Get with the program: AI for behaviour change

Much is being written about AI. People in fear of being replaced by robots are quick to point out its failings—no new thinking, can’t make decisions, only tells you what you want to hear. 

But at the Foundry we believe that, rather than replacing our creative craft, AI can do some of the legwork that means we get to a higher-quality output faster.

An area of potential impact I’ve yet to see any thinking around is the use of AI for behaviour change marketing. I’m not an academic expert in this field, but I have made some observations and started using AI to help craft behavioural interventions. Here’s why:

AI is a brilliant research assistant. Where behaviour change marketing is prevalent, such as public health, there are thousands of useful documents available for review, but it can be a challenge to get up to speed on all of it. AI quickly works out what sources are and aren’t relevant to your particular challenge.

You can use it to summarise long documents, highlight relevant content, and even see which particular papers have a high hit rate of useful and pertinent information. This means you instantly know what to read first and what to discard. 

Behaviour change works best when you really understand the target audience, especially (with reference to the COM-B model) whether it’s an issue of capability, opportunity or motivation that’s stopping them adopting your desired behaviour.

It’s a great model, but it’s far too dry to inspire creativity. By uploading your audience profile and COM-B understanding of them, AI will shape a ‘day in the life’ picture of their pleasure and pain points. So your creative team quickly gets to grips with their target audience and can craft better comms. 

Once you’ve designed your behavioural intervention or campaign, you can use AI to fine-tune the output. Personalisation increases effectiveness and AI allows you to do this at scale. Adopting the methodology of ABM, you can adapt a 1:1, 1: few and 1: many approach depending on how much you know about individuals. 

With high volumes of relevant specific information you can use AI to create a hyper-personalised intervention. For example, if you’re encouraging healthier eating, you could tailor meal plans and recipes based on likes and dislikes, time available and cooking skills. 

If you know less about your audience, you could group them together and provide broader personalisation based on factors like whether they’re cooking for one or a whole family. These are just a few of the simple ways AI can speed up customer profiling, improve targeting and influence behaviour. 

I’m sure this is just the tip of the iceberg. I’m keen to learn more about the capabilities of AI tools and how we can use them in real-world case examples to get better results for our clients. If you’re interested in finding out more about AI for behaviour marketing and how it could help your project, please get in touch.