"We're gonna sell our way right to the apocalypse": AI in the sales process with Adam Honig (Spiro.ai)

Adam Honig from Spiro joins us to talk about AI, CRM, and Scarlett Johansson. We discuss how sales organizations are adapting to the emerging trend of using artificial intelligence to support the sales process, how Spiro is trying to replace the classic sales CRM with a more proactive, intelligent solution, and why AI in sales probably won't lead to global nuclear annihilation.

Adam Honig is the CEO and founder of Spiro. In this show, we discuss: 
  • Why Adam hates old-school sales CRMs (and why you should too)
  • The role that AI-enabled tools like Spiro can play in the sales process
  • Why salespeople (and podcast hosts!) shouldn't worry about AI taking their jobs anytime soon
FULL TRANSCRIPT:

Geordie Kaytes: Adam, can you give us a quick rundown of what Spiro does?

Adam Honig: Sure. So, you know, we really hate the term CRM,  and I think anybody who's ever had the misfortune to actually use CRM probably hates it as well.
And so we don't think of ourselves as a CRM is matter of fact, we think of ourselves as a software application that's designed to really kill CRM. And what do we mean by that? We believe that any time that the sales rep is doing data entry is wasted selling time.
We believe that if you know the software is not proactively doing data entry reminding salespeople who they should be following up with and basically making their life easier so the sales team can be much more productive, you shouldn't be using it.
And that's what's Spiro does, and it was kind of born from the the idea that using all this great AI technology that's out there today, we can read a sales reps e-mails back and forth with customers we can use that data to create contacts, to create companies, to update opportunities. When you send a proposal to a customer, why can't the sales stage just be shifted to "Proposed," for god sakes, and stuff like that.
And so basically doing all of the CRM "shit work," basically, the salespeople have to do, automatically for them in the background with a platform that then learns from that, and makes great recommendations about who the next person that you should call and email is. That's, in a nutshell, kind of what we're about.

Geordie Kaytes: Cool. And how did this all come about? I know you sounded like you had some frustrations with the basic CRM throughout your career?

Adam Honig: Well, it's true. I mean I so I've been working in CRM for about 20 years at this point, and my last company grew to be one of the largest consulting partners of Salesforce.com's before I sold that company 2012.
And we worked with thousands of companies, you know and the story was the same, you know, they would spend millions of dollars trying to get Salesforce to be customized for their needs, and then months later, like 10% of sales people would be using it! And it would be the same complaints over and over again.
We had a whole army of people who focus on what we call "change management," which was basically tricking salespeople into how to use CRM, it was terrible. You know, I would present at conferences, and I would say, "Hey, how many people are using Salesforce? How many people like it? And you know, some guy threw something at me.
That was kind of like level of — it's terrible — what this one guy said to me, "I feel like I'm in a bad relationship with Salesforce. All it does is take." You know, and I'm like God, so you know I spent 20 years of my professional life focused on this one thing that everybody hated it. Ugh! It was terrible.
And so then you know, I sold the company. And that was fine. And I went to go see a movie. I don't know if you're familiar with the movie "Her" at all?

Geordie Kaytes: Yes, oh, with Joaquin Phoenix?

Adam Honig: Joaquin Phoenix, he plays a man in the not-too-distant future who downloads a new operating system onto his phone which is all AI. Now, of course, it's played by the voice of Scarlett Johansson.
She doesn't actually appear in the movie, it's just her voice is enough to make Joaquin Phoenix fall in love with her. And, basically, she reads all this data, gives him advice, helps them out throughout his life, and it's an amazing experience for him.
And I'm watching this movie and I'm like, "Oh my God, this is what sales people need. They don't need Salesforce, they need Scarlett Johansson telling them what to do all day," and so that's how we started Spiro, with that vision. Unfortunately ScarJo wasn't available. She was like having a baby or something like that. So, you know, we never really got her involved in the project. But that idea is how we came up with with Spiro.

