Skip to content

Sales AI: Is the revolution coming?

Liston Witherill
Liston Witherill
13 min read

For more information on remote selling and a complete list of links mentioned in this podcast, visit this remote selling article on our website.

Sales AI: Is the revolution coming?:

Full Transcript

Is love only for people or could you love a computer just as much? Now, I love my dog and I love my cat as much or probably more than some people, but I don’t love my computer. This was the central question in the 2013 movie, Her, written and directed by Spike Jones. The main character of the movie finds the love of his life in a computer program. Farfetched? Well, a few years back, I was at a startup pitch event and one of the companies there made an intelligent support bot that they tested by letting people chat with it online, but they didn’t tell the human subjects that they were chatting with a robot. One of the testers actually fell in love with the chat bot. I’m not making this up. This actually happened. And when they told the guy that he was in love with a robot, he didn’t care.

Is that the promise of AI? A full and complete replacement of humans? Or are there smaller, more specific things that AI will do to supplement rather than replace us? In this episode of modern Sales, we’ll cover AI in sales, what it is and what it’s not to discover the truth behind the AI revolution and whether it’s actually coming.

Welcome to Modern Sales, a podcast for entrepreneurs, business owners, and salespeople looking to have more and better conversations with your perfect clients. You’ll get a healthy scoop of psychology, behavioral economics and sales studies to help you create win-win relationships. I’m your host, Liston Witherill, and I’m pleased to welcome you to Modern Sales.

Welcome once again to Modern Sales. My name is Liston Witherill, and this is the sales AI series on Modern Sales where we’re taking a deep dive into the world of artificial intelligence, and specifically, how it will impact sales. See, AI, it’s pretty confusing, or at least I find it confusing. So in this series, I want to demystify some of it. I want to take some time to learn about it, its potential impact and how it’s being used today, or if AI is even a thing in the first place. I’ll talk about its likely impact on jobs and what we can expect from AI now and in the future.

But seriously, this series, it is a little selfish. I wanted to know more about AI and I thought you might, too. You see, when most people talk about AI, they’re talking about a very small part of it that’s not around now and may take 50 years to come to be. In fact, it may never happen at all. I’ll tell you what it is right after this short break.

Welcome back. The big promise of AI, it’s this thing called general intelligence or artificial general intelligence, which is the biggest fear and promise of AI. It’s Skynet from Terminator. It’s what destroyed the world in The Matrix. It’s what Joaquin Phoenix fell in love with in the movie Her that I mentioned at the top of the show, and it’s the brains behind HAL 9000 in the 1968 film, 2001: A Space Odyssey.

Obviously, the media, entertainment industry, they’re very worried about AI. It’s sensational. It grabs your attention and it stokes fear in some of us, but it also stokes fear in people who really know what’s going on, like Elon Musk or Stephen Hawking. They’ve come out publicly against some of the dangers of AI, but they’ve also talked about some of the benefits of it, about some of the promise, and a lot of other smart people have done the same thing.

But here’s the problem. Artificial general intelligence, the main thing that we think about when we hear the term AI, it doesn’t exist, at least not yet. But if you do a Google trend search to see how AI is trending as a search term, you’ll see there’s a lot of interest out there. So if AI isn’t really general intelligence, what is it then?

Before I get into that, I want to mention really quickly, I want to give a big shout out and thank you to my good friend, Dan. He helped me out a ton in creating this episode and this series, so big shout out to Dan. This series would not be possible without his efforts. So thank you, my friend Dan. He is building a chat bot right now and he is one of the leaders in the space, so he was instrumental in putting this together.

So what is AI? Most textbooks define AI as, and I quote, “Any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.” Right, so what the hell does that mean, I’m sure, right? So instead of looking at AI generally, let’s look at sales specifically. Let’s break it down. It’s a device, which could be a computer program, it could be a robot, like some of the videos you may have seen about Boston Dynamics. It could be your phone. I have Google Assistant in my phone and it seems pretty smart even though it’s not, but I’ll get to that later.

But if we look at and dig in what AI does specifically rather than generally, it starts to make a lot more sense. Basically, when most people say AI, what they’re really referring to is machine learning, algorithms and data science, which is to say that the term AI is a mainstream pop term. It’s not a term that has much specific meaning. Now, if you go to the Wikipedia page, what’s interesting is they say AI is basically all the stuff that we haven’t done yet. As soon as we’ve done it, we no longer consider it cutting edge and therefore it’s not AI. So I’ll let you define it however you want to, but again, I want to focus on how this is going to help us in sales, whether you’re in individual or at a really big company.

