Chatbots are everywhere. They answer our support questions, help us book hotels, track orders, and – if lucky- solve our problems. Companies are pouring resources into chatbot technology at record rates, with the global market expected to reach $25.88 billion by 2030 (Mordor Intelligence). The promise? Faster, more efficient customer service, available 24/7.
But when you ask users how they feel about chatbots, you don’t always get rave reviews.
The frustrating reality
We’ve all been there: You need a simple answer, maybe you’re updating a shipping address or checking on a last-minute flight change. A chatbot pops up and cheerfully asks, “How can I help you today?” You type in your request, only to see:
- “Here’s a link to our help center.”
- “I’m sorry, I didn’t quite understand that. Can you rephrase that?”
- A long, generic answer that doesn’t address your specific question.
You try again. And again. Soon you’ve repeated yourself three times with no resolution. At that point, the chatbot stops being a convenience and becomes a barrier.
So what separates a chatbot that feels intuitive and even fun from one that makes you want to abandon the conversation altogether? It’s not just sophisticated AI or fancy features.
It’s conversational design.
The hidden skill that makes or breaks a chatbot
Most chatbots fail simply because they’re not designed like conversations.
Instead, they’re built as decision trees, marketing funnels, or ticket diversions – all of which make sense from a business perspective, but not from a human perspective. Real people don’t interact with chatbots the way they fill out forms. They type the way they think. They interrupt themselves, they change their minds, they misspell things, they assume the bot understands the context.
Without a conversation designer to shape these interactions, chatbots misunderstand intent, force users into rigid flows, and create frustrating dead ends. The result? A poor user experience and, ultimately, lost trust in the brand.
And trust? You usually don’t get a second chance at it.
We’ve been trained to lower our expectations
Let’s face it: most of us don’t expect much from chatbots. We’ve been conditioned to see them as clunky, unintuitive, and often annoying. Even if we do get help, it might be after multiple rewrites, endless button clicks, or a desperate search for the “talk to a human” option.
But it doesn’t have to be this way. Some brands are getting it right, proving that chatbots can feel natural, helpful, and (dare we say it) even fun. Their secret? They don’t just build chatbots—they design them. Conversation design isn’t an afterthought; it’s the foundation.
User-centric chatbots is a mindset shift
If you’ve ever sat in a meeting where the first question about chatbots was, “How can we reduce support tickets?” you’ve seen where the trouble starts. Automation can certainly improve efficiency. But when you start with efficiency instead of user experience, you end up with chatbots that serve the business first, not the user—and frustration inevitably follows.
A chatbot should be useful before it’s efficient.
The best ones nail three critical areas:
1. They solve first, sell later
One of the reasons people get frustrated with chatbots is that they feel like salespeople in disguise. If someone asks for help with their bill, don’t immediately push them to upgrade their plan.
Example: Banking done right
A user wants to check his account balance.
Bad Bot: “Would you like to sign up for a credit card?”
Good Bot: “Here’s your current balance. By the way, we also offer savings options if you’re interested.”
2. They feel like a conversation, not a script
Chatbots should grasp intent, not just keywords. When someone asks, “Can I check in early?” the bot shouldn’t just spit back policy info. It should recognize the request (early check-in), ask for reservation details, and provide a genuine answer, or hand off to a person if needed.
Good conversation design anticipates the different ways people might phrase the same request. Some say, “I need to send something back,” others say “Can I return an item?” or “I got the wrong size—help!” If your bot only recognizes one of those, the user experience collapses.
3. They offer an easy way out
Few things are more frustrating than being stuck in an automated loop. A well-designed chatbot knows its limits. If it detects user frustration or a complex issue, it should seamlessly hand things over to a human, without forcing the user to beg.
Example: E-commerce chatbot
User: “I was overcharged on my last bill.”
Bad Bot: “View your recent transactions here.”
Good Bot: “I see there’s a billing issue. Let me connect you with an agent who can help.”
Knowing when to let go of automation can be the difference between a user who swears off your chatbot and one who appreciates it.
Here’s how the role of the conversational designer comes in
Some believe that more powerful AI will solve chatbot frustrations. It won’t. You can have the best natural language processing in the world, but without solid conversational design, users will still feel lost.
Conversational designers combine user research, copywriting, UX strategy, and an understanding of human psychology. They identify points of confusion, craft language that sounds natural yet professional, and make sure there’s a clear path for users to get the help they need, especially when the bot can’t solve something on its own.
When done well, conversation design solves more than just immediate problems:
- Higher satisfaction: People who get quick, clear answers walk away with a better impression of the brand.
- Lower abandonment: A user-friendly chatbot means fewer frustrated abandonments.
- Better conversions and retention: Customers who feel heard are more open to offers and more likely to return.
- Gain valuable insights: Well-designed chat transcripts reveal what users care about and guide future improvements to products, services, and marketing.
Make it work in practice
Still on the fence about chatbots? Start small:
- Pick one core use case
Focus on a single high-frequency need—like checking shipping status—and make that interaction seamless. - Test with real users
Encourage teammates and pilot users to approach the bot with real questions. See where they get stuck or frustrated. - Iterate and refine
Chatbots aren’t “set it and forget it” tools. Regularly review analytics and feedback to refine conversation flows.
If you already have a chatbot that causes more grief than gratitude, small conversation design tweaks, like clarifying prompts, acknowledging when it’s stuck, or rearranging the order of information, can produce big wins in user satisfaction.
The bottom line: Better conversations, better experience
As AI capabilities grow, chatbots will only become more sophisticated. But the heart of a great interaction remains the same: understand what users need and guide them with empathy and clarity. That’s the difference between conversation design and a simple Q&A script. It’s a promise that the user’s time and concerns are respected.
This shift often sparks broader benefits, too. When you embrace conversation design, you inevitably simplify complicated returns policies, streamline shipping information, or clarify product features, which are improvements that boost the overall user experience well beyond the chatbot itself.
Do you want more insights like this directly into your mailbox? Good news, our Product Design community shares them monthly in our Product Design Newsletter, to which you can subscribe here.
Are you looking for a partner to help you with your AI use cases? Contact us, we would be glad to see how we can help you.