Geordie Kaytes: So I think there's a lot of misconceptions about — and given that you took it from a sci-fi movie — there's probably some misconceptions about what AI is capable of these days. Do you ever run into people think that it can either do more than it can, or they just don't even understand what it's capable of?

Adam Honig: Oh, like every day. Like any time people say, "Oh, you're in AI! I need an AI to help me make better meals at home." I'm like, okay, that sounds great. Or you know, "I really need AI to be my assistant to help me do everything in my law practice," and I'm just like, whaaaat... You know, I don't know what that's all about. What we use AI for, let me just give you an example.
So our AI, we use a part of AI that's called "natural language processing." And it basically, when you make a call with Spiro, it's listening to your call so we can take notes for you, and it can help you create reminders about things that you told a prospect you're going to do. And I know sales people should be able to do that by themselves, but it's great to have help.
That's a difficult problem to solve, you know, and it's it's in a very narrow domain, because the kind of things that salespeople will do as a next step, I mean, "We're going to set up another meeting," you know, "We're going to send you some information," "We're going to prepare a quote," there's a finite number of things that will actually happen so at least we're playing in that, and that's a hard problem. But to have like this generalized AI C-3PO just walking around, you know, giving you advice and telling you what to do? That is some crazy town stuff man.
That kind of problem is super hard, and I think it's going to be a long time. So if people are saying, "Oh, AI is going to take over my job. They're not going to need me to be a salesperson because there's going to be AI doing it," forget about it, that's going to be a long, long time before there's AI-powered podcast hosts who could just do your job here . I don't think that's coming at us anytime soon.

Geordie Kaytes: This job's not that hard. (Laughing) So,  the AI itself, how much you have to train it? How much do you have to tell it what your sales process is, versus how much does it kind of detect it and learn on its own?

Adam Honig: Yeah, so the way that we built Spiro, so — all AI technologies, are essentially experience-based, and so they learn by building a model of what they predict is going to happen and refine that model over time — so we when we first launched Spiro, we trained it with 15,000 sales reps who are using the product for over a year to make sure that we had all of the algorithms tuned and everything working properly. And fortunately for those folks, they all got to use the product for free while we did it and and we'd literally quiz them on how well Spiro is doing.
So every time we said, "Hey, you know, the data says you should call Joe," they would call Joe, and then we'd say, "How good was that recommendation?" We got the user validation up to about 93% and then we felt like that was very good, about as good as it was probably possible to make it. And that's when we decided that we were ready for Primetime with the product.
So there was a very long training process, in your words. Now, when you start using Spiro right out of the box, the algorithms and the approach that it's going to be using is based upon that data set that we used. But, as you use it more, Spiro learns. And so, for example, if they take somebody who said — just a crazy example — somebody who's selling Boeing 777s, their sales cycle is probably like 10 years. Or, you got a guy who's selling cars, it's about a day. So, Spiro's going to learn that sales cycle by seeing when deals closed.
And the frequency of interaction, just to give you one example, is going to change based upon how long that sales cycle is. Maybe if you've got a year-long sales cycle, you need to speak to your prospects every month or so, until it's getting closer to the sale. Maybe if you have a one-week sales process, you better be calling them every day, otherwise, it's never going to happen.
So Spiro starts to learn those things about the individual sales team's approach to things, and it folds that into the way that it recommends action.

Geordie Kaytes: How does it deal with when you've got to talk with more than one person over on the client side, when you've got a team of buyers?

Adam Honig: Yeah. Well, this is really interesting. So, a very common problem in sales for for many many organizations, of course, is that you and I go meet with Staples, and we're trying to go sell them something, and there's five people there, but there's one person that we think is really the buyer and we focus on her.
Now, what Spiro does, because it's automatically capturing contact information —  like in this particular example, it knows because all of these other people were in the meeting invite for that meeting that we had, it will go ahead and create them as contacts in Spiro — and as you're starting your reach out to Staples in this particular example, it's rotating around the different people that it thinks that you should talk to, and it notices the responses that you're getting from different people, and it's making sure that you're getting the full picture of the deal.
Now, you can basically literally say to Spiro, "Hey, listen this person, you know, they're not involved in the sale. I don't ever need to talk to them," whatever. But by default, it's going to make sure that you've got account coverage, to make sure you really know what's going on.

