In today’s episode, I explain how AI will enable synthetic influencers and better influence identification. Large language models can ingest vast information to surface relevant influencers. Brands should partner with AI firms to build custom models filtering by niche and demographics. Tune in to hear my predictions on AI’s transformative role in influencer marketing.
Can’t see anything? Watch it on YouTube here.
Listen to the audio here:
- Got a question for You Ask, I'll Answer? Submit it here!
- Subscribe to my weekly newsletter for more useful marketing tips.
- Subscribe to Inbox Insights, the Trust Insights newsletter for weekly fresh takes and data.
- Find older episodes of You Ask, I Answer on my YouTube channel.
- Need help with your company's data and analytics? Let me know!
- Join my free Slack group for marketers interested in analytics!
What follows is an AI-generated transcript. The transcript may contain errors and is not a substitute for watching the video.
In today’s episode, Caitlin asks in the realm of B2B influencer marketing, how do you envision artificial intelligence playing a significant role in the coming years? What specific applications or advancements you believe AI will bring to this field? Okay, so there’s a bunch of different ways.
Here’s the thing, artificial intelligence, particularly generative artificial intelligence is all about language, language and images, which are the two areas where as at the time of this recording, the most capabilities exist, large language models like the ones that power tools like chat GPT, for example, can generate language.
So there’s a couple different ways to think about this.
What does an influencer do that involves language? Pretty much everything, right? So you will see the advent of the synthetic influencers, a synthetic influencer, this is very popular in B2C, and it will probably become popular in B2B, where you have a personality that isn’t real, there’s no human being behind it.
There’s a team of humans that maintain the infrastructure, but that personality does not exist, there is no person there.
And those synthetic influencers have thus far been very popular.
And they’re especially appealing to brands because, well, your influence is never going to show up drunk, you know, or have compromising photos put up on Instagram or something like that, because they’re not real.
They they simply do as they’re told.
So that’s a big part of ways that you might see generative AI put to us, but the bigger and more obscure and arcane way to think about influence when it comes to AI is the language model itself.
Language models are trained on a gigantic amount of text, they are processing, among other things, things like common crawl, which is a huge repository of the public internet.
As a result, they see more and have more data in them than any other tool available to us.
And as long as they’re prompted correctly, you can extract out information from that massive meta index, or you can drill down and supplement these tools with specific data that you may have internally.
So for example, let’s say you do a survey of all your customers.
And you say, who is most influential in our space, maybe this industrial concrete space and you go when you read a bunch of blogs, and you get a bunch of papers, you have a bunch of news articles, you cram those into what is known as a vector database, which is a special kind of database that large language models like chat GPT can talk to that they speak the same language specific kinds of numbers, they turn content into numbers, and then they do a bunch of math on those numbers.
When you do that, if you do that, if you build that tool, and I would expect the leading influencer marketing software companies out there tools like Analytica, for example, to build that system, you will be able to ask large language models who is influential in a space and get an answer is coherent, that is relevant to that specific space, but also has the broad knowledge of the internet of everything that’s been published on the internet and public spaces.
Again, these models take in the public web, they take in books, they take in all the academic papers on sites like archive.org.
And as a result, what you end up with is a pretty good knowledge base of what people what people are out there, what is influential out there books, podcasts, etc.
That you may not be able to get from traditional influencer marketing tools or even influencer marketing agencies, because let’s face it, no one human being can keep that much data in their head, but a language model can.
So what companies should be thinking about what they should be looking at is working with the leading edge, the cutting edge, language model folks and firms things to build influence models that can do this kind of work that can extract data from language models, which are really just statistical models of the public internet, among other things, and come up with influencer conclusions that way, because language models inherently can take and filter much more precisely, and much more granularly than traditional influencer marketing tools, you could say, for example, show me the influencers in b2b, industrial concrete, but I specifically want to focus on influencers who are BIPOC, black indigenous people of color, who in my space is is that you can ask those questions of a language model and get information, you would then have to go corroborate the information, you want to make sure it’s not a hallucination, you want to make sure it’s not the language was not making things up.
But you can ask those questions in ways that traditional influencer marketing software doesn’t even know is a question.
So that’s what I expect AI to provide to the influencer marketing space and in the coming months, not years, months, because every release of these models that comes out, the models get smarter and smarter.
At every release of the technology, particularly the open source side of AI gives more capabilities for all of us, if you’re technically savvy, to be able to construct custom influencer marketing tools that fit the specific needs that you and I have.
So that’s the answer to that question.
It’s a really good question.
And it’s very difficult to predict the future.
I mean, we can’t even predict four weeks out much less years, but where the tools are today, we can gather this information we can we can use what we know for today.
And with the types of applications we can build on what exists in market today, if something new comes out tomorrow, we can adapt to it.
But conceptually, from a strategic infrastructure perspective, this is where influence marketing could go.
If the leaders in the space are willing to invest and embrace these generative AI technologies.
Thanks for the question.
If you enjoyed this video, please hit the like button, subscribe to my channel if you haven’t already.
And if you want to know when new videos are available, hit the bell button to be notified as soon as new content is live.
You might also enjoy:
- How To Start Your Public Speaking Career
- How I Think About NFTs
- Understand the Meaning of Metrics
- B2B Email Marketers: Stop Blocking Personal Emails
- It's Okay to Not Be Okay Right Now
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers