Waiting Is an AI Decision Too

Why practical, hands-on AI coaching helps teams start safely and create measurable value

I have a growing frustration with AI decisions.

The technology moves faster every month, yet decisions inside many companies seem to take longer and longer.

There is interest. There are meetings. People test ChatGPT, Claude, Gemini, Copilot, and other tools. Leaders know they should do something. Still, the decision is postponed.

I understand the hesitation. Companies have budgets to protect, people to support, data to secure, and regulations to respect. But waiting is not a neutral position. It is also a decision, with its own costs and risks.

The problem is no longer awareness

AI adoption is rising, but practical use remains uneven. Eurostat reports that 19.95% of EU enterprises used AI technologies in 2025, compared with 55.03% of large enterprises. The gap is not simply about interest. Smaller organisations often face tighter budgets and less access to the right expertise. Eurostat

The OECD reports a similar pattern. Across OECD countries, 40% of large firms used AI, compared with 11.9% of small firms. It also found that 50% of SMEs surveyed across four G7 countries said their employees lacked the skills to use generative AI. Finance, skills, data, and access to practical guidance all affect adoption. OECD

This helps explain why so many discussions stall. The question is too often framed as:

“Which AI system should we invest in?”

A better first question is:

“Which part of our work should we improve first, and how will we measure the result?”

Why are AI decisions taking so long?

Five concerns seem to explain much of the delay. Current research points to a sixth.

1. “We do not have the budget”

Sometimes there really is no available budget. But often, “no budget” means that nobody has yet built a clear, credible business case.

A company does not need to begin with a large platform, a long transformation programme, or a costly technical project. It can begin with one recurring task, one team, and one measurable result.

How long does the task take today? How often is it repeated? What level of quality is required? What must a person still check? What would a useful improvement be worth?

Those questions turn AI from an expense into a testable business decision.

2. “AI changes too quickly. What if we invest in the wrong tool?”

This fear is reasonable. Models, features, and products change constantly.

That is exactly why I do not build coaching around one button or one product. I teach working methods that remain useful across tools: giving clear context, defining the expected output, working with reliable sources, checking results, protecting sensitive information, and fitting AI into a real process.

The durable investment is not knowledge of this month’s interface. It is the team’s ability to think and work effectively with AI.

3. “What will AI do to people’s jobs?”

People can be curious about AI and worried about it at the same time. Ignoring that fear does not remove it. It often pushes AI use into private, inconsistent experiments without clear rules.

Practical coaching makes the discussion concrete. We examine tasks, not job titles. We identify repetitive work, research, drafting, comparison, preparation, and administration that AI may support. We also define where experience, responsibility, empathy, approval, and human judgement must remain.

McKinsey found that formal training can reduce anxiety and increase confidence and usage. This is why involving employees is more useful than presenting AI as a decision already made for them. McKinsey

4. “We do not know where to start”

Do not start with the full list of things AI can do. Start with the work people already do.

Look for recurring tasks that consume time, create frustration, slow customers down, or produce inconsistent results. Then score possible use cases by value, feasibility, user readiness, risk, and the effort needed to adopt them.

This removes much of the noise. The goal is not to use AI everywhere. The goal is to use it where it makes work meaningfully better.

5. “We do not know how to start”

This is where a hands-on AI coach is useful.

My sessions are not technical lectures. People work on their own tasks, with their own examples, in language they understand. We test, compare, correct, and improve together. They leave with something they can use, a clear method they can repeat, and a better understanding of what AI can and cannot do.

The feedback I hear repeatedly is simple: the sessions are practical, easy to understand, and useful from the first meeting. People tell me they learn a lot from the first session. Many say they wish they had started earlier.

6. “What about privacy, accuracy, and regulation?”

This concern is supported by the research. Deloitte found that regulatory compliance had become the leading barrier to developing and deploying generative-AI tools among its surveyed organisations, rising from 28% to 38% across four survey waves. Deloitte

The answer is not to ignore the risks or wait for every question to disappear. A practical first step can define which tools are approved, which information may be used, what needs verification, who approves the output, and when legal, compliance, security, or technical specialists must be involved.

