Tricky Balancing Act: Bots are Efficient, Humans are Essential

Chatbots are a powerful tool for assisting small and medium-sized businesses (SMBs) and banks with customer interactions on their platforms. They offer significant cost-effectiveness by reducing overhead, and provide scalability by handling multiple conversations simultaneously.

Additionally, chatbots can gather and analyze customer data to help organizations improve their services and tailor their offerings. There are important limitations to what chatbots can do. They have a limited understanding of nuanced or complex queries and often lack the emotional intelligence necessary to build trust and connect with customers who are seeking more personalized support. Our panelists will discuss how to strike the right balance between automation and human interaction to deliver a seamless and satisfying customer experience for the small businesses they serve.

Transcription:
Transcripts are generated using a combination of speech recognition software and human transcribers, and may contain errors. Please check the corresponding audio for the authoritative record.

Nick Miller (00:08):
Hello? There we go. Welcome back. Just checking the room for echo. It's pretty good. So hello, welcome back. I'm Nick Miller and I'm joined on stage by Ashleigh Ashbrook and Chip Higgins to talk about the bot-human interface, how to strike the right balance between automation and human interaction. And the question is, well, why the three of us? Well, the answer to that is each of the three of us has a connection to the conversation. My firm Clarity trains bankers to have the expertise and the knowledge they need to add value at the point of conversation. So that's the completely human side of it. Ashleigh at Michigan State University Federal Credit Union has developed a bot that is very well received by their members and staff. She'll talk more about that. And Chip is a former banker who started a company and built a large language platform that gives advice to business owners.

(01:06):
So we've got between the two of them a stair-step function in terms of what's available or possible in terms of the human-automation interface. Ashleigh and her team developed and launched Fran, an automated assistant with a designed personality which clients and staff have responded to very favorably. She directs a team of about a hundred people who work on the electronic side, the digital member experience. She's been recognized for this work. Earlier this year, she joined the faculty of the Bank Automation Summit speaking on what's new in digital assistants. Chip reinvented himself after a 35-year career in small business banking to write a book, Bisix, which applies the laws of physics to growing successful small businesses, and started a company to address the frustrations of small business owners in terms of accessing reliable information, new ideas, and a thought partner. He's created a large language model based on all of that, which you'll see a little bit of later.

(02:07):
So Ashleigh, let's start with you. You've developed two bots, one retail and one small business. Let's start with the retail one. It won the most likable AI agent in this year's Boost AI conference, so it's a very strong virtual assistant. What was your goal, thinking about the human-automation interface, in creating Fran?

Ashleigh Ashbrook (02:29):
Well, thank you so much for having me here, Nick. I'm really excited to be able to talk about how we really have created this path, walking with human and digital together to create a powerhouse combination where we can gain some efficiencies, look at right-sizing our staffing, and also allowing for additional service components for all of our members on the consumer side and the small business side. We're about a little over five years in with developing our virtual agents. We're trying to move away from the word "chatbot" because bots can have a negative connotation, especially if you have any bot attacks coming or info security people here—we hear that word a lot. So we're trying to make sure that we position this as something that is really an agent. They are an extension of our human team, and that was exactly what we were trying to accomplish when we started to build out Fran.

(03:22):
She has a very extensive backstory. She has a connection to our roots that go back to when we were founded in 1937. She's named after our first manager, which would've now been a CEO, Francis Lesnieski. And we told her story quite a bit to all of our employees and all of our members. They were in full awareness that she was a virtual agent, she was here to help, and that she had a connection and tie to being an extension of who we are. You could get the same level of service from Fran, our virtual agent, as you could from a human agent if you walked into a branch and were asking questions, contacted the call center, or if you contacted us on live chat.

Nick Miller (04:00):
So how did you think about the interface between the automation and people? When you were thinking about the design of that, what were your guiding principles?

