Conversational Interfaces and Chatbots: The Decisions That Matter
Interfaces where users interact through written or spoken conversation, including support chatbots and AI assistants embedded in products.
Priya Raghunathan
Why does it matter?
Chat can make some tasks easier, but poorly scoped bots frustrate users and hide paths to human help.
What are the numbers?
- Scope bots work best for narrow, well-defined tasks
- Handoff users need a clear route to a human
- Expectations users should know they are talking to a bot
- Accessibility chat widgets must be keyboard and screen reader accessible
What should I do?
- Define what the bot can and cannot do
- Say clearly that users are talking to a bot
- Offer quick replies for common tasks
- Hand off to humans with the conversation context
- Review transcripts to improve responses
What should I avoid?
Avoid:
- Bots that pretend to be people
- Endless loops with no escape
- Blocking contact options behind a bot
- Launching without testing real questions
When should I get help?
Short answer Bring in help when adding chat to support or sales, or when bot satisfaction is low.
Where this comes from
- Nielsen Norman Group — The user experience of chatbots
- Microsoft Learn — Guidelines for human-AI interaction
The figures and practices above come from the sources listed.
Working on something like this?
We take on UI & UX Design work for teams who want it done once, properly. Tell us what you are building and we will tell you honestly whether we are the right studio for it. Start a project.
Where to go next
Spotted something wrong? Report an error on this page. We correct on the page and say what changed.
