Fix the Real AI Adoption Pitfalls - Hosted with Clarinet

Why Adoption Stalls on Ownership, Not the Model

AI spending is not the problem. McKinsey's most recent global AI survey found that nearly two-thirds of organizations have not yet begun scaling AI across the enterprise, even though generative AI use has become nearly universal (McKinsey, November 2025). Gartner's numbers are more blunt: 88 percent of HR leaders say their organizations have not realized significant business value from AI tools (Gartner, October 2025). And MIT's widely cited GenAI Divide research found that 95 percent of generative AI pilots deliver no measurable profit and loss impact at all (MIT NANDA, reported by Fortune).

For HR and People leaders, this tension lands differently than it does for the rest of the business. HR is usually carrying two AI jobs at once: rolling AI into its own function while also being asked to support adoption everywhere else. That dual burden, and what actually separates organizations making progress from those stuck in pilot mode, was the focus of a recent Achieve webcast built around new research from Clarinet, the AI adoption firm that led the session. Their data lines up with the broader research almost exactly. Of the leaders who named a specific reason their AI rollout fell short, the most common answer was not model quality or integration. It was readiness: skills, training, and strategy (Clarinet, State of AI in the People Function, June 2026).

Ownership Is the Ambiguity Nobody Has Solved

If there is one root cause underneath the value gap, it is ownership. SHRM's 2026 State of AI in HR report found that 52 percent of organizations do not involve HR in AI strategy or vision at all (SHRM, April 2026). Gartner's research points to the same gap from a different angle: AI deployment decisions are frequently made without any involvement from HR, according to Gartner's Eser Rizagolu (Gartner, December 2025).

Clarinet's session offered a useful reframe for HR leaders trying to close that gap. Treating AI ownership as a single job, one name, one new title, is a common failure mode. Ownership actually spans distinct layers: setting the organization's direction (senior leadership), translating that into role-specific expectations (managers), building AI internally (engineering), and developing employee fluency (often learning and development). The goal is not one name across every category. It is a name for every category. The long-term destination, per the framework, is that AI ownership eventually dissolves into how each function already works, the same way no company today employs a dedicated head of internet.

Policy Without Guidelines Is Failing, and Shadow AI Is Filling the Gap

Roughly half of organizations now have some form of AI use policy, but SHRM found only about a quarter of HR professionals believe their policy is future-proof (SHRM, April 2026). Gartner adds a related data point: only 7 percent of organizations give employees any guidance on what to actually do with the time AI frees up (Gartner, October 2025). Employees are not waiting for that guidance to arrive. Microsoft's Work Trend Index puts unsanctioned "shadow AI" use at 82 percent of employees (Microsoft, February 2026), and MIT's research describes a shadow AI economy thriving inside the large majority of firms even where official pilots have stalled (MIT NANDA, via Fortune).

Clarinet's framing for why this keeps happening is simple and sticks: a policy is like a seat belt, it exists to prevent the worst outcomes and is usually written for legal and liability reasons. Guidelines are like GPS, they tell people specifically how and where the organization wants them using AI day to day. Roll out a policy without guidelines, and leaders end up asking employees to embrace AI while handing them a document that mostly lists what not to do. Pairing every policy with a practical, specific set of guidelines is one of the more actionable fixes leaders can make this quarter.

Managers Are the Missing Multiplier

Even where AI use is real, most organizations have not equipped the people managing it. Only 8 percent of HR leaders believe their managers currently have the skills to use AI effectively, and just 14 percent of organizations provide managers with any support for integrating generative AI into daily work (Gartner, October 2025). That gap shows up in a very specific, very common complaint: AI slop, the generic, low-quality output that lands on a manager's desk and gets waved off with "it looks like you used AI for that."

Clarinet's reframe, drawn directly from the session, is worth repeating for any manager effectiveness program: the problem is almost never that someone used AI. The problem is that the output was low quality, usually the result of a vague, context-free prompt. The fix is teaching people what a good output actually looks like, which is a coaching and performance management skill, not a technology skill. That single distinction is doing a lot of the work in organizations that are actually moving the needle on AI employee engagement, versus those generating more frustration than value.

Depth of Use Beats Frequency of Use

Daily AI use is not the same thing as AI proficiency. Clarinet's own fluency assessment benchmark found that roughly two-thirds of employees fall short of what the firm considers AI proficient, even though many of those same people use AI every day. The pattern lines up with McKinsey's finding that today's AI high performers, the organizations actually capturing measurable financial value, are far more likely than typical organizations to have redesigned workflows around AI rather than layering AI onto existing habits (McKinsey, 2025).

In practice, this means most employees are stuck in what Clarinet calls ad hoc AI use, starting from scratch on every prompt instead of building durable, reusable context through saved projects and skills. Moving organizations along the curve, from AI curious toward AI enabled, AI first, and eventually AI native, is less about buying a better tool and more about building that muscle deliberately, function by function.

What This Looked Like in the Room

These findings anchored a recent Achieve webcast, hosted by Zech Dahms and led by Diane Sadowski-Joseph, Co-Founder and Head of Product at Clarinet. The session opened with a live pulse check on how far attendees felt AI adoption had actually progressed inside their own organizations, and the range of answers made the readiness gap immediately visible to the room. From there, Sadowski-Joseph walked People and HR leaders through a practical sequence: how to set a shared vision that has both leadership mandate and grassroots buy-in, how to assign ownership across the layers described above rather than to a single overwhelmed person, and how to spot and correct the specific failure modes, from policy-guideline confusion to tool fixation, that quietly stall progress.

The format stayed hands-on throughout. Attendees worked through a live exercise for envisioning what an AI-native version of their own function might look like, fielded real-time polls on where ownership actually sits in their organizations today, and asked questions that shaped where the session spent its time, from measuring AI's real impact to handling AI slop as a performance conversation rather than a technology complaint. It is one example of the kind of practical, research-backed programming Achieve runs regularly for HR and People leaders working through exactly these questions.

Where This Leaves HR and People Leaders

The research is remarkably consistent across McKinsey, Gartner, SHRM, MIT, Microsoft, and Clarinet's own data: AI adoption problems are almost always human problems, not technical ones. Ownership left ambiguous, policy without guidelines, managers left unequipped, and proficiency mistaken for depth of use are all fixable. None of them require a better model. They require an HR function willing to lead the conversation instead of waiting to be looped in.

That is precisely the kind of leadership work the Achieve Leadership Network exists to support. Members get direct access to sessions like this one, working frameworks they can bring back to their own teams the same week, and a community of HR and People leaders solving the same problems in real time. If closing the readiness gap in your own organization feels like the next priority, learn more about joining the Achieve Leadership Network.