Ashleigh Ashbrook (04:08):
Well, we really wanted... Okay, I think we're back to normal. I'm going to try to speak a little bit lower. Hopefully that wasn't me. We were keeping in mind all of our core values that we have as an organization that we expect all of our employees to follow. We wanted Fran to do the same thing. We wanted her to create a positive environment when she was interacting with our members. We wanted her to be able to embrace and drive change. And our members really did embrace this very well. They took on to Fran, using her and asking questions, and we wanted to make sure that they could get service twenty-four seven. So Fran is available at all times. She does not take a day off. She doesn't have PTO. We do not pay her overtime. We really love her as an employee.

(04:54):
HR really loves her and she's there to answer any questions. The benefit from it, too, is that any question someone asks Fran that she doesn't fully understand or can't execute on—if it's within business hours, she'll transfer them to a human agent. In the backend, my team managing her knowledge base and building out all of the intents or questions that we're putting in her knowledge base is looking at that on a daily basis. So we're going through and making sure that we're putting in more information if there are questions that she doesn't know how to answer. Every time we have a member come back, they're getting more trust and more loyalty built in because she can continue to answer one more thing or expand on one more answer or be knowledgeable in one more scenario that they didn't have before.

Nick Miller (05:39):
So talk a bit about the personality side of it. There's a backstory, but a bot's a bot, right? So how do you create a personality?

Ashleigh Ashbrook (05:48):
Yeah. I was going to say your virtual agents can be anything that you want them to be. It really depends on what your brand is, what your voice is, and what you're looking to get across to your clients or members. For Fran, we wanted her to be professional, friendly, and thoughtful, and we wanted her to be responsive. So she very much fits in with our brand voice. We do tend to stay a little bit on the conservative side. If you're looking at other companies outside of the financial services industry, they will maybe have some more cheeky language or a little bit more gimmicky back-and-forth. We wanted to make sure that everyone coming to us knew that this was very professional. They're entrusting us with their money, with their finances, with everything they have. So we wanted them to first understand that she's here to help serve you, but she also can tell a joke.

(06:39):
She knows some things. We are connected to Michigan State University. For anybody familiar with that, our colors are green and white. So if someone were to say "go green," she will immediately respond with "go white," which would be the appropriate response if you are a Spartan fan. Some of those things we've built in because we know those are things that our members who visit us in the branches and talk to us regularly will have an interaction with our staff as well.

Nick Miller (07:04):
So we have a couple of pictures of Fran. Can you talk us through what we're seeing?

Ashleigh Ashbrook (07:09):
Yeah. This here on the screen is our pillar website. We are moving all of our business account members to a new small business banking platform called Pillar, powered by Nimbus. So they're a partner with that. We also partner with Boost AI to power our virtual agent. Right here, you're going to see the home screen for our website. And at the bottom right, you'll see the icon where you can chat with Fran. That'll be there all the time. You can click on her whenever you want to ask her anything and she'll pop up.

(07:41):
And then if we move... the next one, she's going to have a couple of standard responses that she comes up with. "Hi, I'm Fran." She's introducing herself. She's going to tell you right away that she's a virtual assistant and what she's an expert with. This bot model right now we are building out specifically for our small business members. On the retail consumer side, she's built out with several thousand intents right now and handles about 75% of the conversations that start with her in a live chat, and only about 25% go to a human agent. Since we don't have all of our members transferred over to Pillar right now, we're starting in the same small capacity we did when we launched Fran in our retail environment. She's got about a thousand intents or questions built in that we started with.

(08:27):
The three hot links in dark green that you can see for Join Pillar, loans, and apply for a loan are some of the top asked intents or questions we get on a regular basis. Before you even have to ask anything, you can click one of those and expand it, or you can ask her any specific question you would like to know more about. Right now you're interested in business checking accounts. She's going to give you a link for the checking account. When you click that, it'll expand down to all of the checking account options we would offer any of our members. You can click on each one and it will give you a description. As you want more detail, you can continue to click in and build out more information. The key to that is that you want to be able to provide small chunks of information and allow your clients to continue to click through it.

(09:12):
They're typically looking at this on their mobile phone, so their view is much smaller. If you're giving them a huge paragraph of information, they can't read it. So these are really manageable chunks to read through, and then they can determine, "Do I really want to look at rates?" or "Do I want to go to how I open an account?" because they already know they want a small business checking, and they get to choose their own adventure on what's next.

Nick Miller (09:34):
So we're here talking about the agent-human interface. Talk a little bit about how the members have responded to this, how they feel about it, and then your staff.

Ashleigh Ashbrook (09:47):
Yeah. I would say that the members were incredibly trusting when we launched Fran. When we first started with Fran, we actually gave our members an option on our public website if they wanted to chat directly with her or if they wanted to start with a live human agent. That way we could start slowly into the waters of introducing them to the fact that we had a virtual agent, that we were going to be using this technology, that they could trust it, that it was secure, and it would provide responses that were accurate and detailed just as if they were talking to a human. Fast forward, we moved to put Fran in front of every chat interaction. So now before you even get to an agent, similar to what you would do in a phone tree, you would have an interaction with Fran and ask her a question.

(10:30):
You can bypass that and get directly to an agent. That was a core value we had where we know our members want to talk to a human at times because they need to. We are not going to make that a difficult journey. You can choose to do that, but I don't recommend it because it's not going to serve you well in the long run; they don't want to have too many steps or clicks to get to that human when they need it. And then our employees, as you can probably imagine, were a bit nervous at the forefront thinking, "AI's taking over. I'm not going to have a job anymore. What am I actually going to be doing?" In turn, it is actually the opposite. We've seen longer average handle times for all of our human chats and more complex questions and interactions come through.

(11:13):
Fran essentially is taking a lot of the basic, simplistic tasks. Since we have launched Fran, her number one intent has been, "What is my routing number?" The routing number is plastered in every single corner you can imagine on our website and in our app. It is everywhere, but our members aren't going to memorize that. I have it permanently imprinted in my brain because I've been with the credit union for 18 years, but I don't expect them to. So we know she's taking some of those really basic questions that you don't want to pay a senior employee to do, which is reset a password, answer the routing number, or send somebody a form. Now that we've got that baseline foundational setup where the employees know the questions they're getting are really thoughtful and complex, it's challenging to them in their job.

(11:57):
They're actually really excited that we have Fran and Fran is helping with so many of those conversations. We also are able to utilize her after normal business hours and on the weekends where we aren't staffing actual human agents because it is a bit of a challenge for us to find someone who wants to work within the confines of those hours. Fran can continue to do that. And then she does some basic transactions now behind a secure login and we're working on some LLM features. She's continuing to grow.

Nick Miller (12:24):
So it's me. Jimmy Hendrix here. How are you continuing to manage... You talked a little bit about this, but maybe say a little bit more about how it changed your staffing model and then how you're managing the evolution of this going forward.

Ashleigh Ashbrook (12:42):
Well, when we first launched Fran, we didn't make any significant changes to our staffing model. We wanted to see what the outcome really would be—how many members would react to talking to Fran versus talking to an agent. Over time, we've been able to create several efficiencies. We have one specific dedicated digital member service team that I used to oversee, which was called eServices. We had not hired a full-time employee for probably the last two and a half years because we were right-sizing the staffing. We knew the amount of productivity we could have from an employee, and we knew the amount of productivity we could have from Fran. So it allowed us to take some of those resources, pour them into Fran, pour them into training and development for the existing employees, and give them some additional growth and career opportunities as well.

(13:29):
It saved quite a bit on the recruiting and hiring front, and we've really gotten to develop some time to work with the individuals already on the team that know everything that's going on.

Nick Miller (13:40):
Okay. So on an ongoing basis, then, as your members' tastes change and as you gain experience, what's your vision of how you continue to evolve this interface that you've set up?

Ashleigh Ashbrook (13:51):
Well, I would say total world domination, but that's probably a little too excessive.

Nick Miller (13:56):
Good state.

Ashleigh Ashbrook (13:57):
Yes, right. I think we can all think about, do we want AI to completely take over? And the real answer is no. As much as I love AI, I love what I do and all the technology advancements we've had. There isn't really a time where I can see that you won't need humans. The humans are actually what's building the knowledge base behind Fran. You need employees who want to be connected to your organization, understand how the members are thinking, understand what's happening in the world, and can actually train her on the things you want her to be trained on. We have to have that human component to make sure we're giving out the right information at the right time and the right level of service. So it's really a combination of us walking hand-in-hand.

(14:40):
The employees understand that now because they've seen it. They've seen firsthand that we've grown our AI department from employees who worked on the front lines. They were employees helping members every single day with their questions, and now they're building out the models that will help serve them virtually. We want to continue to grow that and continue to create efficiencies in staffing. I'm sure everyone here understands that it takes a lot and we might not have all the qualified candidates we're looking for, and there is some more flexibility in using some of these AI components, but you want to use it in the right way and do what's going to work for your client or member base.

Nick Miller (15:18):
Great. Thank you. We're going to shift a little bit now. The conversation's going to turn from an agent that is limited in terms of scope—not conversational, not particularly emotionally intelligent—to a platform that is those things. Chip Higgins was a banker for 35 years, primarily in a small business space, became very frustrated with the difficulty of small business owners gaining access to information and thought partners, wrote a book, Bisix, that integrates the principles of physics around how you grow a business, and then developed a large language model, which he calls Bisix on Demand, that really is a conversation partner. My first exposure to it was when I asked it about hiring two salespeople, told it that I wasn't having much success with that, and spent an hour in conversation with the model getting as good advice as I would have expected to get from any sales consultant, including tools, templates, and a simulation structure to use in hiring.

(16:24):
Since Google Duplex was announced in 2018, I've had the view that it would be possible for an advice platform that banks could use, since that's the industry I work in. Chip has really taken a strong step forward in terms of creating that. I thought it would be useful for you to see it and get a little bit of sense of what's possible and where this might be headed. As Ashleigh and her team are developing Fran, Chip and others are developing language models that can really give advice across a much broader set of issues than the products and services a credit union or bank might offer. That was long, but it was good. I liked it. So what led you to write the book? Give us the one minute on the book and then we'll get into the model.

Chip Higgins (17:13):
Yeah. I mean, after 35 years of dealing with business owners—I can't tell that whole story—but the part I'm comfortable with in a minute is when I was at Pinnacle Bank and started the Mastermind program. We developed a Mastermind program for the E-Myth Revisited, which a lot of you are very familiar with. It was a way to engage branch people in a meaningful conversation around business. As I attended those groups, what I realized is it's just a very confusing world and people needed some order in their thinking about what they were doing. The book is physics-based, but it's really the momentum formula of physics, and it's all about building and sustaining momentum in your business. What stands out about that, if you don't know physics momentum, is that for the theme of simplicity, it's probably the simplest equation you can find in physics: mass times velocity.

(18:09):
What I really tried to do is develop an envelope relevant to momentum and apply that physics principle to business. It's been very well received. I've got a number of number one Amazon bestseller rankings with the book and it's very easy to read. That was my intent: make it easy to understand and simple to access.

Nick Miller (18:31):
Talk a bit about... we're going to show a couple of slides of the conversation in a moment, but talk a little bit about how you thought through the interface you wanted. This panel is about the human-automation interface. Talk a little bit about how you thought about that.

