Peter Dun is the Co-Founder and CEO of Feathery, a company specializing in developing form-building solutions for developers and product teams. At Feathery, he leverages cutting-edge AI and no-code technologies to provide wealth managers with flexible and personalized onboarding tools that seamlessly integrate with their tech stack. With a background that includes degrees in artificial intelligence from Stanford University and previous roles such as Growth Lead at Robinhood, Peter has been instrumental in driving innovation in form automation and client onboarding processes.
Here’s a glimpse of what you’ll learn:
[02:21] Peter Dun discusses how Feathery automates wealth manager client onboarding
[03:21] Feathery's unique flexibility in account opening solutions
[07:34] Insights on no-code platforms versus low-code and their application in Feathery
[15:39] Peter talks about Feathery's capabilities in document intelligence and AI utilization
[23:16] The less recognized aspects of artificial intelligence
[28:22] The future of AI in overcoming current limitations
[33:12] Peter’s leadership style and the value of a great partner in business
In this episode…
As financial institutions seek to enhance their client onboarding processes, the integration of a flexible and customizable tool presents a compelling opportunity. Streamlining the customer experience without sacrificing unique branding and established workflows is essential for maintaining competitive advantage. Leveraging artificial intelligence, how can organizations innovate their account opening procedures, transforming them into a more efficient and personalized journey for customers?
Artificial intelligence expert, Peter Dun, uncovers how his no-code, low-code platform equips wealth managers with the agility and control needed to design a client onboarding experience from scratch or customize existing workflows. Drawing from his background in AI, he discusses how this platform streamlines account opening processes and intelligently processes documents, adding a new layer of efficiency for financial advisors. Peter brings insights into the power of no-code tools and the thoughtful application of AI in specific business contexts, providing a glimpse into the future of customer-focused platforms.
In this episode of The Customer Wins, Richard Walker interviews Peter Dun, Co-Founder and CEO of Feathery, about enhancing client onboarding through innovative technology. Peter discusses how Feathery automates wealth manager client onboarding, its unique flexibility in account opening solutions, the no-code versus low-code platforms, and Feathery's capabilities in document intelligence and AI utilization.
Resources Mentioned in this episode
Quotable Moments:
"We automate client onboarding for wealth managers, making it efficient and seamless with our flexible, high-converting client experience."
"Feathery was built as a no-code, low-code tool, giving you full control, similar to a web designer."
"The industry as a whole is becoming more sophisticated with technology every day, becoming more self-serve with the tools they use."
"Document Intelligence uses AI to read your documents and pulls out relevant information based on the info you define in plain English."
"AI looks like it's thinking and reasoning, but it's actually just very good at pretending to be thinking and reasoning."
Action Steps:
Embrace no-code and low-code tools: These platforms empower wealth managers to create customized client experiences without extensive coding knowledge.
Explore AI-driven document intelligence: Using AI to automatically extract and process data from documents can significantly enhance efficiency.
Prioritize client experience customization: Utilizing flexible tools allows businesses to tailor the onboarding process to reflect their branding and meet specific client needs.
Integrate AI in data conversion: Leveraging AI to convert traditional forms into digital formats can save time and improve user experience, tackling the challenge of modernizing data intake processes, and making them more efficient and client-friendly.
Foster collaborative partnerships: Working with synergistic partners can enhance your service offerings and address gaps in your capabilities, overcoming the challenge of lacking in-house functionalities and allowing businesses to focus on their core strengths.
Sponsor for this episode...
This is brought to you by Quik!
At Quik!, we provide forms automation and management solutions for companies seeking to maximize their potential productivity.
Using our FormXtract API, you can submit your completed forms and get clean, context-rich data that is 99.9% accurate.
Our vision is to become the leading forms automation company by making paperwork the easiest part of every transaction.
Meanwhile, our mission is to help the top firms in the financial industry raise their bottom line by streamlining the customer experience with automated, convenient solutions.
Go to www.quickforms.com to learn more, or contact us with questions at support@quikforms.com.
Episode Transcript:
Intro 0:02
Welcome to The Customer Wins podcast where business leaders discuss their secrets and techniques for helping their customers succeed and in turn grow their business.
