The Reason 95% of AI Pilots Fail

Hello,

This is Simon with the latest edition of The Weekly. In these updates, I share key AI related stories from this week's news, list upcoming events, and share any longer form articles posted on the website.

Following up on last week's edition, where I spoke about a potential leak in the AI bubble, a lot is being said about the numbers in the MIT report stating that around 95% of AI projects fail. In many ways, I'm not surprised by this at all. With the boom in Generative AI and the mass availability of LLMs, everyone now has a real opportunity to build something with AI. The problem for me is what exactly are people creating. In the early days of this newsletter, I spoke a lot about the fact that AI is not just Gen AI and LLMs.

I focused on this in my very first post:

Traditional AI has been around for a few decades and is being used successfully across many enterprise sectors. I've spoken about using Machine Learning in finance to identify fraud, price insurance quotes, and in retail to predict purchase behaviour and manage stock levels. These are all excellent examples of how AI is delivering efficient and profitable solutions for companies, and only a few individuals had the skills to build and deploy them. With the current tools available now however, everyone is a data scientist and can create solutions way beyond what they would have previously been able to. And this is a problem with Gen AI, there are many uses for it, but quite often it seems like a square peg in a round hole. CEOs are demanding their organisations use it because they don't want to get left behind. Companies are signing up to enterprise deals with OpenAI, Google and Microsoft to use their tools across the entire organisation, which is all fantastic.

But we need to take a step back. As with any project, you must identify the challenge and assess the available solutions. Gen AI may or may not be the answer. I suspect that the 95% failure rate is due to teams and departments chucking a whole range of ideas into the mix, trying to take advantage of the technology to fix something. There will be a lot of square pegs in round holes.

I created a free Notion document that discusses the Seven Steps to a Great AI Strategy I created in January this year, and even with the rapid pace of development in the AI sector, I still stand by everything written in that document. I'd advise you to go and read it, but as a quick summary, the seven steps are:

Adhere to these steps, and the AI failure rate will drop significantly.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI

Anthropic released a browser-embedded AI assistant named Claude for Chrome. Currently, it's in testing with 1,000 users under Anthropic’s Max subscription plan (~$100–200/month).

Why it matters: This is a real-world example of agentic AI making its way into everyday tools—imagine having a smart assistant right in your browser.

In a recent interview, David Luan of Amazon’s AGI Lab explained that Amazon is prioritising reliable AI agents over just language models. They’re building “gyms”—simulated environments—where agents learn through trial and error in domains like CAD and healthcare. Amazon is positioning these agents as the future of computing, not mere tools.

Why it matters: The framing of agents as the next “atomic unit” of computation makes this shift both visionary and tangible.

A new report projects that the autonomous AI and agents market will leap from $7.9B in 2024 to $11.5B in 2025, with an astonishing climb to $477.8B by 2035. That equates to a compound annual growth rate of 45.2%.openPR.com

Why it matters: These numbers show we're not just dabbling in agents, we're investing in them massively, across industries.

Thanks for reading, and see you next Friday.

Simon,

Was this email shared with you? If so subscribe here to get your own edition every Friday.

Enjoying Plain AI? Share it and get a free gift!

If you find this newsletter useful, why not share it with a couple of friends or colleagues who would also benefit? As a special thank you, when two people subscribe using your unique referral link, I’ll send you my "Exclusive Guide: Supercharge Your Work with NotebookLM." It’s a practical, no-nonsense guide to help you turn information overload into your secret weapon.

You can also follow me on social media: