AI For Sales - Smart Chat Bots? With Stephen Lowisz
There is the misconception that automation will cause unemployment, but that isn’t at all true. Automation and smart chatbots are a tool, not a replacement. It is intended to help employees work more efficiently, not replace them. In today’s episode, Chad Burmeister chats with Stephen Lowisz, the Vice President of Sales for Roots Automation. Chad and Stephen discuss the power of AI and how process automation will drive employee and customer satisfaction. Stephen also tackles the common misconception that AI and automation will cause unemployment, as he believes using these tools will benefit employees in the long run.
Listen to the podcast here:
AI For Sales - Smart Chat Bots? With Stephen Lowisz
I’m here with Stephen Lowisz with Roots Automation. He heads sales and marketing. The organization has been around for a couple of years. We’re going to talk a little bit about how their organization leverages artificial intelligence and then we’re going to dig into what their sales team does in leveraging AI as well. Stephen, welcome to the show. Thanks for being here.
I appreciate the time.
It’s great to have you. Where on planet Earth are you located now?
The corporate office is in New York. However, I’m in Detroit Metro out of the burbs. It’s nice and quiet out here.
I like to share with our audience a little bit more about yourself, about how you got to where you are. The best way to do that is to understand when you were younger, maybe even 6 or 7 years old, what was the thing that you were most passionate about? When you woke up in the morning and you said, “I got to go do,” what was that?
I’m a little sadistic. I’ve always been in sales, even when I was younger. I was walking around 7 or 8 years old with pooper scooper flyers, then it turned into dog-walking flyers and then into landscaping. I come from an entrepreneurial family. The bug rubbed off on me. I saw the progression that my father had made in his career. Honestly, I was the nerd that’s at ten years old looking at lawnmowers, “How can I make money with this?” From Home Depot online, that sort of a deal. Eventually. I joined the family business though I was the lowest paid person, hardest working person for a few years. I sucked at sales if something clicked. I did my first $1.1 million at nineteen, got to build out some sales and marketing teams as a consultant for companies like ESPN, Citrix, Cenduit. I went split out on my own. Roots Automation was a consulting client of mine that convinced me to drink the Kool-Aid, so to speak, so here I am.
It’s a neat application and it’s cognitive process automation so much more than RPA. A lot of people are starting to hear this term RPA. Tell us a little bit about how that passion when you were a kid now ties into the work that you’re doing.
If I had to do it over again, I’d be a developer. I’m a straight-up nerd but being in an entrepreneurial family, you learn how to sell, not how to code. The cognitive process automation and all the big fancy buzz words I get to tour around make me feel a little bit smarter than I am. I’m blunt. With that being said, most people are now hearing of RPA, Robotic Process Automation. Think about it like a piano. Have you ever seen a piano go on autoplay, keys and pedals go up and down? It’s because something is telling it, “If this occurs, then we’re going to do A, B or C.” There’s nothing smart about it. I can go on my computer and set up a program to add a contact in Salesforce over and over again. That’s helpful. Think of it like Zapier as well. If this happens, then that.
What we’re doing is two generations ahead. It’s cognitive automation. We make digital co-workers or bots that can think. If you’re familiar with NLP, there’s an emerging technology called NLGU, which is Natural Language General Understanding. It’s not just a bot that read the AI for Sales Podcast. It understands this is a sales podcast. “It’s on iTunes. Chad Burmeister runs it.” It can pull in all this data and context for you that robots can’t. It can read, think, ask questions and say, “Chad, I think this is the solution to a problem. Can you clarify for me? Now I’m going to wash my hands clean and never ask you again because I’m learning as I go.”
Technology is the facilitator to get you to a point, but it is never meant to replace that point.
I met with a guy years ago in the bay area who had acquired some code that had that capability. It was acquired from an Israeli firm that would listen to terrorists. It had to have cognitive understanding rather than surfacing, “It’s this person and here are their coordinates.” It was able to read a paragraph, not just a sentence. It’s interesting to see because now that company at the time deployed this technology. I’m like, “Let me go and see if I can play around with it over the weekend.” Similar to you, I have an MBA in Computer Information Sciences but I’m not a programmer. I thought I could go in and do that. It turned out it was a very deep amount of code base but it didn’t happen quite that easy. When you work with companies, I assume there is some work that has to occur in order to implement these cognitive bot-powered capabilities. What does that look like? If a company comes to you and says, “We want to automate these processes.” Two questions. First, what’s the value of doing that? The second is how hard or easy is it to turn something on like this?