Richard Walker 0:16
Hi, I'm Rich Walker, the host of The Customer Wins where I talk to business leaders about how they help their customers win and how their focus on customer experience leads to growth. Some of my past guests have included Jason P Carroll of Aptive Index, Brian Thorpe at Wealth Tender and Kate Guillen of Simplicity Ops. Today is a special episode of my series on artificial intelligence. And today's guest is Peter Dun, co-founder and CEO of Feathery. And today's episode is brought to you by Quik!, the leader in enterprise forms processing. When your business relies upon processing forms, don't waste your team's valuable time manually reviewing the forms. Instead, get Quik!. Using our Form Xtract API, simply submit your completed forms and get that clean context-rich data that reduces manual reviews to only one out of 1000 submissions. Visit quikforms.com to get started. Before I introduce today's guest, I want to give a big thank you to Chris Mills, the CEO of the Pacific Financial Group, for introducing me to Peter and the team at Feathery. Go check out their website, at tpfg.com, where they help financial professionals become champions for their clients.
All right, I'm excited to talk to Peter, and in full disclosure, Peter is a resale partner for Quik!, so we've been working together for a while. So I'm excited to introduce him to have on the show. Peter Dun is the co-founder and CEO of Feathery, the platform that enables automated custodial account opening through a flexible, high-converting client experience. Before Feathery, Peter was a growth lead at Robinhood and completed his bachelor's and master's degree in artificial intelligence at Stanford. Peter, welcome to The Customer Wins.
Peter Dun 1:58
Great to be here Rich. Appreciate you having me.
Richard Walker 2:00
Yeah, I'm excited to talk to you. So for those who haven't heard this podcast before, I talk with business leaders about what they're doing to help their customers win, how they built and deliver a great customer experience, and the challenges to growing their own company. Peter, I know I understand a lot about your business, but let's help our audience understand more about your business. How does your company help people?
Peter Dun 2:21
Yeah, totally Rich. Great question. And I think you kind of intro Feathery a little bit, but we automate Client Onboarding for wealth managers that includes the data intake process for the end clients and also the ultimate account opening process behind the scenes with their custodian and also connecting with their tech staff, right? Their portfolio management systems, like an Orion, their financial planning tools, like an E-Money, Money guide. Their CRM is like a sales force or like dynamics. Um, so we're able to intake, find data, synchronize that with their staff, and also get that data into their custodial and systems, of course, use power by Quik! and automate that entire onboarding process for the client.
Richard Walker 3:07
So Peter, I don't know if this is a crowd of the market or not, but there's a lot of different new account opening solutions out there. So what led you to say we have a better solution, the different solution? What is it that's unique about Feathery that people should pay attention to?
Peter Dun 3:21
Yeah, so I'd say, like, the biggest thing about Feathery that I think makes it stand out, and is the reason why we have a different approach is really the flexibility of the tool. I think there are absolutely account opening solutions out there today, but they all tend to be pretty out of the box in the sense of, they'll give you some pre-built templates you can use for a certain workflow, but you have very limited control over actually customizing that workflow, customizing the logic behind the flow, like maybe you want to add additional questions, you want to layer in your risk survey questionnaires. It gets very hard to do that, especially at scale. If you're a larger firm, you care about your custom branding you care about. You really care about the client experience that you're trying to deliver right like you want it to be unique to yourself optimize. It gets really hard to do that with existing account opening solutions.
Feathery was built as a no-code local tool. What that means is, you can, of course, start from our templates, but at the same time, we give you the full control, like full control, similar to a web designer, in terms of drag and dropping, customizing layout, styling, adding logic, connecting and custom APIs, the level of control you need to be able to scale as your firm continues to grow, as well in terms of clients.
Richard Walker 4:39
So when I started in this industry, 2000 2002 is when we launched our business. A lot of the firms I worked mostly with broker-dealers at the time, but a lot of the firms were in the mindset of, we're building it ourselves. We can do it ourselves best. Enter the 2000 10s, and we've got partners that came out and said hey, we can give you an out-of-the-box, off-the-shelf type solution, which often also meant a lot of professional services to customize it to their needs. I would argue. And I wonder if you see it this way, but I would argue that the there's a shift in our industry. They want to go back to build it yourself without building it themselves, which is why I think Feathery is a really interesting tool, because you're giving them a way to have something out of the box, but also configure and control it in a more scalable manner. So I'm wondering what you think about that.
