A recent MIT report shows that 95% – you read that right: Ninety-Five Percent! – of corporate genAI projects fail to move the needle.

My take?

This is consistent with my last 15+ years' experience in the data field, across big data, predictive analytics, data science, ML, and now genAI:

Companies are often in a hurry to Just Do Something™ with the exciting new technology, so they are less eager to slow down and be strategic about it. Corporate FOMO drives leaders to push hard for genAI and they lose focus on the real goal. The end-result is that they wind up launching Something™ ... but that doesn't go far … and they wind up on the wrong side of this 95% statistic.

Do you want your AI projects to be in that winning 5%? There's no guarantee of success in AI, but these five steps should strongly increase your chances:

1/ Get the leadership team up to speed on AI. Execs who understand what AI can(not) do are in a much better position to decide how the company will use it.

2/ Make a plan. Not only will this add structure to your efforts, it will also tell you when things are going awry.

3/ Get your house in order. All that fancy AI relies on clean, correct data supplied by a solid infrastructure.

4/ Involve your legal and PR teams early. Lawsuits due to allegedly purloined training data can be costly. The same goes for a wave of bad press due to malfunctioning AI bots.

5/ Work with a professional. Someone who has end-to-end experience with AI can dramatically reduce your risk and improve your chances of success.

(If you're a longtime reader and these steps look familiar … yes. These are the same steps I post every few months.)

If you'd like help with making it into that winning 5%, please reach out.