Geordie Kaytes: So, when you've gotten Spiro involved in really complex multi-phase sales processes, it seems to be able to handle those just as well as it can, you know, something relatively simple like selling a car from your example?

Adam Honig: So Spiro is really designed for B2B sales, and we're designed for sales processes that have a sales cycle, so a core value of what we're doing is always making sure that you're following up with people in your pipeline through proactive recommendations. And if you have a very short sale cycle, or where somebody basically walks into the the dealer lot and buys something, you probably don't need Spiro. But if you're selling to businesses, and you've got a sales cycle that goes anywhere from like a month to a year, that's definitely our sweet spot.

Geordie Kaytes: And how does Spiro deal with it once you've got multiple people on your side as well, as in "team sales" kind of situations?

Adam Honig: So, Spiro's always modeling behavior. So, you and I are on a sales team, you're reaching out to certain types of contacts, I'm reaching out to different types of contacts, you know, it starts to learn and understand where people's strengths are.
So, let me give you an example. Every time you make a call with Spiro, you call a contact and when you hang up the phone Spiro asks you, "Hey, how did that go?" And you disposition it, "Good," "Okay," or "Poor," and based upon how your calls go, in the context of the call Spiro starts to say, okay, Adam should be calling these people, or whoever, it starts to put that kind of model together for you automatically.

Geordie Kaytes:
So it basically acts as this air traffic control for your sales team, in a way.

Adam Honig: Exactly. Exactly. We like to think of it as creating a self writing to-do list. You know, every sales person out there, I'm sure they remember their top deals that they're working on, so they get in in the morning, they call their top deal, and then they're like, okay, what should I do now? Well, Spiro has that answer for you.

Geordie Kaytes: And you mentioned some people wondering if Spiro's going to replace their jobs. Besides that, what kind of pushback have you gotten from from salespeople and sales managers?

Adam Honig: Well, you know, I think there's a lot of pushback against CRM. So, in our research and since we've been running the company, there's actually a lot of companies in the world that don't have CRM. And when I talk to sales leaders at these companies, and I say, "Well how — you've got this 50-person sales team — how are you not using a CRM?" And their response to me is typically some version of, "I've seen CRM.  It's terrible." And so we have to overcome this this issue, the the ghost of CRMs of the past, I like to call it, and so that's that's definitely a big issue.
And for some customers, you know, the AI that we have is very dependent upon data. And so Spiro works best when it's connected to your email and understands what's going on in there, it can read your calendar, can listen to your phone calls. We don't — we're not Facebook, we're not selling that data to anybody — but people have privacy concerns and I totally respect that. But you know, of course, there's no human in the loop here looking at your data or anything like that. It's the AI, and it's only there to serve you, but that's an obstacle sometimes.

Geordie Kaytes: Okay, so people aren't super happy about having all their sales conversations be shipped up to the cloud?

Adam Honig: Some people are concerned about that. Yes.

Geordie Kaytes: Anything else that people get a little bit weirded out about, or doubt that Spiro is able to do?

