[ad_1]
As the founders plan for an increasingly AI-centric future, Edward Kim, Gusto co-founder and CTO, said shrinking existing teams and hiring a group of specially trained AI engineers is “the wrong way to go.”
Instead, he said, non-technical team members “can actually have a much deeper understanding than the average engineer about what situations a customer can put themselves in, and what things they’re confused about,” which puts them in a better position to guide which features should be integrated. In artificial intelligence tools.
In an interview with TechCrunch, Kim — whose payroll startup generated more than $500 million in annual revenue in the fiscal year ending April 2023 — explained Gusto’s approach to AI, where non-technical members of the customer experience team write “recipes.” Guide the way to the AI assistant Gus (Announced last month) Interacts with customers.
Kim also said the company sees “people who are not software engineers, but have a little bit of technical thinking, being able to create really powerful, game-changing AI applications,” such as CoPilot — a customer experience tool that has been rolled out around the world. Gusto CX team in June and is already seeing between 2,000 and 3,000 interactions per day.
“We can actually upskill a lot of our people here at Gusto to help them build AI applications,” Kim said.
This interview has been edited for length and clarity.
Is Gus the first major AI product you’ve launched for your customers?
Gus is the big AI function we’ve released for our customers, and in many ways it connects a lot of the point functions we’ve created. Because what you start to see happen in apps is that they get filled with AI buttons, like, “Press this button to do something with AI.” Our message was: “Click this button so we can create a job description for you.”
But Gus allows you to remove all that, and when we feel like Gus can do something of value for you, Gus can show up in an unobtrusive way and say, “Hey, can I help you write the job description?” It’s a cleaner way to interact with AI.
There are some companies that say they’ve been working on AI for a million years but haven’t gotten the attention until now, and others that say they’ve only realized the opportunity in the last couple of years. Does Gusteau fall into one camp or the other?
The big change for me is that when you talk about programming, for most people, it’s not accessible. You have to learn how to program, and go to school for many years. Machine learning has been more difficult to access. Because you have to be a very special type of software engineer and have a data science skill set and know how to build artificial neural networks and things like that.
The main thing that has changed recently is that the interface for creating machine learning and artificial intelligence applications (has become) accessible to anyone. While in the past, we had to learn the language of computers and go to school for it, now computers are learning to understand humans more. This seems like no big deal, but if you think about it, it makes creating software applications much easier.
This is exactly what we saw at Gusto: people who are not software engineers, but have a little technical thinking, are able to build really powerful, game-changing AI applications. We actually use a lot of our support team to extend Gus’ capabilities, and they don’t know how to code at all. It’s just that the interface they use now allows them to do the same thing software engineers have always done, without having to learn how to code. If you want, I can talk about one example of each of these.
That would be great.
There’s this guy who’s been with the company for about five years. His name is Eric RodriguezHe had already joined our Customer Support team and (then) moved to our IT team. While he was on that team, he started getting really interested in AI, and his boss came to me and said, “Hey, he built this thing. I want you to see it.” The first time I met him in person, he showed me what he had built, which was basically a CoPilot tool for our (customer experience) team, where you could ask him a question, and he would give you the answer in natural language. . Just like ChatGPT does, except it has access to our internal knowledge base on how to do things in our application.
At this point, we’re showing this to our support team, and they love it. It has completely changed their workflow and how efficient they are. Basically, any time they get a support ticket, instead of looking at the knowledge base that we’ve created, they actually ask this CoPilot tool, and the CoPilot tool actually answers the question. There is still a human between CoPilot and the customer, but often they can get the response from CoPilot and then copy and paste it to the customer. They check its accuracy, which it often does.
We immediately moved Eric to the software engineering team. He reports directly to me, believe it or not, and is one of the best engineers we have now. Because he was an early adopter of just playing with AI and is now at the forefront of building AI applications at Gusto.
Not everyone is as technically minded as Eric, but we’ve found a way at Gusto to leverage the domain knowledge expertise of non-technical people in the company, especially on our customer support team, to help us build more powerful AI applications and, in particular, enable Gusto to More and more things.
Anytime our customer support team gets a support ticket – in other words, one of our customers contacts us because they want our support team to help with something – and if it happens repeatedly, we already have our customer support team writing a recipe for Juice, which means they can actually teach Gus without any technical ability. They can teach Gus how to guide a client through the problem, and sometimes take action.
We’ve built an internal interface, an internal coping tool, where you can write natural language instructions to Gus on how to handle a situation like this. There is actually a no-code way for our support team to tell Gus to call a specific API to accomplish a task.
There’s a lot of talk right now like, “We’re going to eliminate all these jobs in this field and we’re going to hire AI specialists that we pay millions of dollars because they have this unique skill set.” I think this is the wrong way to do it. Because the people who will be able to develop your AI applications are actually the people who have experience in this field, even though they may not have the technical expertise. We can actually upskill a lot of our people here at Gusto to help them build AI applications.
The scary AI scenario is this top-down scenario, where executives say, “We need to use AI,” and it’s disconnected from the reality of how people work. This seems to be the most important thing, as I’ve built tools to let teams tell you what AI can do for them.
exactly. In fact, the non-technical people who are closest to the customers, they talk to them every day, and they actually have a much deeper understanding than the average engineer about what situations a customer can put themselves in, and what they’re confused about. So they are in a better position than engineers or AI scientists to write instructions to Gus to solve this problem.
I think other people I talked to noticed the same thing. The best AI engineers are actually people who are domain experts who have learned how to write good claims.
When you think about how this will play out over the next few years, do you think the company’s headcount across different teams will look pretty similar, or do you think this will change over time as AI is deployed throughout the company?
I think the role is evolving a little bit. I think you’ll see a lot of our CX people not directly answering questions, but actually writing recipes and doing things like quick tuning to optimize the AI. Everyone will move to the top of the abstraction layer, and that will obviously bring more efficiencies to the company and also a better customer experience, because they will get answers to their questions immediately.
This opens up the possibility for Gusto to do more things for our clients. There is a huge road map of things we want to do, but can’t, because our resources are limited.
[ad_2]