Recently, I helped a friend build an automated bidding lead collection tool. His requirement was actually very simple: based on the company's business scope, search for relevant bidding information on search engines every day and filter out potentially valuable leads.
He had been doing this manually, spending a significant amount of time every day. So he wanted an automated tool to replace this repetitive work.
I initially thought this was a very simple requirement. Web scraping technology is already mature, and using an AI IDE should be enough to write the project. Moreover, since we were already using AI, why not add some LLM capabilities, such as automatically parsing company materials to generate keywords and incorporating LLM evaluation and analysis into the workflow.
So I quickly got started.
I first outlined the requirements with Gemini, wrote a requirements document, and created an interaction specification. Then I placed both documents in the project root directory, letting Codex automatically generate a development plan based on the documents and implement it step by step. Since this tool is a standalone application using Playwright for browser automation, not a large-scale scraper, the implementation difficulty and compliance were relatively manageable.
After a couple of days of trial and error, I finally produced the first version. Users could upload company materials, the LLM would parse the information and generate keywords, then call Playwright to perform searches on search engines. Users could also upload company qualifications, and the system would use the LLM to evaluate each captured bidding piece for matching.
But the results were not ideal.
Many portal website results couldn't be drilled down further, only capturing a title. The result list was mixed with a large amount of news content, and some bidding information was even outdated. After repeated discussions with the AI, I tried to strengthen filtering rules and changed the portal website list pages to targeted scraping, but the effect was still mediocre.
I stared back at the ugly interface and quickly realized where the problem lay.
I asked the AI to help me build a tool, but what came out was a tool from an old era.
Although I had never done web scraping before, I knew that writing rules, cleaning data, and maintaining site adaptation were complex tasks in themselves. And I already had AI, yet I was still solving problems in a traditional way.
So my problem wasn't technical; it was in my way of thinking.
My product mindset was still stuck in the past, treating AI only as a "faster programmer." But AI isn't just for building tools; it is itself part of the tool.
So in the second phase, I started to adjust my direction, reducing rules and structured workflows and shifting toward semantic processing.
Solutions like browser-use are more intelligent but consume too many tokens, so I chose to add a simple layer of intelligent decision-making on top of Playwright, giving the program some understanding when browsing and clicking.
At the same time, I rethought the true goal of this project. Its goal wasn't "better," but "replacement." It only needed to replace my friend's daily manual search process, saving him from spending time on repetitive tasks and allowing him to focus on more critical matters. As long as this was achieved, it would save him 90% of the time spent on this task.
Looking back, many of the things I did earlier were actually unnecessary.
For example, more complex keyword generation or more detailed data evaluation. These tasks didn't need AI to do them for him, because he had been using his own keywords to search and could quickly judge which information was valuable. This shows he already had that capability. What truly needed to be replaced was only the search itself.
Over-developing these analytical capabilities actually led me away from the core of the problem.
People are always the main body of things; tools are just extensions of people.
This experience also made me rethink a bigger issue.
Some scenarios do require AI to participate in value judgment, such as when I use AI to write code, because many technical problems are beyond my ability. But most of the time, people don't need to completely outsource their value judgment to AI.
The tool for my friend is still under development, but the story of this mistake can end here.
Just like many companies implement intelligent customer service to reduce costs, or even to block responsibility. But the original purpose of the customer service role was to soothe consumer emotions, collect genuine feedback, and thereby improve products and service experiences. It was a channel connecting companies and consumers, but later many companies turned it into a "firewall."
Now with AI, it's an opportunity for a new division of labor. Repetitive, mechanical, and procedural communication can be handed over to AI, while human customer service can focus more on understanding consumers, solving problems, and optimizing products. Let professions return to value rationality.
Technology should not diminish human value. What it truly needs to take over are those dead processes that make people more mechanical.