Discover more from B2B Wins by Steve Zakur
B2B Wins #20: AI faces the chasm
OpenAI is the Evil Knievel of AI
AI went from mystic vaporware to something that the average person can interact with within a moment. It was just a few months ago that chatbots were stupid, AI couldn’t be seen, and search was painfully ineffective and dull. No more.
It’s readily apparent that the current generation of these new chat technologies isn’t very smart. They’re great speakers. They’re lousy speechwriters. None know anything.
But the great thing about these chatbots is that they’ve changed the arc of AI forever. There’s been a fundamental shift in how people think about AI and that is everything.
If AI were a tech startup what we have witnessed is its Crossing of the Chasm. It’s time for you to cross with it.
1. What is this chasm you speak of?
It’s hard to believe that way back before the internet was a thing people still walked erect and wondered how to outcompete their competitors. Before the advent of the search engine. the management consultant ruled the information savanna.
The management consultant sometimes grew lonely and thus needed the management guru to keep him company. One of those gurus, Geoffrey Moore, penned a book in 1991 called Crossing the Chasm.
This book helped tech entrepreneurs everywhere understand the dynamics of selling highly disruptive technologies to mainstream customers. Most technologies attract interest. However, few can get beyond that early population of visionary customers.
Here’s the challenge: Once you have convinced a bunch of corporate oddballs to buy your tech your problems are just getting started. How can you persuade the vast armies of mediocre middle-managers, concerned more about their 401K balance and means to pay the mortgage than growing the company they work for, to buy your product? This is where most early-stage companies founder. This is the Crossing the Chasm moment.
Shifting from Early Adopters to the Early Majority requires a change in the company’s approach to sales and marketing. You need new processes, new skills, and likely a whole new message.
This is where AI lived until late November 2022. The trajectory of AI had become unclear because most of us didn’t know what to make of it. OpenAI helped us see the power.
2. The realm of nerds
The vast majority of AI to date has been limited to the realm of nerds. Certainly, the software engineering types have been at it for a while. More recently those marketing folks who are wonkier than the average person were buying tech that included AI mojo. These are those early adopters. They care more about value creation than career risk. But this is a rare breed.
If you had to explain AI to someone who wasn’t an engineer or didn’t speak data you were often met with skepticism. You were speaking a whole different language with strange acronyms and phrases that were impervious to deduction.
Powered by Knowledge Graphs.
“What’s a knowledge graph?”
That’s fuel for skepticism. Doubly true if that enquirer is the CFO.
What has changed dramatically in the past 6 months is the emergence of AI tech that average people can see and understand. This is a huge unlock. Not because these chat technologies are, or are not, good at what they profess to do. The skeptics will always have a field day as problematic behaviors are discovered.
It’s a huge unlock because anyone who watches what these chat technologies do is amazed. To the average person, it looks like magic. But the real magic is that anyone can use it. AI is no longer purely in the realm of nerds. Anyone can now chat with a machine.
3. Was ChatGPT the key?
ChatGPT was the key to unveiling AI in a way the masses could understand. I’m not sure why ChatGPT had its moment. Sure, the technology was better than we’ve seen before but it took off in ways that were unpredictable. Regardless, ChatGPT helped with the nerd problem.
The biggest challenge of AI is that much of the technology operates in a black box. Its decision-making is inscrutable. To be clear, ChatGPT is no different. And yet, it is different.
One of the biggest challenges in AI adoption has been that it's largely an invisible hand. If your CMO tells you that marketing campaigns are better because of AI they’re going to talk about automated A/B testing, generational AI being used in email subject lines, and data cleansing tech. Even if it works, this is all scaffolding and plumbing stuff that never made sense to lay people.
For a variety of reasons, luck could be a big one, OpenAI’s ChatGPT emerged at a moment and in a manner that made it different from all its predecessors. You could now see AI doing something. Text scrolled onto the screen as if created by a hyperactive typist who was also an expert in most things. ChatGPT does a thing that people understand.
OpenAI made the invisible hand visible.
4. Imaginations are on fire
So what now? Imaginations are on fire. How do you capture those sparks and fan them into real change?
Even though the skeptics can now see the power of AI they are also getting an earful about "ChatGPT hallucinations" and "Sydney (Bing) wants to rule the world". While the skepticism remains there’s also opportunity. Time to take advantage of the moment.
Here are a couple of thoughts:
Elevate the AI you’re already doing with decision-makers. The message is “that stuff you see being done in technologies like ChatGPT, that’s at work in a much quieter way in our implementation of X and it’s already delivering Y value”.
Experiment with chat-related technologies in your user experience. Small domain chat models, personalized content, and SEO optimization are places where LLMs can deliver more relevant content to users both on and off your domain without the hallucinations.
There are a number of companies working on Hybrid AI in knowledge discovery and customer engagement. Hybrid AI takes the best parts of conversational LLMs and couples them with reasoning to avoid hallucinations and increase decision transparency. Check out the work being done by Elemental Cognition as an example of Hybrid AI in action.
Look for companies that are specifically leveraging ChatGPT APIs (see also Elemental Cognition above). Despite the fact that it’s fun to play with, ChatGPT isn’t an end-user application. Its power is when it is combined with other technologies to deliver conversational features. Think of it as the voice of some other tech’s brain. Look for brains working on your use cases.
Gather a group of business people and tech folks—they all have to be AI-familiar (or at least AI-curious)—and set them free to determine what tech to experiment with and pilot emerging technologies in your processes and platforms against valuable problems. Don’t leave it to vendors to define your use cases. Seek them yourselves.
OpenAI’s ChatGPT and now Bing have been a significant unlock in helping people understand the power of AI. But don’t get trapped in these technologies or use cases. Use this moment to highlight the work you’re already doing, experiment with technologies that emerge in the next 12 months, and set your people free to run those experiments to find the value.