I landed at the Googleplex with an idea of how to make Google search just a little more relevant for everyone. The opportunity to impact the discovery of information and ads, at the scale of Google was electrifying — but would anyone listen? Every Noogler was expected to spend the first week indoctrinated in class rooms. Instead, I walked the halls of building 43 physically hunting down the heads of search engineering with a PowerPoint outlining my naive ideas.
I soon got an audience with the head of all search at Google. On slide 3 of my PowerPoint explaining some of the math and possible benefits, he politely asked, “Did you come from Microsoft?” I nodded, blushed, and knew this was going to be cringy. “At Google, we speak with code and data. If you can run some of these experiments and get promising results, just reach back and we’ll talk then.” I wasn’t embarrassed, I was invigorated. I’d found my tribe. I hate PowerPoint. A few weeks later we met again, discussed my code and data, and I soon had a team of 10 engineers working on the project to improve web search. I’d learned to speak with data as the best form of marketing.
Everything from countries, companies, products, and even individual people in the world are marketing themselves as “AI”. It’s not you — everyone is confused by “AI Marketing” today. I recently spoke at a CTO summit on the topic of AI Marketing, mingling with CTOs, CEOs and CMO’s afterward in breakout sessions on how to market AI and how to interpret AI marketing. I’m often asked by friends if particular marketing messages are ‘real’ or ‘fake’, or even whether their own company’s AI marketing is ‘legit’. Sadly, I’m occasionally even accused of AI over-hype myself. AI marketing is confusing the marketers and those being marketed to. It is a bit of a mess. Below is an overview of the types of AI marketing, what that marketing implies about the product, and some quick tips to spot what is real and what is hype.
There are a lot of PowerPoint slides, but little code or data out there. Some of those PowerPoint slides, websites and webinars, are from people with real-world AI deployment experience, but those are few. Here are the four major types of “AI” marketing that you will encounter and how to spot the differing levels of ‘hype’:
This class of marketing originates from the confidence of reading a couple of popular books or “AI” or explainer videos on YouTube. People quickly take on the skeptical or overoptimistic stories about the general usefulness of AI and run with it. These folks don’t really know how easy or difficult it is to deploy AI to solve real-world problems, but hand-wave quite a bit. There is often little to no AI behind this marketing.
How to spot the “Naive”? These folks often talk about “needing large amounts of data” but can’t tell you how much for their product or service. They often restate the general examples they have read, versus their own experiences. They often assume AI will either change everything or nothing. They rarely give any product specifics.
Not all this is marketing is bad, the world needs smart non-AI-nerds to communicate and incorporate all this AI. These might actually be the most important people at this point in the AI transformation.
“The False Starts”
These are companies that saw the buzz and promise of AI and immediately started their marketing efforts or even company. Most of these marketing or branding starts were before their engineering teams started working on AI. These teams thought “AI” was something you just add like table salt, or support for mobile. AI has lots of promise, but it isn’t applicable to all problems, and building an AI-first version of an individual feature isn’t always guaranteed to work, let alone the complexity of an entirely new AI-based product. Some of these companies will luck out and catch up with the propaganda, but most won’t. We are already seeing many of these marketing campaigns fail and growing ‘quiet’ regarding AI claims as they are found out.
How to Spot “False Starts”: The marketing often focuses only on the core technical basics of “AI” and Machine learning — because the team only has a basic understanding, but they are very proud of it. The marketing campaigns often feel like a blitz. They don’t talk in much detail about how AI is applied to their specific problem, with specific nuances and specific benefits. They are can appear overconfident with the power of AI and speak of it just being ‘better’, without talking about the practicalities of how it might be better, or worse in some situations, the difficulty of implementing it, or the incremental nature of the value. When asking how it works, what flavors of AI they are applying and how, they will often say it is just “proprietary”, or only let folks that aren’t technical speak to it. Lack of any details is a sure sign the product doesn’t actually have AI in it yet — but they are hopeful.
I’d suggest you also be kind to these folks. I think both the marketers underestimated the complexity and timeline of deploying real AI. And keep thinking of the poor product engineers waking up to ads proclaiming their experiments are already working, and knowing it's not true. Don’t bet on these folks but don’t throw tomatoes.
There are bad actors out there, willfully and knowingly making false marketing claims. These folks don’t really care what “AI” is, it is just a ploy to get more attention and somehow a few more bucks. Unlike the “Naive”, many of these people know enough to know what they are saying isn’t true.
How to spot “Charlatans”: The charlatans make wild claims about the magical abilities of their AI-based solutions. If the claim doesn't seem real— it isn’t. If it was, they wouldn’t be talking to you — the AI would be doing all the work, and wouldn’t need them to tell you about it. Eerily, many also point out how other AI-marketers are actually the false prophets in classic diversionary tactics. These are often people new to the industry they are working in, and claim credibility from a different, obtuse, area of expertise.
Ultimately, their marketing dollars also will disappear because word of their value, or lack of it, will eventually circulate in the community. We just need to be patient. Calling them out by name just brings you to their level. Avoid them at all costs.
“The Real Deals”
The real deals in the world of AI marketing are pretty easy to spot, but perhaps not to fully understand technically. Google, Microsoft, Facebook, Amazon, open.ai, deepmind, and a bunch of smaller AI-legit marketing efforts are obviously leading the AI-bandwagon. The power of the AI in their products, sometimes to a fault, can be easily felt by users worldwide, and they fall over themselves to explain how it works so as not to allow the fear of AI hurt their businesses.
How to spot “The Real Deals”: The real deals make humble claims about incremental, but useful progress in their products by applying AI. They realize that engineering AI systems are difficult so they share code and data and learnings to attract the next wave of legitimate AI engineers to work on their problems. The biggest clue is that they are winning in their competitive categories and getting better day by day. Real AI is transformative — you can see it.
Ultimately AI is difficult to apply to specific applications but will be game-changer for almost every industry, just like mobile, and cloud. This technical transition will leave many companies and people behind, and the winners will be the real-deal AI players.
Improve your Hype Radar
Above are some good guidelines to identify the hype in AI marketing, but the best way to cut through all the hype is to simply invest a few hours of learning. If you are somewhat technical, I recommend watching a few YouTube videos from Andrew Ng. They will quickly give you the context to make heads and tails of what ‘real AI’ is. If you are non-technical, I highly recommend the recent book by Melanie Mitchell, A Guide for Thinking Humans. The book is an approachable and near-complete explanation of modern AI technology and has a very practical bent. If you can’t spare the time to watch Andrew Ng scribble on the whiteboard, or listen to Melanie Mitchell’s new book on Audible on the way to work, you will forever remain at the mercy of AI marketing.
The beginning of every great transformation in technology brings out these same marketing patterns. Feel free to indulge in the chaos of it all, as I do, but, be careful of which AI you bet on.
Jason Arbon . CEO @ https://www.test.ai