Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
AI: An Unlimited Army of Interns, A Conversation with Alfred Whitehead at Klick Consulting
48 Plays9 months ago
Touching the hottest topic at the moment - Artificial Intelligence - Alfred Whitehead, EVP of Applied Sciences at Klick Consulting, joins Rusty Ray at Alantra to unfold applications in drug commercialization, particularly OTC, and its impact on personalization in marketing and product development.
Recommended
Transcript
Introduction to 'Crossroads by Elantra'
00:00:06
Speaker
Hello, you are listening to Crossroads by Elantra, our podcast that focuses on healthcare. I'm Rusty Ray, and I head up the healthcare investment banking team at Elantra here in New York. Today, we are finally touching on a very hot topic, AI.
Guest Introduction: Alf Whitehead
00:00:21
Speaker
And we have as our guest, Alf Whitehead. Alf is an EVP of Applied Sciences at Click Consulting. I'm not so tech savvy, so this is going to be a very educational episode for me as well.
AI in Drug Commercialization
00:00:32
Speaker
We will be discussing AI applications in drug commercialization
00:00:36
Speaker
particularly OTC, and more interestingly, how that impacts personalization of those products from both a marketing and product building perspective. In our recent consumer health report, the personalization of things was a real big trend that we're watching for. So we're excited to dig in. AI has been in the news a lot and not always for good reasons.
00:00:56
Speaker
but maybe just for our listeners, Alf, can you give us a little bit of context about what is generative AI and just help us kind of level set what we're about to embark on in this conversation?
Generative AI Explained
00:01:07
Speaker
Great. Yeah, I'd say generative AI, Rusty, is sort of the most active field of research in the AI right now. And what it is, is really a set of models that play this game of completing the sequence, completing the pattern. So if you've reused a chatbot, for example, that's been trained on human language and you give it some language and it thinks about, Oh, okay. What's a likely way to complete that pattern? And that turns into a conversation.
00:01:31
Speaker
And that's what gets generated in generative AI is that completion to the sequence. Now, this is going to be incredibly powerful. We've started with language, which has given us a whole new way to talk to computers, but it's going to work for pretty much any other sequence. It can fill in the background in your photos, or it can figure out how a genome should be completed in order to figure something out once we get that training in place. Very exciting, right?
00:01:55
Speaker
And along those lines, I guess it's come to the fore so much more recently. What do you see as being different today is say even 18 or 24
Foundation Models and AI Setup
00:02:05
Speaker
months ago. I mean, how has this conversation really taken off with your clients and how do you think about generative AI today versus even kind of recent past?
00:02:15
Speaker
Yeah, I've been in the space for a number of years and I think the biggest change now is we have what we call these foundation models which are a eyes that can be used in other context like hiring an intern into your business has this baseline level of education and common sense and knowledge of the world and then we teach it sort of the specifics of the business on top of that and so that's been a radical change in how we deal with these things.
00:02:38
Speaker
We're now able to set up AIs incredibly quickly. We're able to teach them in the same way we teach humans. You no longer need a whole bunch of computer science experts to do this. Subject matter experts can work with it directly. It's super exciting for us.
00:02:52
Speaker
Thanks. That's really cool and interesting. I appreciate the background for novices like me in the computer sciences. So that helps me sort of understand what we're trying to embark on in terms of our conversation.
Background on Click Consulting
00:03:04
Speaker
But maybe before we get into the weeds on commercialization and how AI is helping to redefine that landscape, maybe just give us a little bit of background about what Click does, what you've done historically, and then
00:03:19
Speaker
Yeah, let's maybe transition into how you're using AI for your current clients and what clients are demanding of you on the AI landscape. But maybe let's start with a background on click.
00:03:29
Speaker
Yeah, absolutely. Click is a privately held company. We started about a little more than 25 years ago, born out of the web. And we really focus on helping life sciences companies commercialize their products. That can be pharmaceutical, OTC, med device, anyone in that space. If you have a product you want to take to the market, we can help with everything from promotion to strategy to all of those things
00:03:52
Speaker
Do your clients know what they're asking for our clients savvy with a and they say this is what i want this is how i want to do it or are you having to do a fair amount of instruction and saying hey this is what is even capable of forget the myths out there in the newspaper let's talk about how we can really help you are you really having to start at a more basic level in some cases.
Misunderstandings of Generative AI
00:04:14
Speaker
Sometimes i mean i would say there's been the name generate a is a bit of a misnomer when it comes to how folks in the world of business and commercialization think about the word. So i would say most people come to us and they say hey can we use generate a to do our copyright can use it to create the text that we're gonna use in promotion.
00:04:33
Speaker
And yes, you can. But when you think about what a copywriter actually does in their day job, there's only a tiny amount of their job that is actually typing the words that go into the advertising. And so we've really been taking a much broader view. AI is effectively an unlimited army of interns we've added to our company that can help us out. And so we think about everything that goes into creating that promotional content throughout its
AI in Research and Regulatory Tasks
00:04:57
Speaker
lifecycle. We've really focused on it. And so we've been using AIs to go through and do things that
00:05:02
Speaker
like reading every scientific paper in the space with a specific lens and summarizing it and comparing these things, looking across the entire competitive landscape of every other product in there.
00:05:13
Speaker
reading all of their reviews and understanding all of that. That's the sort of power AI can bring to that research phase. Yes, it can help with the writing, but then after you've done the writing, you usually have to go through some sort of regulatory clearance process. AI can be very helpful here and understand how the promotion fits into the various FTC or FDA or whatever the appropriate regulator's point of view is.
00:05:34
Speaker
in order to make sure that we can actually get that content out to the customers. So that's been super helpful. And AI actually helps us orchestrate that entire process. One of our most successful projects has been a set of AI representatives, we call them, that take on different points of view of people working on a project. And they can use that point of view to evaluate project plans and get us to a better planned project at the beginning.
00:05:58
Speaker
That's really interesting because it sounds like you're using AI or your clients are expecting the use around AI to generate so much more data, richer data that leads to better outcomes.
AI in Market Research
00:06:13
Speaker
So often we hear about AI as sort of a replacement for the types of work that folks do. And you described some of that replacement as a team of interns, but the value add, the real sort of
00:06:27
Speaker
technical areas where there's regulatory and other kinds of checks on language that would be maybe disallowed by the FTC, for example, that's still something that falls on an individual to check. Was that expertise or do you expect AI to sort of take on those kinds of roles in the future too? I think the way we're looking at AI is kind of a next generation spell check in some ways.
00:06:52
Speaker
So one example might be market research I've been experimenting with. When you're trying to get the voice of your customer, you do a bunch of market research, maybe focus groups, you get a bunch of feedback from real consumers. And AI can synthesize that as a persona that you can then talk to in the future so that you can sort of bounce ideas off of it as you're working, even though your focus group isn't meeting right now.
00:07:13
Speaker
That's really impressive. I mean, just the concept of having an ad hoc focus group on demand would just be incredibly powerful. If you're shifting strategy or tweaking language or positioning around certain brands, whether it's for new launches, new product development and the like, that to me is extremely powerful and very cost-effective and probably something that in the past would have been taken months and months to organize, whereas
00:07:40
Speaker
Here it sounds like it can be done in a matter of hours. Hours sometimes, sometimes even minutes. Occasionally we find these AI use cases that are literally just a prompt. You go to one of these AIs and say, hey, I need you to do this. Can you do that for me? And it does. It's kind of magical.
00:07:55
Speaker
That's really cool. You've given us a couple of examples, but I guess more holistically, how has Click been using this with your clients?
Personalizing Content with AI
00:08:05
Speaker
Is it really just on initial commercial launch and around that customer experience and getting that feedback from the customer? Or are there broader use cases around the overall commercialization effort as it relates to going back to AI?
00:08:20
Speaker
It's a good question, Rusty, and I think when you look at where most companies are today, 2023 was a big year of experimentation. Most companies have been just figuring out how it can fit into their business and what it can do for them. So we've worked with a number of OTC companies working on things like personalizing content, for example.
00:08:37
Speaker
So if you have as one of our customers does a database of hundreds of expert written articles that may be appropriate to a certain type of consumer. Can we have the AI take the consumers need read the articles and restructure them and repackage them in a way that's personalized for that specific person.
00:08:55
Speaker
And it turns out that works really well. And we think that sort of personalization is going to be a trend in marketing. A lot of our clients are looking at how can we take these AIs and use them to really improve the personalization. But you also have to know quite a bit about your customers in order to meaningfully personalize to them. So how can we create and build that relationship? You also have to be able to deliver these programs through a channel, whether that's on the web or somewhere else. And you have to be able to measure how well this is going.
00:09:25
Speaker
over time so that you can figure out are you actually getting the value out of here? Or are you just creating sort of a chat pod that no one talks to? Yeah, a lot of noise, as it were. One of the things you mentioned a moment ago is talking about personalization. And I think I agree with you in terms of the personalization of the message. But what we're seeing in the OTC space, especially is personalization of products.
Personalized Consumer Health Products
00:09:48
Speaker
You have vitamins that are tailor made for an individual or an individual type.
