As I wrote in my Substack, Your Biggest Competitor Just Launched Their AI Strategy. And That’s OK, the goal is not to be first. The goal is to be deliberate. This means building consensus before you start.

Your business unit leaders will see this as a golden opportunity. Mention an AI strategy, and everyone suddenly has an opinion. But as we have seen, cynicism and negativity quickly set in.

You saw it before. The dot-com boom. The Great Financial Recession. The crypto revolution. During these black swan events, technology was positioned as the preserver of enterprise value.

You also know that technology adoption is messy. Hype and reality rarely align. But the companies that won the day didn’t move first, they moved smart.

Build a proof of concept, for the many, not the few

Before you commit to an AI proof of concept, understand a few realities. 99% of all AI projects fail. That is why you should aim to build a POC within 30 days. It’s not arbitrary. It is a forcing function.

Anything longer than 30 days and the organisation moves on. The board gets distracted. Financial priorities change.

 This is the point at which you hire an outside consultant to build the POC. Why? Because if your organisation had the internal capabilities and spare capacity to build an AI, it would have done it already! That’s not to say it doesn’t have the capabilities or skills today - but they’re more than likely tied up.  

Your operational teams are managing other platforms: running ERP implementations or have bonuses tied to other KPIs. Your POC will compete for their time and attention.

So, ask your consultant to start small and get ready to scale. Ask them the difficult but necessary questions within S.AI.L’s AI Implementation Checklist. Your POC needs to prove AI can automate a process or provide new data-driven reports.

The shadow AI problem

You must solve the “Shadow AI” problem. 20% of organisations officially mandate AI tools, yet 80% of Gen Zs use shadow AI to enhance productivity.

Whether you like it or not, free GenAI apps are harvesting your organisation’s data. It does not matter if your employees anonymise business data. Apps like ChatGPT, Perplexity AI, Claude, Midjourney, and more, are smart.

They’re smart enough to work out which organisation the prompts are referring to based on the unique combination of industry jargon, project codenames, and technical metadata in the queries. That data can be recompiled and traced back to your company.

These apps are also popular because they learn. MIT Nanda highlighted that 98% of AI projects fail because they are boring and do not learn from their users. ChatGPT’s minimalist design is intentional. It minimises distractions so users form a psychological bond with the AI, enabling it to continuously learn. That is recursive learning.

Your employees will want to use enterprise AI apps that learn and look sophisticated. So task your consultant with building an enterprise AI app that learns, is fast, and is responsive. Get your teams to trial it, test it, and build trust that it is the solution for them. When you get it right, they will thank you. Shadow AI usage on your network will trail off.

The bottom line

You are about to make a decision about AI. Before you do, understand what you are committing to. The key is to start small, measure carefully, and scale deliberately.

Understand the political landscape, the operational constraints, the resource requirements, and the governance implications. This is not a technology decision. It is an organisational decision.

The winning companies are not those that spend the most on AI. They are those that spend smartly. They understand the organisational reality. They manage the politics, align the incentives, and invest in governance. They move deliberately.

You have lived through enough technology cycles to know this. The winners are not the first movers. They are the smart movers. So before you commit, make sure your organisation is ready.

Make sure your teams are ready. Make sure your incentives are aligned. Make sure your governance is in place. Then move forward with confidence.

We drafted an implementation checklist for evaluating AI consultants. If you’re a P&L owner, contact us for a free consultation. Send us a WhatsApp or email us at hello@sail.legal.

References

[1] S.AI.L, "AI Investment Checklist"

[2] Fortune, “MIT report: 95% of generative AI pilots at companies are failing”

[3] McKinsey & Company, “How finance teams are putting AI to work today”

[4] IT Executives Council, “AI Budgeting and Governance: How CIOs Should Plan for 2026 Investments”