Adam Honig: Well, one of the things that people sometimes get weirded out about is the fact that we've built a bunch of different sales personalities into Spiro. So this is kind of funny, but going back to the whole "Her" thing, we decided what would make Spiro a lot more interesting is if you could basically select the way that Spiro communicates with you.
And so within Spiro, the default is a mode that we call "Coach," and so the messages that you get from Spiro are sort of like a positive affirmation, like helping you through your day and stuff like that, but that might not be for everybody. You might need the "Jewish mother" personality to help you through your day, which is very sarcastic, you know, and says things to you like, "Hey, I noticed you didn't call Mike. I bet your brother would have done it by now," or whatever.
And you know with Spiro — because hey, listen, you know, we're in business, we're helping people sell, but we still like to have a good time here. So just to be interesting, we've built in six or seven different personalities into Spiro to give you sort of a flavor for, you know, whatever kind of personality works best for you. So sometimes that's a little bit of an explanation,

Geordie Kaytes: And what kind of prerequisites or processes or structure does a company have to have already in place to benefit from bringing an AI solution like Spiro into their sales process?

Adam Honig: One of the really great things about Spiro is that we are — because the platform is built with AI and it learns your sales process as you go, there's almost no time to set up or configure it for your organization. And the reason why that's relevant to your question is because there's no prerequisite on "you have to have a sales process," even, because Spiro will kind of figure it out for you.
I mean, people might not have a defined sales process, but they're still making calls. They're still sending quotes or proposals. They're still trying to close deals and so the basics are there. You know if you don't have a product, I mean, of course Spiro is not a very good solution for you — you're not selling anything for god sakes — but if you're doing business, I think that there's no real prerequisite.
We have this customer, one of our early customers is a company that manufactures cardboard boxes. And it's not the most glamorous industry or anything like that. But you know, they, at this point today, they've got about 50 sales people selling cardboard boxes all around the country. Procter & Gamble is one of their customers, and stuff like that.
And every one of their sales teams does things completely differently. And that's just the way that their business works, and they're a traditional business. They've been doing it for a hundred years, they're not going to change, and so Spiro is able to detect that that's just the way things are, and so it works with the different groups at their own speed basically.

Geordie Kaytes: Okay, and can you ask it at the end what your sales process is? Is it able to spit that out in a human-understandable way?

Adam Honig: It's really not, although it's a really interesting idea. So one of the things about Spiro, which it does do — which isn't like in the same exact way that you asked me — but the product comes with a very rich reporting environment, and one of the reasons why sales leaders buy Spiro is because they really want to understand the sales process.
They want to understand the gaps. They want to understand their forecasts, and so on. And so we've built this incredible reporting environment where you can actually literally see that happen. It doesn't spit it out in like a documentary, human-understandable way, but if you want to see the progression of deals through stages, you want to see the timing and how — by region or by group or anything like that — the time differences as deals go through stages, our reporting environment does all of that, and the great thing about it is that since we're basically not relying upon the sales team for data entry, the data actually turns out to be really accurate.
And I think we did one survey for a large telecom company to compare the data collected in Spiro versus their old Salesforce instance, and we actually captured nine times more data than Salesforce.com, just because of the way that we built the product. So, not human readable, but, you know report readable if you know what I mean.

Geordie Kaytes: Yeah. Are there any other — you know, besides what Spiro does today — were there any other avenues that you went down when you were developing it, around applying AI to the sales process, that you decided were not the direction you wanted to go in, out of the gate at least?

Adam Honig: Yeah, so, I'll tell you one kind of dumb example, and one more serious example. We had this theory when we started the business that when salesperson completed an action in Spiro, we should give them a reward. And so, in the initial version of Spiro, you could choose a popular Instagram feed, whether it's like motivational quotes, or Victoria's Secret images, or surfing, or whatever, and after you had a successful call, boom! It would pop up this image. And we really thought that might be a fun way to help motivate folks to make more calls and do more activities and stuff like that. It turned out actually decreased their activity levels. So that was something that we had to kind of pull out of the product.

Geordie Kaytes: Huh. Do you know why?

