technology-enabled pet healthcare company; fractional CFO work across companies including BlueMark, Natrion, Snout, and Translucent AI; and advisory roles at Ambi Robotics and Manatee. That’s an operator’s and CFO’s view of where work, cost, and margin actually sit—which is what makes the prioritization in the assessment stage genuinely practical rather than theoretical.
No Black Swan is also honest about where AI doesn’t belong. If a workflow isn’t a good fit, that’s the advice given, not a tool sold anyway. No percentage fees, no guaranteed results, no strategy decks without implementation.
What to Do Next
If your team is already using AI individually but the company isn’t AI-run, the assessment stage will show you exactly where the gap is and what to fix first. That’s the starting point for most of the businesses No Black Swan works with: AI use that is real but scattered, with no one owning the agenda and no way to know what’s actually changing.
The right next step is a free 30-minute consultation. The conversation covers your workflows, constraints, and what a practical plan would look like for your specific business. Exact scope and fee are agreed after that call—every business differs, and no fixed quotes are given in advance. If you’re curious what the assessment would surface for your team, that call is where to start. You can book it at /dori-fussmann-lqbqlb/.
Where Things Stand
AI adoption data from the closing months of 2026 shows the gap between tool access and genuine operational change widening, not closing. Goldman Sachs’ October 2026 update to its AI adoption tracker found that the Census Bureau’s Business Trends and Outlook Survey put US establishment adoption at 24.6%—up from roughly 20% earlier in the year—with adoption approaching 50% in the information, professional, and finance sectors, and expectations of further growth within six months. At the same time, Deloitte’s late-October 2026 State of AI in the Enterprise survey of 3,500 director-to-C-suite leaders across 25 countries reported that only 28% are seeing AI-driven revenue gains, up from 20% in January, while 81% report faster decision-making—suggesting that speed benefits are arriving faster than revenue impact. The consistent pattern across recent surveys is that more companies have AI tools than have changed how they work: access is spreading rapidly, but the workflow-level ownership required to turn access into operational change remains rare, and that gap is precisely where most growing businesses are stuck.