AI is changing the way developers build software.
Not only because models are getting better, but because the tools around them are becoming more flexible, more powerful, and more integrated into real business workflows.
In this episode of My Data Guest, I spoke with Alan Nichols, CTO and co-founder of Rasa, about how AI is reshaping developer tools, what enterprises need from AI agents, and why building scalable and maintainable systems is still one of the biggest challenges.
Rasa started almost ten years ago with a clear goal: bridge the gap between academic research in dialogue systems and practical tools that developers could actually use.
At the time, many teams were building chatbots for platforms like Slack and Facebook Messenger. But most systems were limited. They could handle simple flows, but they struggled with more complex conversations.
Rasa focused on giving developers the building blocks to create more robust dialogue systems. Instead of hiding everything behind a black box, the goal was to give teams control, flexibility, and the ability to build systems that could be maintained over time.
That developer-first mindset is still relevant today.
With the rise of large language models, the AI landscape has changed dramatically. Before ChatGPT, many enterprise teams were mostly worried about preventing errors. They wanted control, predictability, and strict guardrails.
After LLMs became mainstream, the conversation changed.
Companies started to see the potential of AI systems that could understand language more naturally, assist users more effectively, and create better experiences. The tradeoff is that these systems are less predictable than traditional software.
That means the challenge is no longer only about building something impressive. It is about building AI systems that enterprises can trust.
A key part of this is designing systems that can improve over time. AI agents should not only answer questions. They should learn from interactions, adapt to user needs, and provide developers with the feedback needed to make the system better.
This is where developer tools become critical.
Good AI tooling should help teams understand what the system is doing, where it fails, and how to improve it. For enterprises, maintainability matters as much as raw model capability.
The main takeaway from the conversation is simple: AI is transforming developer tools, but the fundamentals still matter.
Developers need control.
Enterprises need reliability.
Users need better experiences.
And AI systems need to be designed so they can evolve.
The future of AI agents will not only depend on better models. It will also depend on better platforms, better workflows, and better tools for the people building them.
Listen to the episode
In this conversation, we cover:
Rasa’s journey from dialogue systems to AI agents
how AI is changing developer platforms
the impact of LLMs on enterprise adoption
why companies moved from skepticism to experimentation
how to build AI systems with confidence
why maintainability matters in enterprise AI
the future of developer tools and AI agents