Adam Honig: You know, we were never really able to get a handle on that, unfortunately. I mean, sometimes the data is just inconclusive, you know?
So, the other thing that we're working on is — you know, when you work in AI, one of the really important things is allowing users to be able to trust the system.
You know, a lot of the AI is kind of black box, like it's really hard to pin down why it's recommending one course of action versus the other. And so instead of just making it so that Spiro makes recommendations about who you should be following up with and so on, now the users can also explicitly set their own recommendations and criteria for how they're doing that.
So it's like, "Hey, next Tuesday, I need to call all of my customers who have this sort of criteria because we're running a sale," or I don't know what, some sort of reason right. I'm flying to Houston, so I want to call all of my customers in Houston to set up some meetings, or something like that. And so we what we had to do is adapt Spiro to not just be AI, but also allow the users to do their own thing with with the product as well. That was a lesson learned for us.

Geordie Kaytes: So you talked little bit about this is mostly or pretty much entirely for B2B sales. Is there a difference in Spiro for selling services versus selling products?

Adam Honig: Well we have quite a few customers in the Professional Services industry, and there's some nuances involved in that that —especially in reporting — come out.
One of our customers is a large digital agency and they run social media campaigns on behalf of their customers and they sell those campaigns with Spiro. And then they need to be able to forecast the revenue from those campaigns. And I don't know if this sounds familiar to you at all, but you know, to forecast the revenue of a Professional Services engagement, it is spread out over a period in some sort of unique way depending upon what the project is. And so that type of reporting is something that we do in Spiro, we enable for them.
It's just very different than the type of forecast reporting that a product company does, because quite often their revenue is very discrete. "Hey, we sold a widget, and we shipped the widget, so its revenue in this period," versus a project that might span six months that you have to understand the revenue spread of that project over a period of time to make sure that you're reporting your sales properly.

Geordie Kaytes: Okay, so integrating that reporting into the... I hate to say CRM, but into Spiro, is where you see most of those differences — do you call it a CRM?

Adam Honig: Sometimes, depending on the day of the week. We're working on a new thing that we're going to call ourselves.

Geordie Kaytes: Are you allowed to say what it is yet?

Adam Honig: Yeah, I'm not I'm not because we're not sure of it yet, but I know it's gonna be it's gonna be much better than CRM. It's gonna make CRM look like yesterday's VW bus.

Geordie Kaytes: So if Spiro learns your sales process, and then you want to change your sales process, how does that work? Are you able to push top-down changes, or tell Spiro, "yeah, everything we were doing back then is wrong, we're going to do this this new thing going forward"?

Adam Honig: Yeah, so there's a number of different ways that you can do that within Spiro by changing the data itself. And so one of the things that Spiro — like I mentioned, it learns the expected sales cycle, right? — and so it starts to populate expected close dates for deals.
So if you're going to introduce new products, and maybe you think it's gonna have a different sales cycle, you can explicitly tell Spiro, "okay, these deals are going to take longer to close," and Spiro will go, "okay great. I know how to handle that." Or you know, there's different ways within Spiro you can do that.
We're planning on building — and this is you know a project that's going to probably take us the whole year — sort of an explicit interface that allows you to understand what all of the levers are within Spiro, and so our customers can can modify the modulators, if you will, within the product to make things a lot more tuned to what do they want, but that's a little bit more off for the product.

Geordie Kaytes:  Does Spiro take into account any kind of differences between customers? So, what their firm segment might be, the individual personas that you might be talking to over on the prospect side, how does that get pulled into the model?

Adam Honig: Yeah, so I would say that Spiro is very context sensitive, and not content sensitive.
And so what I mean by that is, we don't look to see what somebody's title is to make a recommendation about who to call. What we do is, we see that you're calling that person a lot, and you're having good conversations, and the sentiment of things seem to be going the right direction. We're looking at the contact itself, not their title .
But you know, there is definitely some industries that Spiro doesn't work  that we've learned as well, and so like pharmaceutical sales, for example,  it's just a model that Spiro doesn't fit well within, because people who do pharmaceutical sales, they're not really like taking orders, right? They're not really producing quotes. They're influencing doctors to write prescriptions in a certain way, and stuff like that. And that's just not our thing at all.
So there's definitely some nuances within industries that do really impact things, but in our business, we're really focused on what I call "traditional industries." And so this is manufacturing, logistics, professional services,  industries is that are not often very "tech forward" in their way of doing business, and have often looked at these CRM solutions and rejected them because they suck.

