đ usncan Note: 3 Lessons Learned From Building Our First AI Agent
Disclaimer: This content has been prepared based on currently trending topics to increase your awareness.
Photo by Aidin Geranrekab on Unsplash
UNSPLASH.COM
Having led a business for nearly 20 years, Iâve overseen a lot of product launchesâand somehow, I still learn something from every single one. But recently, my company, Jotform, waded into entirely new waters with the release of our first AI agents.
Our foray into agentic AI has been a long time comingâwe first started batting around ideas for how this new technology could simplify customersâ lives back in 2023, well before the term âAI agentâ became a household term.
The last few months have seen a lot of hype around AI agents, from bombastic promises about their disruptive potential to doomsday prognosticating about them taking over the world. But no matter what side of the discourse you fall on, the reality is that theyâre changing how business is done. As leaders everywhere consider how to implement agents into their own operations, I wanted to share a few lessons we learned from creating ours.
Listen To Data, Not Hype
As with any other product, your original idea may undergo some significant transformation by the time itâs released.
When Jotform first explored AI integrations, the tech world was buzzing with generative AI hype.
During one of our regular hack weeks, our teams came up with an idea that we were sure had legs: AI-powered forms that users could generate and customize, simply by interacting with the form itself.
The team quickly launched a beta version, putting it into usersâ hands to gather real-world feedback. It proved to be a hitâpeople were using the product, and it had a strong rate of retention.
But examining the data also taught us something both unexpected, and extremely valuable: The vast majority of our usersâ90 percentâwerenât using the new tool for forms; they were using it for customer service.
This wasnât what we initially had in mind, but the people had spoken, and it was up to us to listen. Using these insights, we made adjustments with this new use case in mind. The lesson? As with any other product, the success of your AI agent is reliant on your ability to pivot. Donât be afraid to listen to what your customers are saying, even if it doesnât align with your initial vision. The data never lies.
Efficiency Over Friendliness
Even casual followers of AI news have probably heard about the dystopian phenomenon of AI companions, in which human relationships are replaced by AI systems. This turn of events might make leaders assume that their agents have to come equipped with sparkling personalities in order to be successful.
Turns out, users arenât especially interested in whether the agent theyâre working with is a charming conversationalist. What they really care about is whether itâs efficient.
In the past, customers have expressed a strong preference for talking to a human over a chatbot, likely because their experiences with the latter usually ended in frustration. But in our testing, we found that users actually prefer interacting with agents: Not just because of the speedier responses and around-the-clock availability, but the fact that they are less inhibited than they are when interacting with a real person. With agents, users donât have to worry about asking dumb questions, or repeating the same question several times. With the pressure of being pleasant and agreeable off, users are more likely to get the help they actually need.
Whatâs not important? The agentsâ people-skills. In fact, research from Salesforce found that overly human-sounding agents present a liability. âWhen AI agents are given human-like personalities, users may begin to relate to them as they would to a person,â explains Yvonne Gando, Salesforceâs senior director of UX.
âIf the agent canât meet those expectations or misunderstands a question, this often leads to disappointment and a loss of trust in the brand. Itâs more effective to keep AI interactions helpful and approachable without creating a false sense of human connection.â
In short, make your agent pleasant and helpful, but donât try to make it something itâs not: human.
Reach Your Customers Where They Are
As we continued to develop our agents, we realized the value of integrating them across multiple channels, allowing the option to connect over websites, phone lines, messaging apps like WhatsApp, and more. Their 24-hour availability and ability to interact in multiple languages also breaks down the barriers that have traditionally prevented users from getting their questions answered, all while making them accessible. Knowing your customers is key to understanding their needs, and it was important to us that no matter how a customer interacted with a Jotform Agent, the experience felt cohesive.
AI agents arenât just a fleeting trendâtheyâre genuinely reshaping how businesses connect with customers. By listening to data, prioritizing efficiency, and meeting users where they are, companies can create AI agents that truly add value to peopleâs lives. Like any other product you develop, maintain a sense of flexibility and willingness to adapt to changing realities. You wonât regret it.
