Many organisations have slipped into a peculiar form of AI paralysis. They are active enough to feel modern, but not disciplined enough to create value.
One week the team is testing ChatGPT. The next, it is Claude. Then Gemini. Then a niche specialist tool that promises to revolutionise some small corner of the workflow. Boards and directors compare benchmarks, watch demos, attend webinars and forward links to one another with a sense of productive urgency. Yet very little changes where it matters: inside the actual operating process of the business.
The AI model matters less than the workflow built around it. McKinsey’s January 2025 research found 92% of companies plan to increase AI investment, yet only 1% describe their deployment as mature. MIT Sloan found half of performance gains came from user adaptation, not the model itself. The competitive edge comes from building a repeatable system: standardised templates, approved source material, and defined quality gates — not from chasing the latest tool.
“Testing tools is sensible at the beginning. Staying in testing mode forever is just procrastination dressed up as innovation.”— Larysa Hale, Expert Circle
This is the trap of perpetual beta testing. A certain amount of experimentation is sensible. In the early stages, leaders do need to understand what suits their team. The problem begins when exploration becomes a substitute for implementation. Because if you are constantly switching tools, you are rarely building a stable workflow.
Does it matter which AI tool I use?
There is a reason tool-hopping is so seductive. New AI models arrive with fresh claims, impressive demos and the familiar promise that this version will finally make everything click. But the real question for a managing director is not, “Which is the best AI tool?” The real question is, “What problem are we trying to solve, and what process will solve it reliably?”
McKinsey’s January 2025 workplace research, based on surveys of 3,613 employees and 238 C-suite leaders, found that nearly all companies are investing in AI, yet only 1% describe themselves as mature in deployment — meaning AI is fully integrated into workflows and driving substantial business outcomes. The bigger failure is not lack of access to tools. It is failure to embed them properly into the way the organisation works.
Why user adaptation matters as much as the model
MIT Sloan, reporting in January 2026, found that only about half of the performance gains from moving to a more advanced AI model came from the model itself. The other half came from how users adapted their prompts. That means a firm cannot buy its way to superior AI performance simply by chasing the latest model.
Andrew Ng’s work points in the same practical direction. His Generative AI for Everyone and AI for Everyone courses focus not merely on prompting, but on how generative AI can be used in day-to-day work. The implication is clear: AI literacy and working patterns matter far more than endless software tourism.
“The competitive advantage does not come from chasing every new model. It comes from building a workflow that keeps producing value.”— Larysa Hale, Expert Circle
How to build a stable AI workflow: three steps
The first step is to choose for fit, not fashion. Select a primary tool based on the work it needs to support, the comfort level of the team and the kind of outputs the business actually values. Then commit to it for long enough to build competence.
The second step is to standardise the method, not merely the software. The most valuable internal asset is rarely the app itself. It is the company’s briefing library, reusable context blocks, approved source sets, review steps and quality standards.
The third step is to solve the business problem first. If the real issue is slow lead response, inconsistent proposal quality, weak market analysis or overloaded customer support, start there. Design the workflow around the friction point. Then choose the tool that best supports that workflow.
There is an old truth in business technology: an excellent operator using a good-enough system will usually outperform an undisciplined operator using the latest premium software. AI is proving much the same.
Our briefings and programme are designed for senior leaders who want to move beyond experimentation and build robust, model-agnostic AI workflows that create real business value. Because innovation is only useful once it has been turned into a system people can actually run.
Larysa Hale is the founder of Expert Circle and creator of the AI-Driven Marketing Growth Programme, a structured series of briefings and masterclasses for managing directors and senior leaders in professional services. She has spent over 15 years helping founders, marketing directors and business leaders build commercially grounded growth strategies.