I’ll give you what the market has and what we do because I’m a sales guy and what other options do I have here. If you want to go build out an RPA automation, it’s going to take you over $100,000 and about eight months, assuming you have the data scientists and developers. Only 4% of all RPA automation breaks even at any point. It’s typically North of two years which is ridiculous. When you have enterprise organizations with a larger center of excellence, it makes sense and that’s where the Cofounders over here came from, large enterprise environments.
What we do, because it’s intelligent, our bot learns and we’re essentially using pre-built modules and then customizing it to each customer. Instead of eight months, we’re talking about 4 to 6 weeks. The output of that is one bot does the equivalent of 4 to 8 people worth of work. When I say that, usually the thought goes to, “Roots is going and trying to automate the workforce.” No. We like people, which is why we don’t want them doing crap work. I think you’re probably on the same page there. Let’s call a spade a spade, bot. When you’re talking about bots, people get scared and be like, “You’re automating the workforce.” No. We want to give people more meaningful work.
Lori Richardson once said that at an event that I was at. She said that’s what she’s most excited about AI in sales is that it up to levels people and lets them do more meaningful work.
I have a fun stat for you. It’s depressing. Gallup came out with a study in July 2020. Eighty-five percent of the workforce is disengaged, 18% of the workforce is actively disengaged. Meaning they’re actively working against you and what you’re trying to employ them to do, which is mind-boggling. Everyone wants to look at their team and say, “That’s not us. It’s the marketplace staff.” It’s the number one reason most people don’t have meaningful or purposeful work. We take the stance that you have an obligation not just to provide gainful employment but you have the moral and ethical obligation to bring the best out in your people. You can’t do that if they’re doing crap in data entry. If they’re doing mundane tasks that don’t require any skill, that doesn’t challenge them or move them forward. They’re never going to move up. Up-skilling and re-skilling is this nice, fancy idea thrown around at conferences. That’s all it is. That’s the problem we’re trained to solve.
We have a prospective customer that was looking at a data application that requires one-to-one handpicking through the data. The reps on the team were like, “This is awesome because then all I have to do is go look at data all day and I don’t have to sell on the telephone because that could be uncomfortable.” We came in and said, “If you go buy this data tool for $50,000 for the year, you’re going to be doing your reps more of a disservice to them and your business than if you were to bring in the dialing application that 10X is the number of sales calls your reps can have every day.” Sometimes it’s not obvious with change management but when you look under the hood and say, “What’s the ROI of such an investment?” You have to focus on what are the outcomes you’re after when you’re putting in technology like this.
There are two issues with AI in sales and marketing. One, everything is AI these days, whether it is or isn’t. Anything with an if-then statement anywhere in the program is now artificial intelligence. Number two is we’re using AI to make people dumber because we don’t trust our people. That’s the exact wrong approach. If you’re talking about small ticket SaaS sales, you might be able to get away with having a workforce that does the same thing repeated over and over. The same exact process, wording, objection handling. You might be able to get away with it. What we do is enterprise sales. You cannot make your team dumb that is taught how to avoid human interaction.
What you need to do is keep them on the phone. I don’t mean pounding through hundreds and hundreds of calls a day. I’m here in Detroit, so we have companies like Quicken Loans where they’re making 600 to 800 calls a day. It’s nuts. If you go to the HomeAdvisor, they’re making 200 to 300 calls. What we want to do is teach people how to have meaningful interactions and capitalize on every conversation that they have, leaving an impact on every individual prospect, which is not done through technology. Technology is the facilitator to get you to that point but it is never meant to replace that point.
Give me an example of a company. You don’t have to necessarily name their name. Tell me a customer who’s deployed the technology. How did that work in their environment?
I’ll give you a good example. There is a very large lender that we signed up in 2020. They specialize specifically in personal loans, not commercial. They are supposed to be a mostly tech organization. They pride themselves on being a customer-centric experience. The funny part was when they had an application for a loan, they had to check Salesforce manually to see if there was a duplicate. They had to run OFAC checks, Office of Foreign Asset Control to see if this person was on any lists. They had to pull a different source and see if this person on any terrorist watch place, then they had to get the bank statements and try to parse them into the system.
Every bank has a different format for their bank statements, so that’s difficult to do. They’re trying to be a tech focus organization and give this great tech experience. What we ended up doing was automating all that data entry for them. What would take hours to approve a loan application, we get to do in a matter of 2 or 3 minutes. No problem. That’s a perfect example of, “Let’s go automate all the mundane stuff, not automate the human experience and interaction.” Most people have that flipped.