Chip Higgins (18:46):
Yeah. I think in all my work in small business, it always came down to scalability and accessibility, but still keeping high quality. I think that's been the main theme. I thought the book presented all of that, but I've been paying attention to AI and coaching because I'm a coach, and I saw an article from the International Coaching Federation. It basically said all their testing of AI coaching models revealed that the AI model is equally competent to a coach, other than with extreme emotion. I was like, well, if coaching normally costs $1,500 a month and most small businesses want it but they can't afford it, is there a way to get my book in a coaching format? Regarding the coaching philosophy—and I enjoyed you talking about training an agent because the possibilities are endless—my philosophy was problem identification, a good conversation to define what the problem is, understand where they've been on that problem, understand the set of options they're considering, and possibly introduce more options if they're stuck. And then most importantly, find a way forward.

(19:48):
I think that's where a lot of small business owners get stuck: they want to do something, but they can't figure out how. So that's the mindset of the agent. The other part regarding the personality is that my book has been ingested into the agent. My book is my life's work. It's me. It's everything I think about small business. So it has my voice already. When people interact with the agent, that's what they find. They're basically talking to me.

(20:36):
I think you probably found that, Nick, in your interaction. It's a very hopeful, resourceful kind of tone that it brings, with access to a lot of things. It's a closed system in that it has my book, all the articles I've written, all the blogs I've written. It's closed in that respect. It's not being shared with the LLM world, but it can exit that world and go to the LLM to pull in templates and other information very specific to an industry. It's fascinating to watch it work. It's beyond my wildest dreams what it can do. I've been really blown away by how it operates in the field and the outcome of it.

Nick Miller (21:17):
What sorts of conversations do you see people having with it?

Chip Higgins (21:23):
It's a broad range. I've had people come in just for an elevator speech. They start talking about what they do and they can't say it fast enough. You'll watch the agent formulate the words according to the Bisix way—it talks a lot about energy—so it was energetic and purposeful about what you're trying to accomplish. It will do that. I've had a solopreneur with a side hustle doing an art business where it helped them on some marketing things. So it's very simple on that end, but I've also had some amazing... I mean, I've had one leadership interaction. There's a chapter in the book on inertia that's all about leadership. Someone came in and said, "Basically, I read this chapter. I don't know where I am on leadership." It immediately asked them, "Well, where do you think you are?"

(22:09):
And they didn't know. So it gave them 10 John Maxwell influencing principles and said, "How many of these do you think you have?" They said, "Maybe two or three." I was like, "Well, what do you want to do in your company?" "I'd like to build a leadership culture." To summarize what happened, they designed a book club around "On Fire at Work," which is a great culture book. The agent gave an introductory speech for the kickoff, had emails introducing the idea, and basically equipped them for the entire thing where they were ready to go. Probably the most extensive one I've seen, which really blew my mind, was someone who was struggling with an employee. The whole conversation was about SMART goals. They were trying to figure out, "How can I hold this person accountable?"

(22:58):
That's the friction chapter in the book about getting traction. From SMART down, it was a parsing of how would you define this? Is it really specific? Is it measurable? It was probably... if yours was an hour, Nick, I bet it was two sessions of an hour, but now they have it. They have a SMART goal, a template for the goals, and a template for tracking. I talked to this person last week, and I think they're coming in for the performance plan now. I'll be interested to see how they navigate through it—it doesn't sound like the SMART goals are being adhered to. That's the range of it. There are many chapters in the book. It's about direction, mass, even financial chapters. Friction and leverage are really about liquidity, profitability, and your debt stack. It'll engage in all of that.

Nick Miller (23:53):
Yep. Let's take a look at one of the conversations here. An opening and then...

Chip Higgins (24:02):
If you could go back to that, I want to talk about the whole experience on that first slide. Intention is a big thing in my book. What is the purpose and what are you trying to accomplish for any of your customers? You've got to have intent to it. I'm making a statement here that we're trying to make them extremely competitive, achieve the fullest vision of their business, and achieve longevity—the 50% failure rate is still happening 40 years later in my banking career. So we're making a statement. It's very friendly and open. I'll go to the next slide. This is about energy. Any coach wants to know what's most important to you and what's on your mind today. This is a truncated slide because they do have the opportunity to go to a physics assessment and figure out where they are in the whole business way formula.

