The question of AI adoption often feels like a modern version of an age-old dilemma: what comes first, the cart or the horse?

I was speaking with a client recently who was keen to adopt AI to improve organisational efficiency. As part of a proof of concept (POC), I carried out an operational analysis and uncovered a significant amount of waste embedded in their day-to-day processes. This waste translated directly into excess cost. Costs that had quietly become “business as usual”.

When I broke down these inefficiencies and mapped them against potential AI-driven optimisations, the picture became very clear. Yes, implementing AI solutions would require investment. But that investment was materially lower than the ongoing cost of inefficiency the organisation was already absorbing every single day.

The executive team understood this immediately. They were enthusiastic, aligned, and keen to embed AI tools into the organisation.

The challenge, however, emerged elsewhere.

Executive Enthusiasm vs Team Resistance

While leadership saw AI as a lever for efficiency and competitiveness, the existing team of engineers and analysts were far more sceptical. They raised a number of objections. Some reasonable, some less so. Common concerns included:

• “Our data isn’t ready”
• “The models won’t be accurate”
• “We need to clean everything first”
• “AI will introduce risk”
• “This isn’t the right time”

These hesitations are not without merit. Data quality matters. Governance matters. Context matters.

But there is a critical distinction between valid concerns and paralysing excuses.

Using these concerns to justify a complete lack of AI adoption does not protect the organisation, it actively harms it. Every delay represents:

• Lost competitive advantage
• Continued unnecessary operational costs
• Missed opportunities for learning and capability building

In short, it does nothing to improve the organisation.

Kicking the Can Down the Road

In the short term, teams may succeed in pushing back on AI adoption. They may delay pilots, slow decision-making, or block implementation altogether.

But this is simply kicking the can down the road.

AI is not a passing trend. It is becoming a foundational capability across industries. Refusing to engage with it does not stop its progression, it only widens the gap between organisations that adapt and those that fall behind.

From an individual perspective, the same principle applies. It is far better to adapt skills to AI-enabled ways of working than to refuse the technology or feel threatened by it. The professionals who thrive will be those who learn how to work with AI, not against it.

So What Comes First: Buy-In or the Tool?

This brings us back to the original question.

Does team acknowledgement and acceptance come first, or does the implementation of the tool come first?

In reality, waiting for perfect alignment before taking action rarely works. Cultural change often follows practical exposure. Teams need to use the tool to understand its value, limitations, and implications.

Adoption does not mean blind rollout. It means:

• Starting with constrained, well-defined use cases
• Embedding guardrails and governance
• Learning through iteration
• Improving data incrementally rather than waiting for perfection

The organisation must move in the interest of its long-term success, even if that means initial discomfort.

AI Is Not a Replacement for Thinking

One important misconception is that AI removes the need for human judgment. In practice, the opposite is true.

Context and human guidance are still essential.

In traditional teams, a manager might assign a task without providing much context, relying on shared understanding or informal clarification. With AI, this approach fails. Large language models and agents are highly sensitive to the quality of instruction they receive.

Without the right context and clear prompts, the outcome is often hallucination, irrelevance, or incorrect results.

AI forces organisations to become better at something they have often neglected: clear thinking and clear communication.

Garbage In, Garbage Out—Amplified

The principle of “garbage in, garbage out” has always applied in technology, but AI amplifies it. If incomplete, ambiguous, or incorrect information is provided to an AI agent, the output becomes a game of chance rather than a reliable result.

This is not a weakness, it is a mirror.

AI exposes unclear processes, fuzzy requirements, and poor-quality data that already exist. In doing so, it creates an opportunity to fix them.

The Real Opportunity

AI adoption is not just about automation or cost reduction. It is about:

• Improving operational efficiency
• Raising the quality of decision-making
• Forcing clarity of intent and instruction
• Enabling teams to focus on higher-value work

Organisations that recognise this and act decisively will move faster, learn faster, and outperform those that hesitate.

The question is no longer whether AI will be adopted but who will adapt early enough to benefit from it.