I’m finding it harder to get excited about building another online tool.
Not because software has become less useful. I use more of it than ever. But increasingly, I want to ask for something and get the result. I don’t particularly want to find a website, create an account, and learn where someone decided to put the export button.
As a user, that feels like progress. As someone who builds these things, it makes me uncomfortable.
The useful part might survive without the website
Say you build a tool that helps people compare services. You spend time making the results readable, the filters useful, and the experience pleasant.
Then an agent comes along. It gathers the options, applies the user’s preferences, and gives them a recommendation. Maybe it uses your service along the way. Maybe it finds another source. Either way, the user never sees the interface you spent weeks building.
Your work might still be useful. But being useful inside someone else’s workflow is a very different position from having customers who deliberately choose your product.
Will they know your name? Will they subscribe? Will they ever find the other things you built?
That’s the part that worries me. The capability can keep its value while the business around it gets weaker.
We can already see a bit of this
Pew Research Center looked at Google browsing activity from 900 US adults in March 2025. When an AI summary appeared, people clicked a traditional search result in 8% of visits. Without a summary, that figure was 15%. Links inside the summaries got clicks in just 1% of visits.
That doesn’t prove agents will replace browsing. Search summaries and agents completing tasks are different things, and the study doesn’t establish that the summaries alone caused the difference.
But it does show that people can get something from the web without visiting the websites supplying it. Now extend that possibility from answering a question to completing a task.
I don’t have an emotional attachment to filling out forms. If my agent can compare plans, convert a document, or pull together a report, how often will I open the individual applications involved?
Agents can also interact with existing websites through a browser, so this doesn’t depend entirely on every service building a dedicated integration. Clear information and reliable ways for agents to act could become an advantage. They won’t automatically give the builder a sustainable business.
Experiences feel different
There are things I go online to do because I want to be there.
Playing a game. Making music. Exploring a strange interactive project. Talking with people who care about the same obscure thing I do.
Having an agent finish those activities for me would often remove the reason I started.
That gives me a bit more hope about building online. There’s still value in making something people want to spend time with, where participation matters.
The distinction isn’t perfectly clean. A creative tool can contain tedious work worth delegating. A practical dashboard can help someone explore something they couldn’t see before. Reliable infrastructure, unique data, and trusted services can remain valuable without anyone looking at their interface.
But it changes the question I ask about an idea: which parts would someone gladly hand off, and which parts would they actually want to keep doing themselves?
Who gets to choose what gets used?
We might move from trying to rank in search results to trying to become the service an agent selects.
That choice could depend on quality and reliability. It could also depend on default integrations, commercial agreements, or which providers are easiest for the agent platform to work with.
I don’t know how that settles. You could build a good service and still find that the route to its users belongs to someone else.
For an independent builder, that’s unsettling. Making something useful is already hard enough without also wondering whether people will ever know it exists.
What I want to build now
I’m still interested in building online. I’m just less convinced that another destination website is always the right shape for an idea.
If I’m building a capability, I want to understand why it would remain worth choosing when an agent can compare it with everything else. If I’m building an experience, I want to understand why someone would want to participate.
Right now, I’m more drawn to that second kind.
Something you’d still want to open, even if your agent offered to handle it for you.