Geordie Kaytes: That's funny, because they're not "tech forward," but they're adopting AI into their sales process. That sounds like...

Adam Honig: I know! I know, that's the whole thing. So we're leapfrogging those other people. I don't know what to say. But since the AI — like the core of the AI is, at Spiro, it's making this  type of sales software usable by people, you know, it's kind of getting the software out of the way.

Geordie Kaytes: Now, you talked a little bit about your plans for the future. Just broadly, even outside of Spiro, where do you see AI going in the sales process, and supporting the sales process over the next five or ten years?

Adam Honig: Well, I think there's so much work to be done. Google has done a great job of laying out some of their vision for the way that AI will be used, and I firmly believe that all business applications, even beyond sales, will be using AI in very short order.
And I think for me, the vision of using AI is really all about proactivity. Like, almost every bit of software that we're using today is dependent upon the user initiating something, and I just think that's backwards. Why can't your accounting software be — you know, when you come in in the morning and say, "Hey, this journal entry looks wrong because I analyzed it while you weren't here," or whatever the thing is.
And so this level of having work come to you, as opposed to you going to the work, I think is going to be the biggest revolution of how AI is going to impact people.

Geordie Kaytes: And you see people kind of freaking out about that, because of a loss of control, or do you see that being just something that people will get used to? How do you think we're going to adapt to this socially?

Adam Honig: I know that there's people who are all worried about AI and whatnot, but I don't really feel like there's a lot of fear of AI. I mean,  everybody's seen the Terminator, right? So they know it can go wrong. But  the examples of AI that are being used today are so pedestrian, I don't think that people are really worried about that so much. Not the people that I'm working with.
And I think in most cases it comes down to the fact that you can show a return. It's like, "Hey, we can use this AI and it's going to save you time or money. Would you be interested?" And people are like, "I don't care whether it's AI or a bunch of stones. If you can save me some money, I want to talk to you."
So, I think I think it's going to be fine. We're gonna sell our way right to the apocalypse. (laughing)

Geordie Kaytes: Do you think there's anything that AI will never be able to do, at least in relation to supporting sales and the sales process?

Adam Honig: Well, you know, we do a lot of work with salespeople and we publish a lot of content. We've got a great blog by the way. I'll just give it a boost for a second. So we publish all kinds of amazing sales content at Spiro.ai/blog. We write funny blogs. We write serious blogs. Our most popular blog is "Seven reasons why everybody should marry a salesperson," people love the blogs.
Anyway, so we write a lot about sales and selling and everything like that, so we spent a lot of time about rapport building, and about making a human connection with somebody. Really listening, and trying to understand what their needs are. And you know, just based upon how hard we're working in AI to do these very basic things, I just really believe it's going to be forever before AI can empathize with you and ask you the right questions, or really be the role of a salesperson. I mean, I for one am just on the side of, that is just never going to be what the AI is going to do.
I mean, chess. You know, winning at chess is a big win for the AI community, right, but compare that to the thousands, millions of permutations that happen every second during a sales call. I don't know how we're going to solve that.

Geordie Kaytes:  Alright, well, thank you. Thank you so much for giving me so much of your time today.
Is there anything that you'd want to close off with, let people know about especially where they can find you?

Adam Honig: Yeah, sure. So, I mean if you're interested in learning more about Spiro, our website is Spiro.ai, and we have a lot of information there about the product, about, you know, as I mentioned before, many many funny blogs on the side.
You can certainly sign up for a free trial or a demo right there on the website as well.

Geordie Kaytes:
Cool. All right, Adam. Well, thank you so much for your time today.

Adam Honig: Yeah, my pleasure, good talking with you. 

Join our newsletter

Got it. You're on the list!
© 2019 Geordie Kaytes