If we get specific, let’s break apart, what are all the jobs that need to be done within sales? I would say research, prospecting strategy, direct communication, which is like the one-to-one, I’m sending emails, I’m talking to people in meetings, I’m going and doing pitches, that includes negotiation, persuasion, change management. Then there’s the project management side, collaboration planning, resource utilization, organization, and then there’s the management and training side within companies. Surely there’s more, but those are the major ones.

And right now, AI can contribute to each of these fields, but it can’t completely take any of them over. To prove that point, let’s take a look at two of the most common AIs, and I’m using air quotes when I say that, but I’ll just keep referring to it as AI so we’re on the same page. But two of the main ones you’ve probably heard lots about are chat bots and lead generation assistants. Now, those can mean lots of different things, but just know that each of them has its own set of requirements, which brings me to the next big point. AI is not a single thing. Usually it’s a collection of technologies that work together to do “smart things”.

I’ll get into what each of those components is, but first, I wanted to let you know about another podcast that you might enjoy. If you’re a regular listener of the show, and thank you if you are … please hit subscribe if you haven’t already … you’ll probably also really enjoy the B2B Sales Show from Sweet Fish Media. You might want to check out the episode on effective account multithreading with Peter Chun, VP of sales over at Lucid Chart.

So back to AI now. Let’s talk about the component parts. I mentioned chat bots and lead generation. Let’s start with lead generation assistants, because they’re a little bit easier to understand. Really what they’re doing is not complicated at all. I’ve used a few tools that promise to help personalize outreach by surfacing valuable research. So essentially what it did was I would put in someone’s email address and then the “AI” would go take that email address and resolve it on the internet with sources of information and then feed that information back to me and give me a few options of personalized sentences that I could use in my outreach. Here’s my takeaway. They were all horrible. It took me more time to dig through all of the crappy research that they surfaced for me than it would take me to just look up each lead on LinkedIn and make my own decisions. And so it totally wasn’t worth it.

One use of a lead generation assistant could be research. Another use could be in creating your lists in the first place, which I would also define as research. But let’s say we define lead generation simply as a program that can feed us contacts who are a good match for us to prospect. Simple enough, right? This isn’t AI so much as it’s an algorithm. I mean, you can even argue that if it is AI, then LinkedIn is AI because you can set up a search and it will tell you new people who match your search all the time in sales navigator. It’s pretty easy, right?

So the component parts you need in place for this are just a database, a definition of the leads you want back from your query and a way to deliver those leads. In LinkedIn, you would just get a notification message. Other lead generation platforms will push the content right into your CRM with people’s first, last name, email address, even direct dial phone numbers. But I’d hesitate to call this intelligence because the system isn’t making any judgment calls whatsoever. It’s just running a program, an algorithm, right? It’s following a set of rules that I fed it. It’s really not doing anything that resembles intelligence. So in that case, I would say lead generation, just not there. There is no AI really that I am aware of.

Now let’s look at something that’s a lot harder, and that’s chat bots. I’m not going to name names of companies, but if you want to sponsor this podcast, please reach out. There’s a lot of companies making chat bots, and the promise is when someone comes to your website, they can fill out instead of a form, a chat window where they’re chatting with a bot that can quickly respond to their queries and maybe even book meetings with them or point them in the right direction and get them closer to the middle or bottom of the sales funnel.

I’m going to get a little technical here. Hang on for dear life. Thank you again to Dan for this discussion. I’m going to dig more into chat bots in a later episode and get more specific about them and the chances of them taking your job. But for now, we can use them as an example of where AI is today, like what is the state of the art right now and where it has to go. Chat bots have three main parts. First is the natural language processing. So anytime language is coming in, it needs to figure out what does it mean, what is the context of this language, what is the person’s intent? The second is a handling layer or logic engine, and lastly, connectivity. What happens to this information once we get it?

So let’s pull it all apart. If someone visits your website and begins to chat with a robot, that robot needs to be able to understand language. That’s the language processing function. It needs to make sense out of what you’re telling it. Next, the robot needs instructions on what to do based on the intent of the person chatting with it. That’s the logic engine and because language is such a messy contextual business, this turns out to be a really freaking hard problem to solve. Maybe the chat bot is supposed to be booking a meeting or linking the person to a help article or giving them some dated content so that you capture the lead, whatever. It needs to make the right decision.

And finally, it needs a mechanism to move and store the data and usually that’s just passing it right into your CRM. Maybe that would trigger some sort of email follow up. Maybe it would put a phone number in your phone queue or assign it to someone on your team, whatever it is, right? Basically, again, there’s nothing really intelligent happening there so much as a series of steps you’ve already told the system what to do, right? It’s not going in there and making decisions for you, which is all to say that a lot needs to go right for it to work and a lot of humanness and a lot of human intervention needs to be part of the system.

