It’s Time to Experiment: Adapting to the New World of AI in Marketing
Ever hear of the change curve? It’s a model that change management experts use to help people anticipate and understand their feelings and reactions during significant change or upheaval. They’re common in industries that have undergone significant transformation, and there are several versions, but they all basically look like the one below which is from the Kübler-Ross model:
In preparing for AI at Rise Interactive, I’ve been on somewhat of a parallel track studying change management. And as I look at this model, I couldn’t help but notice that overall sentiment and conversations around AI have seem to have followed the change curve.
Which brings me to experimentation: I’ve been conducting a little experimentation of my own recently that I think explains input-ready data in relatable and practical terms. I was playing around with the Browse with Bing beta plugin for ChatGPT, which at the time (it’s no longer live) linked to Bing, Microsoft’s search engine. Here is a screen shot, which is pretty self-explanatory:
Call me impressed.
All I did here was ask ChatGPT a simple question, and pulling from the web, it came up with what was probably the best description of our company I’ve seen in a long time.
As you can see, it had bullet points. It had descriptions. It had multiple layers. It was written in human speech. The descriptions included citations, and when you clicked on them, they took you directly to the web page where the platform found the information. Immediately, I walked over to my SEO team, showing them what ChatGPT came up with and wanting to know from them:
- How do we understand how it decided where to pull the information from for citations?
- How do we set up our website so that when AI models land there, the information they pull is what we want it to be?
While they were on that, I wanted to see what else it could do. I asked ChatGPT for some competitive analysis, and without showing you what it found, let me assure you that it did a pretty darn good job here too.
Here was all this capability I unlocked with little more than a working theory of things I wanted to test and a credit card to cover the cost of a $20 monthly subscription.
What this experiment proves already
What this one simple experiment shows is that it’s essential to have website content that is updated. These AI tools will reveal exactly where they’re getting their information (like you see with SEO results), whether this information is accurate or not. If we don’t get ahead of this, what they come up with might not be what you want. And if this is the case, what does this say about your company and how you’re managing your brand?
Now it’s your turn to experiment
More than anything, I think this experiment shows that as marketers, we need to get our internal act together. Think about all the pages of your website that right now show up on page 10 of Google that could potentially be at the forefront of an AI search about your brand in the near future.
The information we get out of AI platforms will only ever be as good as the input we’re providing them, which is why we need to be shining up our input-ready data and working with—not against—these platforms in the new world of marketing. Without a strong content strategy, these AI platforms can’t read, can’t learn and can’t organize information in the way you may want them to.
We need to start asking ourselves and keep asking ourselves, “What’s the right input? What’s the right prompt?” If it’s easier, start where I did and ask your preferred AI platform the same questions about your own company.
We can continue to compare notes as we take the journey along the change curve together. If you haven’t already, let’s connect so that we can continue to learn from each other.