Peter Dun 5:30
Yeah, I mean, 100%. I think that's a big reason why I started Feathery, is because I believe the industry as a whole is becoming more sophisticated with technology every day, and part of that means that they're able to be more self-serve with the tools that they use. There's no reason why a financial advisor or an operation specialist can't go into a tool like learn the kind of basics and fundamentals of web design and be able to actually really take your destiny into their own hands with the types of experiences that they're delivering, is the reason why I started Feathery. And as a little bit more of a some context before this, when I was at Robinhood, we built everything in-house. So we took the opposite approach. We got a full engineering team dedicated to onboarding.
We had product managers, and designers, and it was actually really painful to build, and a big reason was because of the constant iteration and the changes that we have to make on the onboarding flow. My first years at Robinhood, we ran like, 50 plus growth experiments. They were really painful, but they were also really valuable. We ended up doubling our conversion rate just from being able to optimize the flow, right, like, depending on if you're iOS or Android, or depending on, like, your financial history. We're gonna personalize the flow, personalize the questions we're asking you. And because of that, I kind of was like, how can we give this Robinhood life experience with the ability to move fast? But kind of serve that to the rest of the wealth industry?
Richard Walker 7:18
All right? So I want to dive a little bit deeper on this low code, no code kind of framework or this concept, because I think a lot of people, unless you're in tech, they don't really know those terms and really how it translates to them. So what does that mean to be, first of all, no code?
Peter Dun 7:34
Yeah, of course. So being no code means you are able to build out, maybe, like a web experience, like a website or like a custom application, without having to actually write code yourself. You're able to use, like a drag-and-drop framework, right, just click and drag point and click to be able to build out a website or some application that would otherwise required like an actual developer to write, write code to do.
Richard Walker 8:04
Okay? So it's a tool set that maybe has templates of existing designs. But really, I mean, like we do this with DocuSign, right? You open up a document, you drag a field over, now you can sign that form. Yeah, exactly. Program that field, you just had to drag it over. So it's similar to that kind of concept, right?
Peter Dun 8:22
Yeah, DocuSign, right. You can drag it. You can, like, click on it, and then you have, like, style bar opening on the right side. And then you can be like, I want to make my borders, like, a little thicker. I want to change my background colors, like, I want to bring in, like, my firm screening into this whole design. It basically lets you do that without having to type it into.
Richard Walker 8:43
So then what does low code mean?
Peter Dun 8:47
Yeah, so low code is kind of the level above no code in which, like you might be doing something a little bit more complex somewhere, and so you might have to write a little bit of code, a little bit of basic code, maybe a little bit of JavaScript, or a little bit of a custom language. Is common with to list a few examples, what website builders, like a Wix or web-flow, they allow you to maybe drag in, like a custom code block, and then you can paste in your own HTML to that code block if you want to run like, show something custom in the middle of your site, for example.
Richard Walker 9:27
So does your product require people to do low code?
Peter Dun 9:31
Not at all. Not at all. So it's, it's very much no code. You can build all of your designs and experiences without writing any code. I'd say about, you know, 85 90% of our wealth management clients don't touch any code at all. It's really that kind of, like, last 10% of like, really, like, more complex use cases where, like, someone actually wants to be able to write code. Maybe you have some really custom, like. Logic or calculations you're trying to do. Maybe you're trying to call some like, you know, talk to some custom system on your back end that we don't have a native integration to already, where people start caring about writing code a little bit more.
Richard Walker 10:17
Okay, so I don't want to give people the impression that you're a Website Builder, but the web is the tool, right? HTML is the tool that you're living in. And it sounds to me that you're helping customers build out web-based experiences to perform the workflows, the document management, things that they're trying to do in their business. Is that a fair assessment?
Peter Dun 10:39
Yeah, I think it was also forms to an extent, but very much like kind of digital web-based forms for us. So imagine like a Google Forms, right? Like a Google Form is a digital web-based form to use to collect information, maybe similar to that, but you're able to build out like the exact type of client experience you want. For that forum, you're able to embed into your website or host sound Feathery if you want, and then that whole kind of flow is automated behind distance.
Richard Walker 11:14
Okay? So you're helping people build nicer, more user-friendly, new account opening and Client Onboarding type experiences in the framework and the style that they want with very little to no code to do it.