Just as if you were to hire a person to sell for you today or you were trying to improve your own sales skills, you would need to train to do that. So if I’m hiring someone, I need to train them on the way we sell here and if I’m trying to improve my own sales skills, I need to go out and find out what are some of the ways that I can improve my sales skills. What’s the right process? What are the right words that I can use? What do clients mean when they say X, Y, or Z? What are standard objections? How can I negotiate? On and on and on and on, right? The same is true for a chat bot or an AI can’t figure that out on its own. You have to tell it, and we’re not at a place where it can just figure it out and optimize it.

Now, there was recently a “meeting of the minds” of some of the leading AI researchers who are all academics and they met at a very nondescript hotel conference center. And what they said is the big missing piece is systems being able to train themselves on how accurate their guesses are. Right? For true intelligence, typically we have to make assumptions and start projecting into the future. We need deductive reasoning, we need logic, but ultimately it relies on us making judgment calls, and those judgment calls being reliably good. And AI is just not there yet.

What are the basic uses of AI? Well, I talked about the basic sales functions, so the best uses of AI fall under those categories. Just as a refresher, those functions are research, prospecting, strategy, direct communication, project management and of course training and performance management. And I think there’s a lot of opportunity there.

Now for the biggest teams, data enrichment will be a huge use of AI to help teams target buyers that are most likely to buy right now. I see this as a really, really good use of AI based on, let’s say, how forthcoming the other team is with communication and with scheduling meetings, how many people are involved in the sale, what external signals can come from the company. There’s been a tool for years now that allows you to predict how soon a customer of a software product will churn out based on the average lifetime of customers of that product, so it’s not really intelligence but still useful information that people aren’t very good at gathering.

Imagine for a second if software could go out and collect data from all across the internet and put together new insights that would inform you about your buyers. I think that’s a really big opportunity. But I would also say there’s an inherent tension here. The problem as I see it is that there will be no longterm advantage to that and the reason is if one company can do it, another can too. The technology evolves. It has an impact for sure, but the companies that nail the fundamentals will preserve the largest advantage in the marketplace. In other words, these AI technologies, yes, they will be helpful, but they won’t be a sustainable advantage. Eventually, as I said, instead of being state of the art, they’ll just be regular old tools that everybody uses.

Now, one perspective I hadn’t considered when I conceived of this series came from a LinkedIn comment that I got when I posted about this episode and that comment said that AI will disrupt sales, not because it will be so good at selling, but because it will be good at buying. Now, I think that’s a really, really smart insight. However, I still think that there are intangibles. I think that for it to be really good at buying and for it to be reliable, people have to have faith in its choices and for them to have faith and its choices, it has to mimic human decision making and for that to happen, my friends, humans have to be so good at programming the AI that it practically duplicates their choices, and I just don’t see that we’re going to be capable of getting that granular. I could be totally wrong. The longer time that goes between this episode and you listening to it, the more likely I am to be wrong, but I’ll go ahead and hang my hat on that.

I think in the long run, AI will not be a sustainable advantage as it becomes more democratized and cheaper, which brings us to the next question. How good is it? Well, it’s really good if really good people are there to master the tech and program it for the specific use cases needed at each company, which is to say that the software itself is just a tool. I can go to Home Depot, I can buy the best tools, I can spend more money than any other carpenter. I can buy all the books. I can get all the best materials, but a local carpenter will still do 10 times better work than me because she knows how to use the tools better than I do. It’s not about necessarily the quality of the tool, it’s about how she uses the tool. And I think it’s the same exact idea here. AI will only be as good as the operators of the AI until we reach that point of general artificial intelligence. And when we do, I have to say all bets are really off. We have no idea what would happen next.

Coming up in the rest of the series, I’m going to cover three major questions. Who’s going to use AI? Will it take your job or could it do the sales function for you? And in the last episode of the series, I’ll be covering how it’s used right now. That’s it for the first episode in the Sales AI series. And next week, we’ll talk about what AI has in common with the banking industry in the 1960s.

If you aren’t already subscribed to this podcast, please do so. Hit the subscribe button. You’ll get notified of each new episode. It’s really as easy as that. You can also sign up for my email newsletter for daily sales insights at It’s totally free and it’s linked in the show notes. Thanks to everyone who makes this podcast possible. Juan Perez is our editor. Mary Ann Nocum is our show assistant. Our show theme and ad music and some of the episode music are produced by me, yes, by me, Liston Witherill, and all other show music is by Logan Nicholson at Music for Makers. Thanks so much for listening. I’m Liston Witherill of Serve Don’t Sell, and I hope you have a fantastic day.

Modern Sales Podcast