I talked to a few guys who left one of those organizations and went to a competitive organization. It’s funny how small the world it is. It very much could be the same group of people. That’s the secret sauce. Their unique, competitive advantage in the market is the ability to make those decisions within seconds or minutes.
I have a fun one for you. We have a very large insurance broker, focused mostly on the commercial side in Ohio. That’s as far as I’ll go here. To get one quote makes them 24 hours. I don’t mean a day. I mean, 24-man hours to put in front of their customer and present it. Here’s what happens. They accept a submission from an agent, they send it to 4, 5 or 6 different carriers depending on what the policy is for. Every carrier has its own SLA. You have no idea when you’re going to get the quotes back. It takes them eight hours to take these giant quotes that the carriers gave them and put it in a presentable format for the customer. They show the customer, sit down, walkthrough, choose, “I want this quote.” Then they take this big XL back and transpose it onto a Word document, so it’s more of a full proposal. The accuracy is only 85% because it takes them an entire day of data entry, looking at Word and Excel and trying to compare the two, so they have yet another eight hours to review it.
It’s three business days to get one proposal out. These are licensed agents that should be going and selling it. Don’t get me wrong. We do more than financial services and insurance. That’s where we focus but illustrate the point across the board. People aren’t trying to automate prospecting and there’s the right way to do that. I’m not against it. There are right ways to automate lead scoring, lead nurturing, etc., using AI but it’s not meant to replace people. That’s what we’re aiming for.
I remember back to FedEx versus UPS days. At the time, FedEx had it down to where it could be a day, especially in a more complex deal UPS was more ground transportation when it comes to delivering the quote. It would take them two weeks. I come from moving fast on a speed boat concept. I’m like, “A day? Are you kidding me?” Compared to the relative another choice of two weeks, it was like, “That’s super-fast.” The same thing goes in nowadays. You can go in and you can buy something online. All that algorithm is expected to be running in the backend. It’s going to be interesting to see where things had over the course of the coming days, weeks, months, and years.
In large part, the focus needs to shift from, “Let’s automate everything for the customer,” which is good in the right applications. I’m not anti-doing that but we need to focus on automating things for our own employees and teams. Happy employees provide a happy customer but not only that they can service more customers. They can sell better. There’s more joy in what they do. They are more profitable for you. It’s funny we talked to a lot of banks, financial services, healthcare and insurance is where we prioritize for the most part here at Roots. Banking is interesting in particular because you have banking apps.
I use VIABUY. We have Zelle. You have all these different assistance that they’re providing, but then you look at the back office and they’re still doing everything on paper. That’s not specific to be VIABUY. That’s a blanket statement overgeneralized. Banks are still doing their reconciliation on paper, everything in Excel. Their loan officers are doing all the submissions by hand, underwriters have to go ask the loan officer to get something from the loan applicant and then it’s a game of telephone back and forth. Automate the back office. It may not be sexy but it is what gets you the customers in the end.
Where do you think the industry is headed? You’re two levels ahead of the competition. Will they catch up and then what’s the level beyond that you’re working on?
We want to automate all the mundane stuff, not automate the human experience and the human interaction.
Everybody’s going to catch up. Sales are going to take a nosedive. Let me explain. Things like GPT-3 are coming out to automate the way that you write and speak. You have programs that do nothing other than, “Let’s go with the shotgun approach and see what we can hit by sending out as many emails, proposals and all these different things as possible.” Speed is needed. Forget about Millennials. They are 40 or 45 years old sometimes. Talk about Gen Z coming up and their 20 to 25. Speed is necessary. I’m for it but they also need hyper-personalization more than a merge field flipping out their first name and company.
We’re going to see a nosedive where people who don’t understand this are going to hurt, fall behind, realize that this spray and pray model only gets you so far then. As we tailor the AI and all these methodologies to a better customer experience not a faster customer experience, then we’re going to see a large increase in the success that salespeople have but the reality is, conversion rates go down year over year. The average in almost any industry is roughly 20% of qualified deals close but they give or take a couple of percentage points obviously. That’s sad. Eighty percent of your time is a complete and utter waste. I rather focus on how do we capitalize on, “What we have? How do we convert at higher rates?” Not just, “Let’s go a spray and pray.”
“Don’t spam your Tam,” has been a phrase out there for a long time. Some of the big companies have been able to get through using that approach and yet dozens of companies are coming out of the woodwork saying, “Look at our metrics. Can you tell us where we’re going wrong?” They show me their emails sent per day, social touches and dials. They say, “Read the team leads. What’s going on? Why do we have only 2 people in green, 3 in yellow and 18 in red?”