(24:50):
But basically, this encouragement: I'm ready to go, I'm here twenty-four seven, unlimited access, and let's do this. In this one, they describe their position: a small sign shop dealing with late AR. They think they've defined the problem here. So it reinforces a business principle of being consistent and systematic with overdue accounts. I think that's a fair thing to do. We start talking about messaging. You can do a polite reminder or persistent follow-ups, giving you a general tone of what you ought to accomplish. Go to the next one. Dealing with late payers and keeping the relationship alive even when people are cash-strapped. Here we go to the coaching thing. What are you doing now? This is a question: how are you handling this now? Would you like a reminder sequence or some type of change in terms for people paying you late?

(25:48):
The person replies here. They talk about increasing prices for late payers; they're making calls when they pass their due date, but they don't have time. I thank them for informing me on that and affirm what they're doing. We go to the next one and start talking about leaning on technology. We talk about how accounting software will do this for them if they'll get the message right on the front end, and upfront payments, deposits, and that type of thing. We move on to something they can do. What's the next step for you? Would automated reminders or a deposit requirement be a practical next step for your business given your workflow?

Nick Miller (26:33):
Truncated very quickly.

Chip Higgins (26:33):
Truncated. I know I've got limited time. I'm going to fill in the blanks about what happens from this point on with the agent. When they reply to that, the agent goes into resourcing mode. In this situation, the agent got into whether they were using QuickBooks or another platform, was prepared to offer other platforms, and even got into designing the message according to what that business owner wanted. It was a very successful interaction from my standpoint. They came in and got resourced on that topic. What was interesting about this one is it took a 90-degree turn. They're still in the financial world, but we started getting questions about buying a piece of equipment and how to do that. The agent basically prepared them to go talk to the bank, talked about the equipment and what it was going to do for their business, how they could talk to the bank about what they're going to accomplish with that, and gave them a checklist of how to prepare for the bank with that financial information and everything else.

(27:28):
It was a very complete interaction. That's what I find. I'm really pleased it's fulfilling that promise of going from the problem all the way through to a solution.

Nick Miller (27:41):
One of the things I saw in Google Duplex and am now seeing with Chip's is that if you think about the advice given in this example and others, the question it raises for me is: how does this fit with a business banking model? If we want our bankers to be trusted advisors and we've got a consistency issue person-to-person across the geography, could a model like this—not Chip's model necessarily, but a model trained to give advice across a range of issues—be a tool to support bankers? Like Fran, could it be available twenty-four seven to answer questions like, "How do I hire salespeople?" or "How do I go to a bank and apply for equipment?" When Google Duplex came out in 2018, I thought that was going to be the end of the world as we knew it, and it hasn't evolved quite that way.

(28:34):
It's been a little slower to change. But Chip, since you've managed small business banking sales forces for years, I'm just curious, how do you see it? Generalizing from the kind of model you've developed, how do you see this playing out in the banking industry?

Chip Higgins (28:52):
I struggle with all the issues I heard this morning at three different banks, and they're the same issues. I think we have a channel that's very crowded and very difficult to populate with people who are knowledgeable and consistent in their messaging. The first thing I believe is this is an absolutely essential thing for a bank to solve the problem. You have expertise and you can pick whatever... my whole thing is about momentum and business success. There may be another application of this, but what you're building is an expert in the field available twenty-four seven with unlimited access that's completely human. When I worked with gearing ratios and I thought about, well, we've got these branches in Nashville, but we've got that one down in Hohenwald. What are we going to do in Hohenwald?