00:09:53
Speaker
You have certain skin creams that are more and more based on skin types or other genetic factors that may relate to certain conditions like acne and things like that. Do you see kind of an even deeper personalization as it relates to these kinds of product types and these kinds of
00:10:12
Speaker
initiatives within the larger consumer health companies, really targeting individuals and sort of bringing them into this personalized product space as well, not just the broader product use. Yeah, I think over the next few years, we're really going to see the way that people select and buy products probably change quite a bit with these new technologies. Right now, the traditional model has been going to your mass market drugstore, you pick a product off the shelf and you take it with you. You've talked a lot about the personalized products, which are usually delivered through e-commerce channels,
00:10:42
Speaker
based on survey data or maybe even genetics if you're getting really fancy and can be shipped directly to the consumer. But the consumer is still the one sort of driving that buying behavior. So one of the interesting questions we have is how do AI agents change that?
AI's Impact on Consumer Choices
00:10:57
Speaker
Will the consumer still be the one browsing the shelf either virtually or physically? Or is that going to be an assistant for the consumer? Is that going to be one that's provided by the manufacturer or by the retailer or by some third party that's in control entirely by that consumer?
00:11:11
Speaker
We've been here before. Product discovery changed quite a bit. When the web rolled out, we suddenly had search engine marketing and search engine discovery. And I would argue that's actually 20 years old now, but it's still AI in its own way.
00:11:24
Speaker
A blunt force version, perhaps, right? Yeah, in fact, less sophisticated. We as marketers and product creators, we got used to telling search engines about our products so that they could be discovered by consumers and might use them. There's going to be another more sophisticated level of sort of decision support, whether that's for the consumer, maybe physicians, depending on the context. And then that will unlock that universal personalized products as well as personalized messaging. How can you navigate all the options to pick what's best for you?
00:11:51
Speaker
That's really interesting. And I want to come back to a point you just made about the physicians. But I guess first, one of the things that we've seen over the course of the last few years is retail has had a tough time, to say the least, especially the drug channel.
AI in Evolving Retail Spaces
00:12:06
Speaker
And so even for those products that are problem solution, where you need it now, think of headache medicine, you don't want to wait for Amazon or Ecom.
00:12:16
Speaker
I feel like as the drug channel continues to narrow and there are fewer and fewer brands on shelves as they sort of reposition into whatever they will become, mini grocery stores or what have you, to drive foot traffic, it feels like AI is going to become only more important because you're going to have to get above and beyond the noise to really reach your consumer. Are you starting to see those kinds of trends take shape?
00:12:43
Speaker
Yeah, I think it's a really interesting point. And I think of it as like a long tail effect. You have sort of your very common conditions that affect essentially everyone, like headaches. And those are probably going to be pretty much every retail point of sale you can think of forever. And then you have this massive long tail of conditions that are a little more personalized. And I think
00:13:02
Speaker
The complexity of number of solutions available is just going to grow and grow. And so having an AI to help navigate that is going to make that digestible for consumers. I think you mentioned the changing retail environment. And it's interesting. We do see drug stores changing considerably, but one of the things they are often adding now is clinical support from people like physicians so that they become sort of a one-stop problem solution place to come, regardless of the regulatory.
00:13:29
Speaker
We've talked about commercialization and AI. I guess when I think of that, my own biases are new brands, innovators. How are some of the older, larger brands, let's say the Tylenols of the world, to use an example, are they sort of attacking the problem in the same way? Are they viewing it differently because they have a different relationship but a longer storied relationship with the consumer or how does some of the larger brands think about this?
Trust in Brands vs. Retailers
00:14:00
Speaker
Fundamental currency that it all comes down to is going to be trust. Who is the trust of the consumer? And I think some of those story brands like your Tom and All's of the world, they have a sort of very direct relationship with consumers where that brand name has a trusted place with the customer who's buying. And I think that's the ideal place for a brand to be, is sort of where they are. So how does AI help them? Well, it may help them as their product line becomes more complex, or as they have more specialized products for different consumers.
00:14:14
Speaker
application vis-a-vis.
00:14:28
Speaker
But fundamentally, they've got the trust. Now, the question is for the folks who are not on that level of brand connection.
00:14:36
Speaker
Is the trust going to be built between the consumer and the brand or is it going to be built between the consumer and the retailer? If you can imagine a retailer offering hundreds of solutions, if you can build an AI that provides access to that, whether it's coming from your own stable of products or maybe from somewhere else, that may be an advantage for a more narrow AI that can only pick from a limited set of recommendations.
00:14:59
Speaker
On that frontier, I totally agree that there's probably a fight that's about to be set up between the retailer and the brand owner as they try to sort of vie for the trust of the consumer and the retailer becomes more of a care provider in air quotes, I'll say, and advising consumer on how to satisfy their needs. Does that marginalize the brand owner in any way, shape or form? And if you think it might,
00:15:25
Speaker
Does that lead a lot more folks to say, well, look, if we can start to target our consumers better and e-commerce at a place where everybody's comfortable with it, does that drive a more DTC model for many of the call it personalized or tail type products that are
00:15:41
Speaker
not so immediate need. Do you think that could be a trend? I definitely see the need for some of those long tail brands to have a more direct relationship with the consumer. I think there's going to be competition there. The retailers are going to want to try to grab that as well. Where I think the rubber hits the road for the brands is on repeat purchases. How can you support your consumer
00:16:01
Speaker
reason away that brings them back. And if you have a model that can achieve that, then the loyalty I think is going to fall to the brand as opposed to the retailer. And that I think is really going to be the key to unlocking that for brands. That's really interesting. Just to kind of pivot back to something you said a few moments ago on the physician or care provider.
