Lately, I’ve been deeply immersed in the world of data: thinking about it, discussing infrastructure, and pondering data foundations. The more I reflect on how technology has evolved and how people consume information on new platforms, the more I realize that data should be the most important component of any strategic focus within an organization because it not only can help drive insights, but it can help with discovery on platforms, old and new.
In my former role, I used to talk a lot about data and how organizations need to have a strong data foundation to execute any task, marketing or otherwise. Now, more than ever, this foundation is critical. As I’ve explored the trends disrupting search and consumer behavior, the importance of data has become increasingly clear for any strategic placement an organization needs to reach a patient or consumer.
This week, I read an article about ChatGPT and Shopify that made me think about how important data is – and will be – to healthcare organizations and to how people discover healthcare now and in the future. It wasn’t healthcare-specific, though. It discussed code strings in ChatGPT suggesting OpenAI is setting up for users to “discover, select and purchase products without leaving the chat window…” by integrating Shopify’s checkout flow directly into ChatGPT. This means going to ChatGPT to tell it what you are looking for in natural language and for ChatGPT to deliver options that you can buy, all within the chat. It also means not visiting a website, at all. (Agents are already doing this on our behalf – if you’ve used any Deep Research experiences on Gemini, ChatGPT or Perplexity, you can see that these tools are reviewing websites and producing amazing research without the user having to comb through the sites individually.)
This requires a vast amount of data to be successful and ensures the “machines” have access to that data. Data like: product inventory and availability, product details, descriptions, images, reviews (to name a few). It might even eliminate the need for a website in a traditional sense, because all an organization will need to do is assume someone is using ChatGPT to search for a product and it will need to ensure ChatGPT can find the product and apply an LLM to that agent’s search to drive a more personalized result on ChatGPT. What’s the point of having a traditional website if all you need to do is surface the data behind the website to these platforms?
This could fundamentally change how we interact with the internet. (Why do you think so many organizations are expressing interest in purchasing Google’s Chrome browser?)
This also means it could (and likely, will) impact healthcare.
Of course, this won’t happen overnight. People still use Google more than ChatGPT: for years, people have developed ingrained and well-established behaviors to use Google to find healthcare information. It’s no secret that 7% of Google searches are healthcare-related searches. I did a quick back of the envelope calculation to determine that there are 35B healthcare-related searches on Google per month.
(7% of 16.4B US searches per day = 1.15B healthcare searches per day x 30 days per month = ~35B healthcare searches per month)
That is a LOT of healthcare searching going on, even though Google is supposedly losing traffic and attention to other platforms like ChatGPT.
So why does this matter in the context of data and shopping on ChatGPT?
If OpenAI is launching a shopping agent via a direct integration with Shopify, I think it’s worth considering the possibility that they just might launch a healthcare booking agent in the future. And if OpenAI does it, why wouldn’t Google? Or Perplexity?
While ChatGPT’s daily “searches” are far less than Google’s, its adoption is rapid and (as I’ve mentioned before) behaviors are rapidly changing. People like natural language search. They like being told a direct answer rather than having to sift through blue links and doing the research themselves.
Assuming a ChatGPT search is similar to a Google search (which it is not, but bear with me), ChatGPT is gaining some momentum here: what if 7% of ChatGPT searches are also healthcare-related? We don’t know how much people are “searching” on ChatGPT per day, but we know that there are 800 million monthly active users of ChatGPT, so if we were to assume that 7% of those users are searching for healthcare, we’d see 56M healthcare “searches” on ChatGPT per month. (Again, not apples-to-apples, but at least it anchors us a bit).
It’s not Google-level, but 56M is still significant. It shows us that millions could be using ChatGPT for healthcare information (if we assume that people use ChatGPT the same way they would use Google). If OpenAI facilitates purchases for retail in the near future, why not facilitate healthcare “shopping” in the future, too?
Healthcare organizations must rethink how they reach consumers and how critical their data is to that journey. This data fuels information discovery and powers the machines that will be leveraged. I’ve said it before: your healthcare data, especially provider data, is your inventory. And it’s never been more important to make sure you have command over it. For example, if your data lives in your CMS – can the data in that CMS reach ChatGPT? What about a wearable like the Meta Ray-Ban glasses? What about Alexa+? What about a future platform we don’t even know about or use yet?
I’d argue no.
And I’d also argue that if you continue to think that your website is going to be the place where people do research to find a provider or learn more about a health condition or diagnosis – that’s not in the cards either. It’s time to think about radical changes that are needed to adapt to patient and consumer behavior and how to optimize for the machines rather than for people.
The website as we know it today is dead on arrival.
The agents are coming.
And the agents are eliminating the need for a traditional website. The agents are hungry for data that lives on the website, but don’t need the website in its current form. People soon won’t, either.
People are shifting their attention and their behaviors to places where it is easier to find information – where they don’t have to sift through endless blue links. Technology is shifting too, by reducing friction and facilitating conversations and conversions. Without understanding your data and how it needs to feed the places where people discover information today and more so in the future, you’ll be left behind. You won’t be part of the conversation, and you won’t connect to the platforms driving the patient discovery and conversion.
For example, if you are building a search tool on your website today, are you already living in the past or are you preparing for the future where people may never visit your site to search for a provider or a condition? Are you preparing for the interim where some people will still come to your site to find information, but also sharing your data in a way that will surface data on ChatGPT, Gemini, Google and more?
The future is arriving fast. What you build today will probably be obsolete in six months. For example, if I hear one more leader say “we’re going through a website redesign”... What’s the point? What leaders should be thinking about is how people may never visit a website in the near future. And, about strategies to construct a data lattice that feeds platforms and technologies with your data. Machines will access your data in the future. So how do you construct a data architecture that is flexible enough to deliver the right information to the machines, assuming someone may never visit your website ever again?
Imagine: instead of OpenAI integrating code snippets from Shopify, they instead decide to work with EPIC to integrate code that allows for search and scheduling directly via ChatGPT. What if Google did this, too?
There is a real chance that a patient never visits your owned properties to find healthcare information or to book an appointment – but instead, leverages the power of agents to take action on their behalf.
How will you make sure the information and technology that used to live on your website is now alive and well on other platforms and accessible to the agent?
Will you and your data be ready for the change that is rapidly in front of us?
Thought-provoking @Carrie Liken! It begs the question as to what will happen with HIPAA ultimately. When people start using AI agents to search for healthcare diagnoses and services, it will become nearly impossible to enforce HIPAA - like the Meta pixel placement issue on steroids.