Peter Dun 11:27
Absolutely. Common workflow that we follow is sometimes customers will come to us and they already have the design in mind that they want for this intake experience, right? And they'll show us like a figma file that they have, and they're like, oh, this is how we want it. And one of the most effective ways we've been able to get prospects like that aha moment is, we'll have our services team recreate, like, a page of like that design that they brought to us in February. And with every other tool you're gonna have to make tradeoffs, it's probably gonna look nothing like kind of the initial experience that you envision, but with Feathery, because they can see that like one-to-one replication in Feathery, you're like, oh, crap, this tool can actually like, let me realize my vision for what I want my onboard and close to do.
Richard Walker 12:16
Yeah, look, I had that aha moment a year ago when we first started talking, because you and your partner were talking to us with the customer. You said, Hey, can we call your APIs and generate forums within the system? It was like, a week later, maybe two. You're like, yeah, it's done. Isabelle, it was so fast. I was so jealous of how fast you guys could move. But I think that's really the benefit of a platform like yours, right? It's speed and it's agility to do things.
Peter Dun 12:44
Yeah, absolutely. And in our platform, we built out the fundamentals that allow us to move faster as well, right? Like, do we want to build on new integration or something we already have, the components we already have, like the kind of, like, under the hood, like, back end logic that does, like the network request and stuff so, it's easy for us to move quickly here in terms of building out new integrations and new components on top of the platform.
Richard Walker 13:12
All right, let me ask you a challenging question. What can't you do? What is hard? What is good use case for Feathery?
Peter Dun 13:19
Yeah. So one obvious thing is we can do Quik! right? It's why I think we're such synergistic partners. We are very much focused on like the software experience, the kind of digital experience of creating these customer journeys, but we don't get into a lot of the business logic of wealth management, right? So, like, we don't do any financial planning on our system. We talk to your financial planning tools, and we'll get your data in and out of the tool. But we're not going to do any financial planning ourselves. We're not going to do any portfolio management. We're not going to do any management of the forms that the custodial forms, the insurance and financial forms are using day to day. Like, we don't have that field mapping in our system, but we're able to leverage bus and class tools like Quik! to be able to accomplish that. Like we're just responsible for getting the data in and out from the client to the systems as efficiently as possible, and kind of transforming it where it needs to be.
Richard Walker 14:23
So I was talking to a customer today that is not happy with the different compliance tools that are available to them, and they're thinking about building their own with a partner. Would somebody come to you and say, hey, we need a compliance tool. Is that something you would build?
Peter Dun 14:36
Absolutely um, yeah. So I know we started this conversation on account opening. But the cool thing about Feathery is, I think you really lock kind of the true value when you start thinking about your data intake solutions across the board, beyond just account opening, right like, let's talk about Client Servicing. Let's talk about off-boarding. Let's talk about compliance. Most of an advisors are regular interactions with a client involves some sort of data gathering, and there's no reason why it just has to be during the onboarding process, like, for like, a compliance workflow you could absolutely be in taking information from the customer, powering the tool with like, compliance forms from a quick bring in your own custom forms that need to be pulled out and processed. That's all totally possible.
Richard Walker 15:28
All right, so far, we've been talking about what you can do, but we haven't talked about artificial intelligence. Where does AI fit into your business? What is the AI doing?
Peter Dun 15:39
Yeah, totally. So the biggest place or so there are two places where AI fits into our products. I'll talk about the first place, and then go second. So the first place is we actually automate the process of converting kind of more traditional PDF, paper-based forms into a Feathery form. We actually have a free online school you can check out at feather.ai is very simple. You upload the PDF, and then you click convert, and then our AI system runs in the background, and then it outputs you a Feathery reform, digital Feathery form based on the PDF that you entered into it. The cool thing is, we've actually digitized a number of the kind of quick provided PDFs, so to help customers skip the need to kind of build it again in-house and connect it to the Quik Form. So basically, we run some of the Quik Forms through this process, we are able to build up our library of like digital, Feathery forms connected to the Quik Forms in an automated way, using this tool.
Richard Walker 16:51
So you're giving them a better experience. I mean, look, nobody wants to look at the form. You're giving them a web-based experience to capture the data and still marry it to the PDF on the back end for signature.
Peter Dun 17:02
Yeah, exactly, exactly. And, you know, like, a lot of times they'll bring in, like, you know, their custom forms, right? And so Quik will obviously, like, build up, like, the field mapping there. But then on our side, we can just view it through AI model. We'll get out the Feathery version of that form, and then they don't have to manually build it again themselves in Feathery.
Richard Walker 17:24
So you said there's a second place that AI is being used. That was one.
