GenAI's Playbook Paradox

Most organisations are struggling with generative AI (GenAI) both from a governance as well as an ROI perspective. One example: we have deployed Copilot or ChatGPT, but how is this unlocking greater efficiency? Email summaries do not cut it—particularly when these same tools are also being used to generate lengthy AI emails. Here an AI is creating more work for another AI!

As a result, broader conversations around organisational AI transformation are being slow-walked until there is better line of sight to value. People want a playbook before they stick their neck out—particularly at a board level. Yet, there is a paradox here, in that so much of the hype around GenAI relates to the opportunity to disrupt industries. By definition, when playbooks have been written, published, and established, this window of opportunity is closed.

This is an issue that private equity in particular is struggling with at the moment. Go-to-market (increasing efficiency) and technology (driving cost-out) are well-trodden areas. Meanwhile, there is far less clarity when it comes to AI. I spoke to one operating partner who confided that a majority of their AI work was related to explaining ChatGPT to portco boards. While that matters too, it is not going to drive enterprise value in the way that core process transformation will.

So how should organisations resolve this paradox?

First, the absence of a playbook does not mean an absence of common sense. While a lot of money is being poured into use cases exercises, it often does not take Ivy League consultants to figure out what is broken. Walk into a customer care centre and agents will talk your ear off about manual case reviews. Talk to merchandising teams and they will vent about the poor quality of product data. Sit with creative teams and witness how much effort goes into storyboarding and creative testing. The list goes on, and in all these areas GenAI has a role to play.

Second, like with many new technologies, experimentation is key to success. Still, in many organisations, running an experiment only means allocating budget. The rest of the project governance remains in place. In order to be successful, experiments need to have far greater degrees of freedom—from architecture reviews, procurement exercises, and certainly the usual launch criteria. Trying something new has to come with the commensurate risk appetite.

Third, while playbooks or best practices might be absent, companies can take a leaf out of other industries' books. For example, tech-forward SaaS players can teach retailers a lot about the role of AI in orchestrating real-time customer journeys. Every organisation claims that they are different, but as Kennedy observed: "What unites us is greater than what divides us." So in business too.

Perhaps most importantly, organisations need to realise that everyone is trying to figure this out. Do not let the LinkedIn bluster fool you—Dunning-Kruger applies to all these newly minted GenAI experts as well. Behind closed doors, I have not heard anyone claim AI is a "solved issue". And yet, we need to figure this out. As Robert Frost once wrote: "He says the best way out is always through. / And I can agree to that, or in so far / As that I can see no way out but through." None of us can avoid artificial intelligence's impact, so organisations need to forge ahead.

—Ryan

Image by ChatGPT.