There’s a couple of tools now that have sentiment analysis in it. I have a 65% open rate, 20% reply rate but 19% of it is negative. They’re telling you to screw off. We look at superficial metrics and think we’re doing well but then we wonder why the revenue isn’t coming in. Sentiment analysis is one of the greatest applications for AI, so you can start to understand not even just what works but what works for this persona versus this persona? Why? What is the difference? Start to be scientific about it as humanly possible?
One of our customers or partners, FlowEngine out of Austin, Texas. They’ve got a co-op built of dozens of companies who say, “We’ll give you our exhaust file of all the calls we make, all the emails we send, all the social.” They’ll tell you, “Which phone numbers in the last seven days picked up their phone?” If you can say, “These are the 20% that are the picker-uppers, these are the 20% that will never pick up, then there’s probably some in the middle that is maybe picker-uppers.” You could start to segment a list and say, “I can’t afford to call the never picker-uppers and maybe I’ll only call the middle ones once or twice but the 20% that pick up all the time, as long as they’re in my decision buying authority, then let’s call those people.” It’s the same thing if you’re going to do an email. Why would you go to ones who never reply to an email? Is the data out there? How do you aggregate it all and present it to the rep at the time they need it?
What you said spurred a thought as well and it’s not even specific to AI. The original sales approach that we had in any industry was, “Let’s go take twenty accounts, stalk or call everybody in that organization and try to bust in there.” Now we have a, “Let’s go spray as many people as possible. Let’s a good target every CEO or CMO under the sun and hope that it works.” What I teach my team to do is persona-based selling. I can’t target the CEO of a bank the same way that I’m going to target the CEO of an insurance company or the CMO at that same bank. It’s not even just what messaging works. It’s what works for exactly what persona? How many degrees different is this persona from this persona? It has to become very scientific. AI is certainly helping in that regard or some apply it the right way. What you said about the phone calls is phenomenal. What’s not phenomenal is there are so many tools to say, “I’ll go get you 100 prospects like this every single day and hit them with the same email every single day.” That doesn’t work.
Just imagine if you had the list that said, “These are the 20% that pickup.” Now let’s take it a step further. “These are the 20% that pickup and/or a driver or certain personality types.” If you look at your customer base and say, “Eighty percent of my customer base are this type.” It’s like they pick up the phone and they’re this type. That’s where the power of big data is going to completely change this industry where its hand-to-glove fit.
If you were to combine that with intent data that was accurate, game over. We use some intent data at Roots with a grain of salt. We’ve used several tools, the accuracy is a bit iffy, especially depending on which organization you use. If you add quality intent data which will only get better with big data, which is coming out every single day and you combine that with something that says, “These people pick up. These people don’t.” Game over. You win as long as you’ve invested enough in your team to have a legitimate conversation. I have a fun story. We’re hiring four additional salespeople on our team. I had an interview. The guy has a degree in sales that exists at all four colleges. I get on the phone. After I say hello, I got in exactly 6 words within 11 minutes. I count it. “I don’t care how good your tech is. Your salesperson still sucks. It’s safe to say he is not getting the job.” That’s what we’re combating because we’re making our salespeople dumb by spoon-feeding them everything and not challenging them on their skills.
The latest AI that I’m seeing in that space is there’s conversation intelligence but then there are conversation guides that say, “Here’s what you should say when you hear them use this word or phrase.” Then it pops up, then as the rep reads it on their screen to the prospect, it checks them off. Some people will be able to handle that. When my manager used to write on a whiteboard for me, I was like, “No. I can’t do this. Let’s talk afterward.”
I think it depends on what you’re selling. What we’re talking about is essentially an interactive script. Maybe you follow it word for word, maybe you don’t. I think that works well with small to mid-level types of sales, maybe it’s $10,000, maybe it’s $20,000 ARR.
You’re on the phone 30 times a day with someone inbound may be, and you’re accepting an order for a $50 a month product or something.
Where you’re not going to find success with that is anything enterprise. If you look at the startup landscape, people started out by selling $50 widgets, $100 widgets primarily in sales and marketing because salespeople and marketers are the easiest people to sell and market to. They want to know what’s going on. They’re always looking for the next best way to get in front of somebody. There are so many startups right now that are selling into enterprise organizations. In Roots Automation, we have a seed round. We will have a series A. We’re selling into Fortune 100 and 500 companies. You can’t have the skillset of a startup to do that. You have to have the skillset of an enterprise sales team, and you’re not going to get that from a script, whether interactive or not. You need to understand emotions. You need to understand what is this person cares about and have the cajones to ask the hard questions that will grab their interest.