I help business teams apply these rules in their daily work. I do not replace those specialists.

What practical AI coaching can change

I keep client details confidential, but I can share several anonymised examples from my work.

In one commercial workflow, preparing a proposal took around two hours. An AI-supported process reduced the initial preparation to approximately 15 seconds, followed by about five minutes of human review. The important result was not the speed alone. The process kept a person responsible for checking and approving the final work.

In a customer-service setting, clearer workflows and improved Zendesk macros helped move customer satisfaction from approximately 75% to 81% to around 90% within one week.

In a fashion-related workflow, a practical AI-supported process produced an estimated annual saving of approximately 600 hours.

I have also coached management and departments across organisations of approximately 100 to 200 people. The work has covered marketing, CRM, e-commerce, HR, finance, customer service, buying, allocation, styling, modelling, and IT.

These examples differ, but the method is consistent:

1. Understand how the work is done today.
2. Select a useful and realistic first case.
3. Test it with the people who do the work.
4. Keep human checks where they matter.
5. Measure time, quality, service, or commercial impact.
6. Document what works so the team can repeat it.

Buying an AI licence is not the same as adopting AI

Many companies already pay for AI tools but have not changed how work gets done.

PwC’s 2026 Global CEO Survey found that 56% of CEOs had seen neither higher revenue nor lower costs from AI during the previous 12 months. Only 12% reported both. PwC links stronger results to clearer road maps, responsible-AI processes, suitable technology foundations, and a culture that supports adoption. PwC

McKinsey’s research reaches a related conclusion. Redesigning workflows had the strongest effect among the organisational practices it tested for reported bottom-line impact from generative AI. Yet only 21% of respondents using generative AI said their organisation had fundamentally redesigned at least some workflows. McKinsey

The tool is rarely the whole problem. The missing parts are often use-case selection, workflow design, clear rules, employee confidence, follow-up, and measurement.

Why my sessions are deliberately non-technical

I am not there to turn business teams into engineers.

I have more than 30 years of experience in sales, marketing, and client-facing business work, and I have coached leaders and departments on AI since 2022. I translate AI into the daily reality of sales, marketing, customer care, finance, HR, e-commerce, operations, and management. I use plain language. I demonstrate the work. Then people practise on cases that matter to them.

Where useful, I can turn business requirements into working prototypes with tools such as Codex, automation platforms, and low-code systems. For production architecture, security, engineering, legal, or compliance decisions, I work alongside the relevant specialists. My role is to connect business needs, people, responsible use, and measurable results.

This approach gives leaders evidence before they commit to something larger. It also gives employees a voice in how AI is introduced.

Not investing in AI is also an investment decision

Delaying a large, unclear AI project can be sensible. Delaying all structured learning is different.

During that delay, employees may continue doing repetitive work manually. Others may use public AI tools without shared rules. Teams may develop inconsistent methods. Managers gain little evidence about what works. Competitors and colleagues keep learning.

Deloitte found that organisational change is moving far more slowly than the technology. More than two-thirds of surveyed leaders expected 30% or fewer of their generative-AI experiments to be fully scaled within three to six months. Regulation and risk were major barriers, but the most advanced initiatives were still producing positive returns for many respondents. Deloitte

The answer is not reckless spending. It is disciplined progress.

Start small. Use real work. Set boundaries. Keep people involved. Measure the result. Continue only when the evidence supports it.

The first step should create clarity, not commitment

A first coaching session does not need to lead to a large contract or a technology purchase.

It should help a leader or team answer practical questions:

• Where are we losing time today?
• Which task is suitable for a first AI test?
• What information can and cannot be used?
• What must a person verify?
• How will we measure improvement?
• What should we stop, continue, or test next?

That is how hesitation becomes informed action.

AI will keep changing. Waiting for it to stop changing is not a strategy. Building the ability to test it safely, use it well, and judge its business value is.

If your team is interested in AI but unsure where or how to start, we can use a free 15-minute conversation to identify one realistic first step: https://bit.ly/Jax15m