Peter Dun 17:27
Yeah, so that was the first place. Actually, the smaller use case, the bigger use case is actually document intelligence in the Feathery products. You know one example, right? Like your customers are bringing you their custodial holding statements from their old custodian, like their Schwab or their fidelity. And today you might be reading through these documents, manually grabbing, like their holdings, their ticker symbols, market caps, cost bases, etc. And then you're manually, manually entering these into your financial planning tool, your E-money, or your money guide, pro feather document intelligence basically uses AI to read your documents, pulls out the relevant information based on the information that you want to pull out that you define in plain English, and then, using the same integrations that power our forms, you can send this data into your end systems, so you can automate the process of extracting data from the holding statements and getting them into your E-Money or Your Money guide.
Richard Walker 18:31
Okay, so given your background, was AI, what was your first idea of using AI in business?
Peter Dun 18:39
That is interesting. Yeah. So, at the beginning it was actually less of, I was thinking less about like, kind of like language and Doctor processing. It was more like computer vision. Especially when I was still at Stanford, I was actually doing research, computer vision research in the medical field. So like, image processing for like, X-rays, scans, being able to automatically detect like tumors and other defects in the human body, like that was The type of work I was doing. Obviously, these days, with the development of large language models, the focus has become a lot more on natural language rather than like the computer vision piece. But at the beginning, and back in the day, computer vision was actually the most, I guess, exciting field of AI development.
Richard Walker 19:42
I think a lot of people hear AI and they immediately think, chatGPT, because consumers have grabbed on to that. But really, AI has been developed over a long period of time, and there's different segments. There's machine learning, computer vision, you're talking about. There's all sorts of different things, but computer vision. Is also really, really interesting, is that what Google was training to figure out what's a cat versus a dog, is that what computer does? How would you describe?
Peter Dun 20:10
Or if you've seen Silicon Valley, it's, it's like a hot dog, not hot dog. It's basically taking an image in and then identifying, there's different flavors of computer vision. Of course, there's classification, like, what is in this image? Is it a cat or a dog? Is a hot dog, or is it not a hot dog? There's object detection, right. And also, this year, self-driving cars, you know, super excited to see all the development there. But computer vision is everything that they're powered off, right? Object Detection. Is there a pedestrian incoming? Is there someone running across the street? Is there another car coming towards you? And at what speed? That's all computer vision. When you're opening your iPhone and it uses face recognition to unlock the phone is using computer vision to verify that you are who you are. KYC, like, know your customer, like, you have to take a picture yourself, take a selfie of one of your ID, yeah, that's all AI-powered computer vision.
Richard Walker 21:19
Okay, I'm gonna ask a really dumb question, because it's just a curiosity that I have, and I think others might have. When you go to a website that says, prove you're not a robot, prove you're human, click all of the images that have a stoplight in it. Are we training somebody else's computer vision model by answering the questions?
Peter Dun 21:38
Oh, maybe. But I'm more willing to bet that the model is already trained enough that it doesn't need like, kind of the user input at that point to be able to like, fine-tune it, like they've already created, like, their own custom data set, where, like, they know the answers to every single data point in the set. And then, like, they'll train on that they'll get to a good point. And then when they actually deploy the model for like use with consumers, oftentimes at that point, they don't have to. That being said with things like LLMs, where like data still is like a very big problem for them in terms of continuing development.
If you don't opt out, there is a good chance that they might still be using your data. Obviously, with Feathery by default, we never use your data for any sort of training or third-party usage. But for a lot of the LLMs out there, you kind of have to look into terms and conditions a little bit to make sure you're not, yeah, it's not being trained on.
Richard Walker 22:44
Yeah, that was gonna be one of my questions for you guys. So are you are using LLMs in your model?
Peter Dun 22:49
We are using LLMs as part of, like, the overall model, but we've ensured that, like, none of the data is being used to actually, like, train on your information.
Richard Walker 23:00
Okay, so Peter, one of my often asked questions when I get to meet somebody who's got a degree in artificial intelligence and spent a lot of time studying it and working with it, what is something you know that the rest of us don't know about artificial intelligence?