00:16:23
Speaker
How does the care provider get integrated into this? Are there risks around that as it relates to AI?
AI's Accuracy and Societal Expectations
00:16:30
Speaker
We've all seen examples in the newspaper of AI making some somewhat hilarious mistakes. And so as we think about consumer health, be it OTC or RX, if you're starting to think about connecting with a consumer and almost guiding them through their patient journey,
00:16:50
Speaker
Do you see risk there that people are sort of coming to their care providers with ill-conceived notions or that care providers are somehow getting miseducated on a certain brand because of the way AI is interacting with a consumer or the brand? I love this question and I think there's a few things going on.
00:17:09
Speaker
Some of it is our social expectation of computers. We've been trained over decades to think of computers as exacting and accurate and mathematical and all of these things. This current generation of AIs is actually a lot more like an organic brain in terms of making mistakes, having little flubs, doing those things that we find socially acceptable in humans. When doctors pass their medical exam, the bar is not set at 100%.
00:17:35
Speaker
We don't expect that from anyone, but we expect it from our computers. And I think that's why people are sort of latching on to AI doing all these weird things. The reality is it's a lot like a human. It's going to make some errors in some sort of predictable ways. It's going to forget things. And I think as that sort of social expectation shifts, that's going to die down and people are sort of going to understand that and start treating it the same way.
00:17:56
Speaker
When it comes to physicians, i think they're really interesting model i mean that used to be historically your one stop shop for solving whatever problem you had to go to the physician and they would figure it out in the access to everything what's the problem with that has been convenience of access and cost of access and there been a lot of barriers actually getting there.
00:18:15
Speaker
I think AI can help with that by acting as sort of a triage. I also think AI support is probably coming for the physicians even before the consumers in some cases. There's been a lot of work done on clinical decision support systems. These are bits of software that integrate usually with EHRs and help doctors decide what to do. I saw a really interesting study at the end of last year looking at
00:18:35
Speaker
physician attitudes towards recommendations for which drug to use. And most physicians use the recommendation from the software and they valued it even over recommendations of their colleagues, which was kind of surprising. The impact of that I think is that physicians are going to get used to having an AI assistant help them with triage.
AI in Healthcare Decision Support
00:18:53
Speaker
I think you're going to see a lot more blending of the products they pick. It's not necessarily just going to be prescription. I think OTCs are going to show up a lot more in the physician's office.
00:19:03
Speaker
as a result of that, because these recommendation algorithms are probably going to be looking at, hey, what's the simplest thing that could work first that's easily accessible? It's like a very modern digital interpretation of a very analog solution, which is a sales force, calling on docs. And we all know that can be very effective. Kind of pivoting back to your client base, if I were one of your clients, I guess I would feel pretty overwhelmed right now. Are people kind of coming with a very sort of broad
00:19:32
Speaker
help me out notion or are they coming with, we got this, we know exactly how we want to handle it and take it on. How are you advising clients to take that initial step?
Guidance and Training with AI
00:19:43
Speaker
I would say the help me out is more common than having a clearly baked strategy when it comes to AI. That's fine, I think. This is all very new. What we recommend for our clients generally is
00:19:54
Speaker
Understand what this technology can do. There is not a huge barrier to entry here, which is I think a big difference compared to some of the past technological revolutions. I can sit down with somebody for a day and I can get pretty much anyone comfortable enough so they can do something innovative with this new tech. I think it's important to understand sort of the tools you have in your toolkit because otherwise you're just running around with the AI hammer looking for a nail to hammer in.
00:20:17
Speaker
Yeah, I think you you raised a very good point often with these newer shiny technologies to use your analogy when you have an AI hammer. Everything looks like an AI nail solution. And that's a wrap. Thanks for taking the time, Alf. Really appreciate it. That was truly fascinating.
Conclusion and Future Engagement
00:20:35
Speaker
I love the analogy of AI to an army of interns, because when you typically read about AI, it just all seems so very overwhelming. But of course, as you said, you got to understand the tool. Otherwise, you're just running around
00:20:47
Speaker
with an AI hammer looking for an AI nail, as they say. If you ever want to hear more about our thoughts or where we're headed within the consumer health space and some of the work we've done, please reach out to us. Thank you.