The Goldilocks of AI Adoption

In 1837, Robert Southey penned a tale that would become a bedtime staple for generations to come. “The Story of the Three Bears,” better known as “Goldilocks and the Three Bears,” tells of a young girl’s adventure in finding just the right porridge, chair, and bed.

But what does this children’s story have to do with Generative AI adoption? Well, everything really.

In the world of Generative AI adoption, many organisations are acting like Goldilocks, sampling every AI “porridge” they can get their hands on. Some are too hot (overly ambitious), some are too cold (overly cautious), but the wise ones are looking for that Generative AI project that’s just right.

So, how do we find that “just right” first Generative AI deployment? Let’s break it down, Goldilocks style.

Too Hot: The Customer-Facing Extravaganza

Let’s take a likely real-life scenario. Your company decides to go all-in on Generative AI, choosing to deploy a cutting-edge chatbot to handle all customer inquiries.

Sounds impressive, right? But like Papa Bear’s too-hot porridge, this approach can burn you.

It’s a high-risk strategy, potentially damaging to your brand and your customer relationships if things go wrong. It leaves little room for error or learning curves.

Then you will need to decide if RAG (retrieval augmented generation) alone gives you the desired outcome or do you need to go through the time-consuming and often expensive route of fine tuning your generation model.

Too Cold: The “Wait and See” Approach

On the flip side, some companies are so wary of Generative AI that they’re practically hibernating. They’re watching from the sidelines, waiting for others to make all the mistakes first. I have often referred to this as FOMF, or fear of moving forward.

This will create problems too. Competitors will be pushing forward and soon customers will abandon in favour of a higher tech solution. Similarly, access to Generative AI assist tools will rapidly become a hygiene factor for new staff considering joining an organisation.

Remember, Goldilocks didn’t find the right bed by standing in the doorway.

Just Right: The Internal, Single-Use Case

Here’s where our Goldilocks wisdom comes in. The “just right” approach to Generative AI adoption is starting with a single, internal use case.

This approach tends to be low risk, high reward. You’re not putting customer relationships on the line, but you’re still reaping the benefits of Generative AI. A great example is putting a bot in front of the mountain of HR or IT policies almost every organisation on the planet has. New staff onboarding becomes a whole lot easier.

Although the temptation is to jump into 10 use cases, a single use case allows for focused effort and clear metrics, as well as a chance for your team to get hands-on experience with Generative AI in a controlled environment.

Goldilocks in action

Remember, Goldilocks didn’t start by redecorating the entire bear house. She tested one thing at a time until she found what worked.

Here are my top three tips for finding your “just right” Generative AI adoption.

Firstly, start small. Choose a use case that’s important but not critical. Think “helpful” rather than “revolutionary.”

Secondly, keep it in-house. Your first Generative AI project should be for internal use only. Save the customer-facing stuff for when you’ve got some experience under your belt.

Finally, measure, learn, repeat. Set clear metrics for success and don’t be afraid to adjust as you go.

So, there you have it. The Goldilocks approach to Generative AI adoption. It’s not about being the biggest or the boldest, but about finding that sweet spot that’s just right for your organisation.

Remember, even Goldilocks had to try a few things before she found her perfect fit. So go ahead, take that first bite of AI porridge.

Just make sure it’s not too hot, not too cold, but just right.