Peter Dun 23:16
I think, well, I mean, like, there's a few like tangents I could maybe go down. I think one is maybe that, even though, like, it looks like it's thinking and it's reasoning. It's actually not doing that. It's just really good at pretending to be, like, thinking and reasoning about things. So, I mean, these LLMs they use like, token prediction, right? Like, they'll basically determine, like, what is the highest likelihood, like, next, like character, it's like output in the sequence of letters that they're outputting to you, and then they'll output that. And so under the hood, it's, you know, just like, a bunch of like numbers and calculations that combine to create this like, very human-like response that they're sending you, but at
The end of the day, they're just very good at replicating words and phrases and sentences that they've already seen through the massive amounts of documents that were processed, whereas, for humans, like when we were babies and we were learning, it was much more firm, like first principles, like us understanding the world, and then be able to formulate that into words and sentences convey the ideas and emotions that we're already feeling. So I think it's a little nuance, but I think there's a little bit of a difference. I think the difference is what, like, the difference is like, LLMs are really great at that. Like, giving like, very nice sounding things like, sounding like, very bad. Professional. But then the area of biggest development today with these AI models is reasoning, like getting them to actually, like think through deeper things that like might be a little more intuitive for a human, because that's where we started from. But for LLMs, it was like the car before the course, they started talking, and now they're trying to think as well.
Richard Walker 25:22
Yeah, yeah. I've had that experience with chatGPT where I've asked it, I asked it to do complex math, and I learned that chatGPT doesn't do math. Yeah, it's an interpretive thing, right? It's predicting text, so it's not really performing math, but I wanted it to start reasoning with itself. Was saying, did I leave out a variable? Am I actually including all the numbers that have been going along with this process to get to a calculation? And I was building a ROI return on investment calculator, I just had this idea. I'm like, hey, let's do this. And the problem with chatGPT was it kept giving me different results for the same inputs, right?
Peter Dun 25:59
Yeah, because, like, you know, it's not actually like, kind of seeing like, oh, like, Here are the numbers. Like, let me do some math under the hood and return to the result. It's like, oh, here are the numbers like, Let me think about like, the types of like numbers. And sentences I've seen in the past and like, oh, I've typically seen like, this number kind of being returned as a response, maybe to inputs that have these numbers, and they'll give you something. So it's like a different, entirely different type of reason, kind of. It's also why, by the way, we're still invested in document intelligence, like, I know, like agents are very big today, right? And a lot of kind of the tech industry is pushing like, autonomous agents.
Who can like, be autonomous salespeople for your autonomous support agents. But I think for me, at least, there's still a big question as to, like, how far this reason can go, like, how much it can actually correct itself and check itself. But document intelligence is one of those things where it requires AI, but it doesn't require a model to be a genius at reasoning through things. It just requires it to be really good at reading things off documents with high accuracy, which I believe we've already been able to reach.
Richard Walker 27:17
Yeah, I had a friend tell me that his company deployed a bot and started talking to customers and relieving them of payments. Like, oh, you don't have to make your payment this month. No, wait, no, you don't have the authority to say that to the client.
Peter Dun 27:29
Yeah, it's crazy like, or, for example, sales force, at least, like, Agent force. And I saw a comment, which I thought was funny. It was like, I mean, they released agent for us, and then they hired like, 1000 people to sell this thing. And then it's kind of like, well, if the agent is so good, like, should it be able to almost like, sell itself at that point where you still need so many humans to do it. And so kind of the nuance between like, what still needs to rest with a human versus, like, what can be automated with an AI bot, which absolutely there are things that can do that, um, it is still like, very, very nuanced.
Richard Walker 28:12
Do you have any, like, long-term concerns or fears or guesses that AI is going to turn out in a negative way?
Peter Dun 28:21
Yeah. I think the biggest thing for me is, I mean, there's another saying where it's like, you always kind of like, overestimate, like college can be accomplished, like, in a year, but you underestimate how much can be accomplished in like, on the three to five-year horizon. That's true of most technology hype cycles as well, including AI, in my opinion. So I think, like, we're still kind of like, in that like, initial hype cycle, like, I think AI has a ton of, like, very real value. I think we're being a little ambitious at the moment as to kind of the use cases that we can, like, put AI into. But it is, like a fact that the use cases are increasing. The complexity of the use cases are also increasing.
Is just at a slower pace than any of us might think it is. So like, I think in the next year or so, like, my take is, like, it's not gonna look like that much different from where we are now. But I think in five years, right? Like, there probably will be a lot more automation. It might require more breakthroughs in these fundamental models. But I think it is inevitable, given, like, the amount of focus and resources we're dedicating towards it.
