Some of you may have been following my adventures and experiments with ChatGPT. It and similar tools are fascinating, and while being very early in their development, offer interesting potential in the coming years.
When makinq queries in ChatGPT, the responses represent faux–dare I say–artificial intelligence. The responses “seem” smart, but aren’t particularly insightful. The tool assembles a collection of well constructed sentences, identifying very high level issues, presenting balanced observations. But nothing new. No insights that cause one to think, “That’s interesting, I’ve never considered this before.”
When I engage the tool in a “conversation,” responding to the observations, but trying to dive deeper, getting more specific and personalized to a scenario, ChatGPT fails. It never gets beyond generalities, it never provides an “Aha” moment.
To some degree, it’s not dissimilar to getting a prospecting call. The SDR has a script, it may be well constructed and delivered, but the moment one tries to dive deeper, the SDR can’t respond. Many of you may recall my adventures with a call when the SDR said, “We think your business is underperforming it’s potential. We think you can do much better.” My response was, “That’s interesting, what am I doing wrong?” The SDR couldn’t respond, the only thing he could do was to ask about scheduling a demo.
None of this was the SDRs fault, it was the fact the SDR was expected to read a script and not have a conversation.
My experience with the tool is similar to the conversation, perhaps even less challenging than the SDR was.
But there are some interesting aspects to the tool. It’s incredibly easy to generate new content. For example, I asked, “What are the 10 areas that could most improve a salesperson’s ability to prospect and engage a customer in conversation? Within a minute I had a nice response with 10 questions. Everything I expected. “Narrow focus to the ideal customer profile; Research the customer;…..” I turned around, using each of the questions to engage ChatGPT. For example, “How should a sales person research a customer?” The answers weren’t surprising–“Go to the website, look for press releases, go to social sites, …..”
I found, within about 10-15 minutes, I created about 130 pieces of content–content that had, more or less, smart words, but not particularly insightful and nothing that would cause me to think , “I’ve never considered this….”
So we are, right now, in a world where the incremental costs of developing content are virtually $0. If we look at it now, without these tools, it takes work on the part of someone. They have to come up with an idea, they have to research it–at least a few Google queries, they have to write it, grammar check, then publish it.
All of this takes time. While the dollar cost may be small, the time investment of the individual may be pretty big. Think about going to a marketing person asking, “I need 10 blog posts on this topic, when can you deliver it?” Alternatively, “I need 10 email sequences on this topic, when can you get it to me?”
There is a pretty hefty cost to developing this. We always have to evaluate, “Could we use that person’s time in a more impactful way?”
With tools like ChatGPT, the incremental cost of virtually all content development is close to $0. We just have to give a few general prompts, then drill down into each of the statements the tool returns, then do it again, and again, and again…. We can generate any volume of content without having to think, research, write, edit. And it costs us virtually nothing.
This means there is no limit to the content we can develop on any subject. We can generate infinite volumes of content in just a few minutes. We can even personalize it in some ways. For example if you look at the questions I asked earlier, I could insert a market–how to we prospect healthcare, financial services, technology…. The responses I would have gotten would probably be similar, but perhaps using industry/market terminology.
For decades, content has been a limiter in our ability to reach out and engage customers. Tools like ChatGPT have eliminated that as a barrier.
But what about the customer? We already know they are drowning in information. We know they are overwhelmed with high quality content, often very contradictory. They struggle to make sense of it, they don’t know how to evaluate and sort through it, finding the most important and relevant information for their situation.
Now, we’ve unleashed the capability to, at virtually no cost, inflict 10 times, 100 times, 1000 times more content on those already overwhelmed customers! And that is what will happen. Even if each of us are well intended, the cumulative impact of everyone doing the same thing and reaching out to customers is enormous.
And what will be the response of the customers? When they are faced with overwhelm, they shut down. The stop searching, they may even just stop projects, thinking the hassle of finding the right information in the petabytes of data inflicted on them isn’t worth it.
We’ve seen customers increasingly leverage digital channels to learn about problems, solutions, getting the information that helps them with their problem solving and buying process. But with infinite amounts of content now available to them, these digital channels become useless. We will force them to look in other places or to do nothing.
Some of you may be thinking, “Dave you are exaggerating the problem!”
But we’ve already seen this in other areas. Think about the old days, when we wanted to send content to a customer. We had to develop it, we had to print it out, we had to stick it in an envelope and affix a postage stamp to it. There was a time and $ cost to this process. We had to be very careful about what we sent, it’s cost. We had to be very focused on who we sent it to, sending it to those we thought might be most interested.
But email, social, and the web changed all of that. We still have the cost of content development, but no cost in content production, no mailing costs, and virtually no time delay in getting that to the customer after hitting send. It has enabled us to constantly send lots of “Can we talk so I can sell you something” requests to 10’s of thousands of customers in just a few minutes. And customers have shut the majority of that down. Just look at open’s, reads, click throughs and how they represent hundredths of percentages of what we drown them with.
And we’ve seen the same with phones. When it cost us a few dollars to make a long distance or international call, we were more focused on what we did. Now, a call costs virtually nothing and we have tools that enable us to make 100s of calls a day.
Customers are shutting those channels down. We are struggling increasingly to catch their attention and get them to respond. But they don’t want to have sales people talking to them,
And now we can do more! Infinitely more! We can inflict endless amounts of mediocre and meaningless digital content on these same customers for an incremental cost of $0.
Buckle in! It’s going to be an interesting ride!
Aftwerword: Ezra Klein has a fascinating podcast on just this topic: A Skeptical Take On The AI Revolution