If you see a company try to “ban” the use of AI, without a coherent plan or alternative, then that is a big red flag imo. It’s an indicator of potential business challenges for them in the medium term. Companies need to sign up for the future or they will become irrelevant really fast. Why not ask ChatGPT how AI could transform your business? The answers are very interesting. You don’t need a consultancy to identify opportunities, but you might need one to help you pick and d

Humans make mistakes and some humans make more mistakes than others. Where we see persistent failure, we tend to inspect more and trust less. This is a patterns that humans have learned to follow, based upon as assessment of risk through probability and impact. AI agents are not perfect, either, and they can make mistakes. They are an unproven risk in your workflow. You should treat them as would a new junior analyst in your team, and ensure you are checking the work. This do

This is going on in pretty much every organisation, I suspect. If your teams do upload pii data to ChatGPT, say, it’s used to train models. This means that the data is broken in chunks and used to refine models at a statistical level. Say you work at xyz plc and upload a customer list. I can not then go to the AI service and say “show me all customers for xyz plc.” (Note that the AI vendor employees can see all the data in its message form.) Ultimately, this is still a breach

If you are rolling out Copilot or ChatGPT to your company and are trying to teach everyone “prompt engineering” then you are heading for a mess. Why? Prompts are critical for accuracy of outcomes. A small, human, variation in prompts will produce huge variations in quality and outcome. This is just a non-starter for structured workflows. Effective prompts and instructions sets are non trivial; some of our scenarios run into 100s of lines. How do you audit all the prompt

Adoption is the number one item in corporate AI projects. By miles. People will only start to use new tools if they understand why, know it’s not a threat and if they are fit for purpose. I would not punish/reward usage in the way being suggested in this article. If the tools are no good then people won't use them. Instead, perhaps punish the people serving up the tools: they should link usage KPIs to the promotion of the business system tool teams. Employee KPIs should

When the Internet/SaaS took over the world, people had to create or repurpose the budgets for tooling. IT spend is still very low in most companies, perhaps <2% of revenue. The global IT market that spooled out from that is very large indeed (say $5T in 2025 according to Gartner). Consider the four following ideas: 1. With AI automating tasks or outcomes, the budget allocated will likely be re-purposed from the labour cost, which will by closer to 30-50% of revenue. 2. AI en

Adoption is the number one item in corporate AI projects. By miles. People will only start to use new tools if they understand why, know it’s not a threat and if they are fit for purpose. I would not punish/reward usage in the way being suggested in this article. If the tools are no good then people won't use them. Instead, perhaps punish the people serving up the tools: they should link usage KPIs to the promotion of the business system tool teams. Employee KPIs should be cu

Around the year 2000 (ish) lots of people said these words: “No-one will buy a t-shirt on the Internet.” I agreed with this. You can’t try it on and returning it was (is) a pain. At that time I thought people might buy a known commodity, like a bottle of Jack Daniels, but the cost and wait time would be too high. Now I mostly buy clothes online. ASOS made a business out of it, along with many other retail channels. My weekly shop is all online. I can even do panic sho

It is the story of new technology since the beginning of technology. You can’t make a cynic an early adopter. Trying to change their mindset is a waste of time. I show prospects the power of AI and some run towards the possibilities and others run away from them. I think the law of diffusion is a useful guide here. Perhaps 15% of any audience is an early adopter. Focus all the attention there. https://lnkd.in/e9isKbT8

If you are rolling out Copilot or ChatGPT to your company and are trying to teach everyone “prompt engineering” then you are heading for a mess. Why? Prompts are critical for accuracy of outcomes. A small, human, variation in prompts will produce huge variations in quality and outcome. This is just a non-starter for structured workflows. Effective prompts and instructions sets are non trivial; some of our scenarios run into 100s of lines. How do you audit all the prompt/outc

Let me know if this sounds familiar pattern for teams trying to drive AI adoption in their business. The Board says “we need to get AI done” so you decided to create an AI committee. You bring in all departments and meet every month to steer the plan. You set yourselves some goals to make an impact in the next six months. The committee meets and starts the journey. After six months nothing has changed. Committee members spend the hour before the meeting doing their tasks and