Until recently, the healthcare communications industry was as settled as any other. Big-named agencies were financed by large holding companies, while small independents brought up the rear. The companies within this dynamic knew what they offered and could focus on being good at it. That was yesterday. Today is very different.

Advances in AI technology have meant these agencies can now provide genuine utility to healthcare communications, as these tools become increasingly accessible. These advances mean there are now lower barriers to entry for developing in-house AI capabilities or using third party applications, such as Midjourney and ChatGPT. Our industry, like so many others, is being disrupted. Disruption means a redistribution of money within the marketplace. Some will win, and some will lose. As a business leader, I am acutely aware that it is the decisions we make today that determine which side of the line we will fall on. But there is no blueprint, only predictions. Pharmaceutical, medical device and technology companies who partner with communication agencies like ours must differentiate between the quality and the noise. This article is designed to help provide some clarity in these confusing times.

In order to achieve this, it is important to separate out the opportunities in front of us, while considering where we are today and where we believe we may be in the future.

1. The adoption of 3^rd party AI applications

Outside of the legal considerations around data sharing, the adoption of these technologies is actually very easy. If you do nothing, your employees will most likely integrate these into your business for you. ChatGPT is the fastest growing application in history, and most creatives I know have been drawn in by generative AI applications like Midjourney and Adobe Firefly.

Where, when and to what extent they are being used is the bigger question, and standard operating procedures around the world are being augmented to maximise the benefit of these tools. The efficiency they promise is clear, but we still require skilled people. Brand and strategy cannot be replaced by ChatGPT because it averages everything out in a world where we require an insight into what will change behaviour within a unique and nuanced scenario. Similarly, creativity is not yet replaceable by machines, not just because we need a very human understanding of the creative request, but because trying to create an image that exists in your minds-eye using language prompts in Midjourney is near impossible!

In the future, generative AI promises to reduce the production time of assets to milliseconds. In our not-too-distant future, a photorealistic image, a video, a letter or even an advertisement could be produced instantaneously. When this happens, we can truly personalise content. I am not talking about a copy tweak or heading change depending on the audience; I am referring to bespoke content for every individual within our target audience based on what we know about that person.

One question yet to be answered when it comes to changing behaviours is whether true content personalisation is more effective than generalised content. The working hypothesis is that the more tailored an advert, the better results it will garner but this notion still needs testing.

Some agencies have already gone all-in on generative AI, positioning themselves as an AI creative agency. For me, this is simply a claim that they have adopted AI quicker than others and not a true differentiator that will propel that agency forward. Is it really possible that there will be agencies in the future who do not use AI? I see this as the equivalent of being ‘the Google Agency’ or ‘the Internet Agency’.

1. Developing a proprietary AI offering

How can AI be a differentiator when everyone is using language models like ChatGPT and generative AI tools like Midjourney? One possibility is that the world returns to some normality. If AI becomes as embedded as Microsoft Word, then it will not be used to differentiate at all. We will return to focusing on who can be the most creative, the best at market research or the best at medical affairs, and AI will be a given. I think this is unlikely because there is sustainable differentiation offered by AI and that comes from having an AI built on a proprietary data set. As the new adage goes, artificial intelligence without the right data is just…artificial. AI is only as powerful as the data upon which it is built. If you have your own data set, you may have a fighting chance. So agencies must be prepared to explain how their data set is unique, how it is processed and what evidence they have to prove it leads to lasting behaviour change.

For our clients, I am witnessing agencies staking claim to the biggest data set as a way of conveying AI quality and dominance. However, my advice would be to approach with caution. Size isn’t everything, particularly when it comes to data. As a client, be sure to ask the following questions:

• Is the data relevant?
• Where are these data sets coming from?
• Have you paid for these data sets and who else has access to them?
• If everyone is working with the same data, then where is the differentiation in your offering?

If the AI offering is predicated on data being supplied by the client for processing, this is not a differentiator, it is a service offering. Be sure to ask why their process is better than their competitor’s.

2. Combining third party AI applications with a proprietary AI offering

The idea that proprietary data derived from, or processed by, large language modelling can provide relevant real-time prompts for generative AI opens the door to practical closed-loop systems. A world where data and generative AI work together to produce personalised content, measure its impact and use this information for iterative improvement, feels scary. What is our role in this new world?

Take some comfort from Purple Bricks, a real estate company that reduced the human element of house sales. The company should have transformed the industry, but it seems that people prefer it when actual human beings are involved. For now, the machines will have to wait to take over.

At Page & Page, we’re choosing option three. If you’d like to learn more about how we’re combining third party and proprietary AI, message me.

By Kristian Webb, Managing Partner