(29:41):
How are they going to be served? I know we have phone centers, but it's a very different level of knowledge that would be extremely difficult to train people in. So I think it's going to be an essential thing moving forward. I think it's going to be a very strong brand differentiator—the way you build it and the way you train it. My agent can be trained however. My agent doesn't get into financial analysis; it will stop you if you try to send it financial information. It doesn't get into legal things—don't want to deal with legalities—but will deal with anything in the book. I think it's really a critical step toward unclogging the channel a little bit in the retail world. And as Ashleigh said, elevating the skill set in other areas for the conversations we had.

(30:28):
I've used Glenn's tools at banks and we all seem to have a lot of tools, but how do people access them? How do they use them? What I would personally love to see, as someone who's left the banking industry now, is that that becomes a job of people helping navigate conversations with businesses about business planning. I think it's very affirming for people to be in a place where they understand it, know how to execute on it, and they become the driver of it. And I haven't lost humanity with mine. There's a weekly email and daily office hours. If they want to contact me, they can. It's not like ChatGPT where you never know. But at the same time, it remembers people and the conversation. You can start building relationships in that conversation. So I think it's a really amazing thing.

(31:16):
Even not to plug my book—you can pick whatever book, E-Myth is a great book too—but I think there's a general base of knowledge about how business works that is really important to establish. The idea of becoming an expert in one book is life-giving. It's not like ChatGPT where you never know. At the same time, it remembers people and the conversation. You can start building relationships in that conversation. It's a really amazing thing. I think there's a general base of knowledge to a business, and having people be able to have that conversation where they become experts in that process is a very powerful thing. So I think it can be both.

(32:00):
I honestly do. But I think this platform solves so many problems I've seen over the years in ways I've never imagined. Some of the things it pulls in from the LLM, I'm just blown away that it's not in my book. Regarding that SMART goal thing, I had the "Four Disciplines of Execution" in my book as the recommended read, and it got right into SMART goals in the most detailed, understandable way for the business owner to get that done. I was just so proud. I'm like, "Now this thing is just high octane." So that's probably the best thing I've done in my banking career, honestly, just to see this whole evolution of technology solve this problem that we all worry about: why do 50% of businesses fail in the first five years? Yeah.

Audience (32:42):
Thank you. I'm with Randolph-Brooks, by the way, from San Antonio, Texas. One of the things that I think about a lot of small business owners is they don't understand how to become bankable. We all understand the folks we deal with: good credit, they do things. But I don't know if your book or some book out there addresses those kind of issues to teach a business owner how to actually become bankable. "This is what your banker is going to look for." It's always about what you put in.

Chip Higgins (33:13):
It naturally has a bias toward that; that's how I spent my banking career. I think the parts of the book where it really shows up is where we talk about velocity and having a direction for your business. That chapter really breaks down what it means to have a direction. For most small business owners, it's just "more." What we go into in that chapter is that there's a segment you're trying to take a piece of, a client experience success thing you want to satisfy, and, by the way, there's a profitability component to this for you to be successful. That sets the tone early in the book about what they need to be worried about. As we get into the chapter on friction—which has a lot to do with organizational traction in a business—it has a lot to do with liquidity and managing margins and how important that is to a banker or investor.

(34:06):
That's something I try to do in the book, too: say it's not just your bank; this is an asset that has value, so protect it. And we get into the leverage part. I admit I made some untoward comments about some of our online lenders, but there are some real painful things I've seen in my life. The idea of having that kind of relationship with the bank is the best possible source of capital besides your own equity. It's a very broad and deep book for 200 pages. It's on Amazon; it's called "The Business Way: Powering Your Small Business to Maximum Momentum."

Nick Miller (34:43):
What I see in this is that when business owners come into a bank, they don't have a business plan. A bank could have a large language model component that says, "Here's how you do the business plan," talks them through it, and asks questions to help them develop it or their marketing plan or financing plan. The point of having Ashleigh and Chip here is to show you two different looks at how the human-automation interface works, and two different possibilities for how we connect people and automation for efficiency on both sides. They'll both be here if you'd like to ask further questions. Thank you for your time.