The politics inside an org in trying to find the right person. The AI’s not going to figure all that out for you anytime in the near term.
When it does, we’re all screwed, then we are all out of a job.
We’ve talked a lot about how your technology enables other sales teams to be effective, where is your team and you personally, what tools and technologies are you seeing that have AI built into them that are piquing your interest?
There’s a couple of things to that. I’m a big proponent of Outreach which is adding some AI capabilities slowly. They have sentiment analysis. A lot of that is automation but there’s some AI being interjected there. I believe you had the CEO of SalesIntel on here.
We signed up with them and it seems to be very accurate data and extremely helpful. There are two things that we’re testing internally that I haven’t seen on the market. I’d love to check something out if there is something on the market. One is, there are plenty of tools. Outreach is one where it will go their LinkedIn profile, Crunchbase link and some things like that but nothing that says, “This is what you were looking for on the LinkedIn profile. Here’s everything that you need. This is the field that you wanted on Crunchbase. Now, here’s the link and go do your own research.” Starting to pinpoint what matters to you the most and being able to sort through that much quicker. You have those relational talk points but you’re not doing the work to go find it. Whereas nowadays, people are satisfied with having a link here, a link there. “I don’t want to click around. I want everything in front of me about this person, so I have talking points.”
There’s a company in Dallas and there’s a reseller of the same company that’s in Ohio, Cleveland area. When we met with their team, they said if they’ve claimed their Google My Business page and there are about 4 or 5 other things, how many star ratings do they have as an individual and as a business? I was like, “How long does it take you to do one of those lookups?” They’re like, “It’s only about 3 or 4 minutes.” The 3 to 4 minutes means every single one individual phone call they make, they have to go research it and they don’t even put it in a data set. It’s just there and gone. I was like, “You mean, we could pay an $8 an hour person to go pull all that or automate it better yet so that when the screen pops and you’re talking to someone, all the information’s there?” They were like, “You could do that?”
If people are doing mundane tasks that don’t require any skill, challenge them, or move them forward, they’re never going to move up.
If you’re selling into small to mid-market, there is a tool called Solidify. It’s been a long time since I’ve used it just to give you the nature of how we sell. It would do things like, “I want people with older versions of WordPress and it indicates that they have a crappy website. I want organizations that don’t have mobile responsive sites.” Same thing for SEO characteristics. That’s pretty cool.
It says, “Build target list for your personalized outreach powered, by data you can’t find anywhere else.” That way, you’re going in with something that’s highly unique and differentiated.
I’ve never seen a platform where you can say, “Show me all the crappy websites out there. Now I’m going to go sell them on marketing services.” That’s a great application for AI. The other one that we’re working on and going to build internally, it’s taken thousands and thousands of points of outreach from different personas, different industries. “Can we lay AI on top to help us narrow down this messaging works with this persona at this size company in this industry? This messaging works with this persona at that size company in this industry?” That’s what we’re trying to narrow down but to get there, you need so many opportunities. One, you need so many points of outreach. You need an incredible amount of data to get there. I’ve seen Salesforce Einstein do a little bit of this but it’s focused on the life cycle or the opportunity life cycle management. Nobody’s focused on, “What makes the best messaging for this person at this time in this company, in this size company.” That’s what we want to solve.
Years in the business as an organization, you’re running sales and marketing. Thank you so much for sharing with our readers. If people want to get ahold of you, how can they do that?
Thank you everybody for joining and we’ll see you next time on the show. Thank you, Stephen.
- Roots Automation
- LinkedIn - Stephen M. Lowisz
About Stephen Lowisz
My expertise in sales consulting and optimization comes from the experience I have working with companies like ESPN and Intuit by the age of 17, the first $1.1M I generated at 19, and the dozens of startups I've helped scale across 5 continents.
I now spend my time as the Vice President of Sales for Roots Automation, a hyper-growth technology startup bringing Cognitive Process Automation to the marketplace at scale. Our technology is able to get companies lean and mean through increasing efficiency by 400-800%.
As the Head of Sales, I'm personally responsible for client acquisition, building and training our sales organization from scratch, and expanding current client relationships. With our recent Seed round of funding, Roots Automation is growing and scaling at exponential rates with no signs of slowing down.