When Chat Is the Wrong UI: Rethinking Conversational Interfaces in 2026
By Burke Holland (@burkeholland), Technologist, author, speaker, and creator at GitHub, focusing on AI-powered development.
As AI-powered development tools continue to evolve in 2026, the chat interface has become the default mode of interaction. From coding assistants to customer support bots, chat is everywhere. But is it always the right choice? In many cases, chat is a square peg in a round hole—and it's time we acknowledge that.
The Allure of Chat
Chat feels natural. It mimics human conversation, lowers the barrier to entry, and requires no learning curve. For open-ended tasks like brainstorming or exploring ideas, it's hard to beat. This is why tools like GitHub Copilot and other AI assistants have embraced conversational UX so enthusiastically.
But chat has a dark side: it's linear, ephemeral, and often inefficient for structured tasks.
Where Chat Falls Short
1. Complex, Multi-Step Workflows
When you need to configure a CI/CD pipeline, debug a distributed system, or design a database schema, a back-and-forth chat is clunky. You lose context, scroll endlessly, and struggle to compare alternatives. A visual builder or a form-based interface would be far more effective.
2. Precise Input and Output
Chat is terrible for entering structured data or receiving precise results. Imagine ordering a pizza through chat versus a well-designed menu. The menu wins every time. Similarly, in development, tasks like writing a regular expression or formatting JSON are better served by dedicated tools with validation and syntax highlighting.
3. High-Frequency, Repetitive Actions
If you're performing the same action dozens of times a day, chat is a bottleneck. Keyboard shortcuts, command palettes, and buttons are faster and more reliable. Chat adds cognitive load and latency.
4. Collaborative and Persistent Context
Chat threads are personal and linear. When multiple people need to collaborate on a task—like reviewing code or planning a sprint—a shared canvas or board is far superior. Chat scatters information across threads and makes it hard to track decisions.
5. When Users Don't Know What to Ask
Chat assumes the user knows what they want. But often, users need guidance. A well-designed UI can expose options, suggest actions, and guide users through a process. Chat leaves them staring at a blank prompt, overwhelmed.
The Rise of Hybrid Interfaces
In 2026, we're seeing a shift toward hybrid interfaces that combine the flexibility of natural language with the structure of traditional UI. Think of an IDE where you can chat with an AI assistant, but also click buttons, drag-and-drop components, and fill out forms. The AI understands context from your actions, not just your words.
GitHub's own tools reflect this trend. GitHub Copilot now integrates with pull requests, issues, and code reviews, offering suggestions inline rather than in a separate chat window. The chat is still there, but it's one tool among many.
Principles for Choosing the Right UI
So when should you use chat, and when should you reach for something else? Here are some guiding principles:
- Use chat for exploration and open-ended tasks. When the problem is fuzzy and the path is unclear, conversation helps.
- Use structured UI for well-defined tasks. If the steps are known and repeatable, a form or wizard is better.
- Use visual UI for spatial or comparative tasks. Diagrams, timelines, and dashboards excel where chat fails.
- Use shortcuts for frequent actions. Don't make power users type commands they could trigger with a keystroke.
- Combine both for the best of both worlds. Let users switch seamlessly between chat and direct manipulation.
The Future of AI Interfaces
As AI models become more capable, the temptation to slap a chat interface on everything will only grow. But the best designers will resist. They'll ask: What is the user trying to accomplish? What is the most efficient way to get there? Sometimes that's chat. Often, it's not.
In 2026, the most successful AI products will be those that use chat strategically—not as a universal solution, but as one modality among many. The future is multimodal, context-aware, and user-centric. And sometimes, that future looks a lot like a well-designed button.
Burke Holland is a technologist at GitHub, where he focuses on AI-powered development. He writes about developer tools, UX, and the future of software creation.
via GitHub AI Blog