Building AI Outside the Chatbot

A post from the founder of Zenhub , Aaron Upright , got me thinking.

In 2022, chatbots were the hottest new thing. Every new YC startup seemed to have launched a copilot or chatbot for some industry or use case—copilots for legal, enterprise sales, robotic process automation, and more!

These copilots merely functioned like chatbots, allowing users to ask questions in the same format you’d use when interacting with ChatGPT. There are several reasons for this. First and foremost, when this new tech was being released, we didn't know any better, so we took the safe bet, using the format people associated with AI.

The world was conditioned to believe that LLMs were best introduced through chat interfaces.

However, as we've moved towards AI agents, we're starting to see some unique user experiences. Teams like Figma have launched “feeling boards.” I don't know what they are called, but they are a novel and different way to work with AI

Still, these experiences are rare now, and despite this shift to agentic tools, users will encounter the same tired design patterns. Of course, hindsight is 20/20. In early 2023 Artemis proudly launched a chatbot app. We thought this was the user experience customers wanted, but we have learned a few things since launching.

AI is Unreliable (duh!)

We all know that LLMs are not deterministic, meaning we can’t control them?the same way we can with a typical SaaS tool. This means your app?will?make mistakes for the user. You can try prompt engineering to make the tool hallucinate less, but it will still produce various outputs. Users need a way to verify the output, make changes or edits, and rerun tasks from specific points.

Transparency is Paramount

A direct symptom of point #1 is that because total control of the tool is impossible, you need complete transparency instead. Avoid black boxes—you must show both the input and output and the steps used to arrive at the final number, answer, etc.

The Magic Is In the Product, Not the LLM

This is more nuanced. We experienced the “Oh shit; this just made me understand Nuclear fusion” moment when we used ChatGPT for the first time. I thought that magical moment came from the LLM, and therefore, we built products with the LLM at the center, hoping users would experience that exact moment in our product. This is met with depressive reactions since it will rarely replicate.

The magical moment came from the chat experience, not the LLM. The real banger was the ability to ask a question and get an answer written to you. However, most tools don't need a chat-and-answer format, and most experiences are inherently worse because of it.

So what does this mean? As LLMs are commoditized and the magic shifts to the AI tool, we need to evolve past the paradigm of chat interfaces.

I personally struggled to see what this next experience was until one day, I was chatting with our CTO, and he casually described how we were redesigning the platform. He described a task list experience, where we surface insights to our users, create tasks to solve those insights, and have the user plan and execute the tasks like they would execute a roadmap in Jira or Linear.

At that moment, it clicked. The user is continuing a behaviour it already has, writing tickets for a team's work for the week; the only thing is it can delegate those tickets to our agents. If the user can follow the entire task journey, it can ask the agent to perform the task later and be notified when it's done. They are not waiting for a chatbot to respond. They are not learning a new behaviour. They are not interrupting their current workflow. They have a full line of sight to the solution the agent takes and can edit any output it creates.

By moving away from chatbot interfaces and towards a more familiar task-based system, we're not just making AI more accessible - we're making it more practical and efficient. Great AI tools should use the systems and tools that humans already love. I don't really need an agent that can do accounting, it would be more helpful if it could just use Quickbooks.

As we move forward, let's challenge ourselves to think beyond the chatbot paradigm and explore new ways of working. The future of AI isn't just about smarter algorithms—it's about more intelligent, more intuitive user experiences that truly empower us to innovate, create, and explore.