Rolling out AI? Hope is not a strategy

A company I know recently ‘rolled out’ Microsoft CoPilot to their team. But there was no training. No strategy. No guidelines. Just an email effectively saying, ‘you now have AI’.

This left me scratching my head. I’m not quite sure where to begin addressing the risks, the tasks that have been overlooked and the potential consequences.

One of my favourite phrases is ‘hope is not a strategy’. That very much applies here. Arming a workforce with powerful tools and leaving them to it in the hope that it’ll work out, is not a strategy.

Here’s what’s likely to happen:

• Nothing (actually one of the safer outcomes)

• Improper use (ranging from the innocuous to the indecent)

• Ineffective use (not identifying the scenarios where it can most add value)

• Doing more, but doing it worse (churning out ‘AI slop’ at a rate of knots)

• Using it brilliantly (but being in the minority and working in a silo with no framework for sharing best practice)

• Using it invisibly, leaving leadership with no idea of uptake or RoI

A better way might include:

• Start with the problem, not the tool: Identify processes that are frequent, annoying, intensive and repeatable

• Set governance and guardrails first: Establish compliance, security and data privacy rules before the first prompt is written.

• Define clear use cases and guidelines: Show leadership

• Educate your team (properly - and also allay any fears)

• Measure (decide what metrics matter to you) and iterate

If you think ‘this sounds like a lot of work’, you're right. It is. But it's important work that separates companies that will genuinely benefit from AI, from those that just have expensive subscriptions sitting idle.