Challenges of Building AI Tools That Truly “Understand” User Intent
Tim Liao
4 replies
In AI product development, interpreting user intent is a critical challenge—especially when users are exploratory or vague. How do you guide users effectively without overwhelming them?
For a long time, most SaaS products have been “expert systems”: the workflows and user inputs were strictly predesigned by developers. If you wanted to tell the system your preferences, you’d fill out a form or select predefined options. Essentially, the user adapted to the tool.
But in the AI era, I see a fundamental shift: the system should actively adapt to each user. Instead of a rigid, form-based flow, AI-driven products can accommodate fluid, natural inputs—letting users express their intent however they like. This is a step beyond “user-friendly”; it’s what I call “human-like” design, where the software meets people on their terms.
Join the Discussion:
I’d love to hear from you! If this perspective resonates with you, feel free to share your thoughts, ideas, or examples in the comments.
To spark discussion, here are a few questions to those great makers and marketing master in product hunt I’m curious about:
- How do you build systems that guide users without overwhelming them, especially when users aren’t sure what they want?
- Have you implemented frameworks or strategies to make AI systems adapt to users in real-time? What worked, and what didn’t?
- What’s your approach to validating whether your AI truly captures user intent, instead of forcing users to adapt to preset flows?
- Do you know of any products or tools that have successfully achieved this kind of “human-like” adaptability?
Replies
Tim Liao@timliao
Minduck
Launching soon!
This is Minduck Discovery’s Series #6: each article is crafted with high quality, sharing deep insights into global AI applications, along with valuable discussions. We’re confident it’ll provide value to the community—check it out!
Don't miss our next article: #7 How Can AI Revolutionize the Way We Explore and Structure Knowledge?
Share
I've had similar questions when working on my product (users write the search request, specifying any details if needed) precisely - I had to understand the intention behind each search prompt, if there are any details and any requirements that are optional or necessary.
My approach is that I **do not** try to use LLM to do all of the analysis, instead, I use LLM to do what's oftentimes called "feature extraction". Simply put, I use LLM as some kind of a "smart pattern matching" tool to extract information that I want to know. Then, typically, I write my algorithms/logic on top of that.
So in my case, I ended up having a processing pipeline that roughly looks like this: "shallow analysis (LLM)" -> "feature extraction (LLM)" -> "normalization, ranking and processing (custom logic)".
That said, it does not give the universal "human-like" adaptability you mentioned since the custom processing logic needs to be written for a specific use case.
TL;DR: I personally use LLM for tasks where it excels, while replacing its weakest points ("real reasoning", "thinking") by implementing my own (more deterministic) logic/algorithm on top of that. I believe that for **reliable** (predictable) behavior, LLM results should be processed by "classical" machine learning or NLP techniques.
@dani_ab Hi Daniel,
Thank you for your detailed response—I truly appreciate the effort you put into sharing your thoughts! I find your approach fascinating and would love to learn more.
1. Could you share more about the type of product or project you’re working on? Understanding your context would help me grasp your concepts more clearly.
2. Why do you prefer not to let AI directly analyze user intent? Considering natural language understanding is a key strength of AI, wouldn’t this be effective for tasks like intent analysis or automating customer support?
3. You mentioned a process with steps like normalization and ranking. Could you share a concrete use case to help illustrate this? My approach is to abstract and categorize user behaviors, then design workflows that match solutions to common patterns.
4. You encourage users to input more detailed needs, but wouldn’t this create friction for users who are less inclined to plan or articulate their thoughts? How do you balance guiding users while minimizing effort on their part?
Looking forward to hearing your thoughts
@timliao Hey Tim,
1. The platform I've built lets users describe what they are looking for, and it then connects them based on matching/complementary needs (so it's a networking platform; you can check more details on ProductHunt; it's called X76 🙂).
2. I do use LLM to perform the initial analysis of the search requests, check them for safety, and extract all necessary information, based on which I'll then perform further processing and the matching logic
3. Yeah, what you've described sounds similar to what I do in my case. I.e., after extracting all the necessary bits and pieces from the users' search requests, I try to categorize and cluster them, after which I can rank the results by relevance to each other within the cluster.
4. I don't specifically try to encourage that. Instead, I try to do the following: if the inputs from the users are short and straightforward, then the results/matches they will get are also pretty generic (which may be fine for many users if they don't want to go into specifics). Those with more concrete and specific needs can write them down and get more specific results for their particular needs. In other words, instead of actively encouraging it, it kind of works in an intuitive way (similar to how it would work if the matching were done by a real person and not by the algorithm).