Why AI in Business Is About Execution, Not Tools

Access to AI is no longer what separates companies. Recent research cited in the piece shows the vast majority of organizations already use AI in at least one business function — up sharply from just a year earlier. What's much rarer is turning that access into real transformation: separate research covering thousands of employees found that while nearly all of them use AI at work, only a small fraction say it's fundamentally changed how they work, with gaps in training, integration, and organizational readiness leaving a large share of possible productivity gains on the table.

The 95% failure rate isn't a technology problem

The piece leans on an MIT study of hundreds of public AI deployments, which found that the vast majority of enterprise generative AI pilots show no measurable effect on profit or loss. The researchers' diagnosis: companies were dropping AI tools into existing workflows without changing how work actually gets done — what the study calls the "learning gap." Individuals often get real value from AI on their own tasks, but that doesn't automatically scale into organization-wide impact.

Better models aren't the fix

When AI initiatives stall, the instinct is to blame the tool and go looking for a better model. But the piece argues the real barriers are almost always organizational: unclear ownership, poor coordination across teams, weak links to daily workflows, and no clear accountability for outcomes. Pilots tend to succeed under ideal conditions — clean data, a narrow problem, a small motivated team — conditions that don't hold once AI expands across a whole organization, where outputs need to reach the right people, at the right time, with clear guidance on when to trust the system and when to override it.

AI changes who owns the decision, not just the output

A central argument here: introducing AI doesn't just speed up an existing process — it changes how decisions get made and who's accountable for them. Klarna is used as a positive example: cutting sales and marketing spend significantly in early 2024 while running more campaigns, not by simply installing a better model, but by redesigning the workflows around content production and agency management so AI fit into daily operations consistently.

On the flip side, the piece flags "automation bias" — the tendency to accept AI outputs without scrutiny — as a growing risk when decision rights aren't clearly defined upfront. Left unaddressed, errors that a human would normally catch slip through, and accountability for outcomes gets murkier over time.

Governance has to be built in, not bolted on

The piece cites survey data showing a majority of legal, compliance, and audit leaders now rank AI-related technology as their top risk concern — yet only a small minority of organizations have a comprehensive AI governance plan. Its argument: governance added after deployment doesn't build trust, it just manages damage. The organizations that scale AI responsibly bake accountability and escalation paths into the workflow from the start, and keep revisiting them as data, regulations, and business conditions shift — since AI systems can quietly degrade in accuracy over time as real-world conditions drift away from what they were trained on.

The takeaway

The piece's core claim: the differentiator isn't which AI tools a company has access to — nearly everyone has that now — but whether the organization has the discipline to redesign workflows, define ownership, and govern AI consistently at scale. Companies that start with a clearly defined business problem, rather than reaching for AI because it's available, are the ones that avoid expensive, inconclusive pilots.