Richard Walker 29:44
I think that's a really healthy way to look at it, Peter, because what I think you're saying, is the models and the large language models and the advancement of AI is further ahead of the use cases. And a lot of people have said this, it's a solution looking for a problem. Right? And what's going to happen is it takes companies like ours to figure out, how do we actually use that technology to further our capabilities and solve the problems of our customers? AI, doesn't know the problems of our specific customers and wealth management, for example, it might have some idea of general problems, but it doesn't know our specifics, because it's not talking to our customers the way that we are. So we have to take that AI tech and use it.
Peter Dun 30:25
We're the translation layer, essentially, yeah, like, there's this technology and then there are these customers, right? Like, who have problems, and we have to be the translation layer to apply that technology to the customer. And as much as we want it to happen, right, like, this isn't something that happens like, over the course of, like, a month or even two months. It takes time to understand people and their problems and to be able to embed technology, especially as powerful and sometimes unpredictable as AI, into it.
Richard Walker 30:59
Yeah. All right, so sorry. One other philosophical question about your business, I want to ask, do you think it's important for customers to know that you're AI-driven, or that you're AI you're built on AI?
Peter Dun 31:11
Yeah, I think it depends, because when we're talking like account opening and Client Onboarding, there's no AI, right? Like you don't really pass through the AI piece of the tool unless, like, you have a specific use case around document intelligence. Once we're talking about pieces that involve AI, I think absolutely, because I think they're like, obviously caveats that come with it, right? Like AI is not like, guaranteed. Like, every single time we'll always give you the exact response you want, right? We can't guarantee you like, right, 95% plus for certain use case based on like, kind of the results we've seen. But AI, inherently, is a little bit of a black box processing under the hood, and customers have to kind of be aware of that, so I think transparency is important, but at the same time, not every use case needs to be powered by AI, so just kind of being aware of that as well.
Richard Walker 32:15
Yeah, I'll give an anecdote story for people. Sometimes I'll take sales calls that we record, take the transcript, render AI, and have it, summarize it, and then I'll say it, give me some customer quotes. I'll go see if those customers said those things in the transcript. And sometimes they don't. The AI is like, well, I think it meant this, so I'm gonna give it a quote, right? So yeah, you have to be cautious of that. Look as we wrap up, I have another question for you, but what is the best way for people to find and connect with you?
Peter Dun 32:45
Yeah, absolutely. So a few ways. You can email us anytime at support@feathery.io, you can also go to our website, feathery.io and we have a chatbot. You can talk to us there. You can fill out the demo form and request the demo. So there are a few ways, depending on what you're looking for.
Richard Walker 33:04
Awesome, great. So here's my last question, who has had the biggest impact on your leadership style and how you approach your role today?
Peter Dun 33:12
Yeah, I'd say without doubt it would be my co-founder, Zack Khan. Zack and I started working together about a few years back, and having a partner in crime, especially as a founder, is so impactful in the sense of like, when it's just you yourself, like, you might be great, but everyone has weaknesses, and it's really hard to kind of be able to spot your own weaknesses and point them out and work on them. But having someone who's kind of like a counterbalance to you maybe has different strengths and weaknesses, and being able to kind of give each other feedback, iterate, it's made me recognize, kind of like my biggest areas for improvement as a leader, that I know for a fact I never would have been able to Sell without Zack, so it's something that I will always be grateful to him for.
Richard Walker 34:04
Man, thank you for reminding me of the value of a great partner. Because you're right. You're absolutely right. A great partner sees you in a different way. My COO and I are partners in our company now. I co-founded with my mom as she's retired, and so Don is effectively my partner and partner, we are yin and yang. I mean, we are very different in so many ways, and it's such a compliment because we can call each other out from totally different perspectives and grow together as a result. Thank you for that reminder. Yeah, absolutely. All right. I want to give a big thank you to Peter Dun, co-founder and CEO of Feathery, for being on this episode of The Customer Wins. Go check out Peter's website at feathery.io and don't forget to check out Quik at quikforms.com where we make processing forms easier. I hope you enjoyed this discussion, will click the Like button. Share this with someone and subscribe to our channels for future episodes of The Customer Wins. Peter, thank you so much for joining me today.
Peter Dun 35:00
Thank you for having me Rich. Was a pleasure.
Outro 35:04
Thanks for listening to The Customer Wins podcast. We'll see you again next time, and be sure to click subscribe to get future episodes.
Comments