Map who is already doing this before you commit a quarter to it

Run a structured competitive scan that shows what exists, what you missed, and where the gaps still are.

For anyone about to start something someone else has already built · 8 steps · 7 min

How it works today.

You start with a Google search for the obvious keywords, click through the first page of results, and skim a handful of landing pages. You check Product Hunt for anything recent, maybe ask a few people if they've heard of competitors, and conclude the space is either wide open or crowded based on what turned up in twenty minutes.

The problem surfaces later. A investor mentions a direct competitor you never saw. A user says they tried another tool first. You realise your search missed entire categories—open-source projects on GitHub, vertical solutions hiding behind industry jargon, or international players who don't rank in your geography. By then you've already committed budget and shaped a roadmap around incomplete information.

Before you start.

All of it has to be true, or step one fails in a way that is annoying to debug.

Access to Perplexity Pro or similar research agent with current web search and citation
A one-paragraph description of what you're building and who it's for
A shared doc or workspace where you can log findings and gaps as you go
Willingness to read ten landing pages yourself before you conclude anything

The steps.

List every phrase a searcher would use if this already existed
Write down the job-to-be-done in plain language, then the category names (both emerging and established), then the feature clusters someone might search by. Include the jargon your industry uses and the plain terms a newcomer would try. You need at least eight phrases before you search anything.

Run the broad search and capture what the agent surfaces

Paste your description and phrase list into Perplexity or another research agent. Ask it to identify existing tools, open-source projects, and recent launches in each phrase cluster. Copy the results into your shared doc with the search date and agent used, so you know later what was covered.

Paste this

I'm researching competitors and adjacent solutions before building [one-sentence description of your product]. Using these search phrases: [your eight-plus phrases], identify existing tools, open-source projects, and recent launches (last 24 months). For each result, provide the name, URL, primary use case, and whether it's commercial, open-source, or discontinued. Flag any that serve [your target audience].

Check GitHub and GitLab for projects the web search missed
Research agents often surface commercial products more readily than repositories. Search GitHub directly using your core keywords and sort by stars, then by recent activity. Look for archived projects too—they show you what was tried and abandoned. Log the repository names and star counts.
04
Ask the agent to surface vertical or geographic variants
Many competitors hide behind industry-specific language or operate in regions where they don't rank globally. Prompt the agent to look for solutions tailored to specific industries, languages, or markets adjacent to yours. This is where you find the uncomfortable comparisons.
Paste this
```
For the same problem space—[your one-sentence description]—identify solutions that target specific industries (e.g., healthcare, legal, education), non-English markets, or adjacent use cases that overlap with [your primary use case]. Include any that rebrand general tools for a vertical.
```
Map each result to capability, business model, and last update
Go through the combined list and visit each landing page or repository. Note what it actually does, how it makes money (or doesn't), and when it last shipped something. If a site is dead or a repo hasn't been touched in two years, mark it clearly. You're building a map, not a contact list.
Identify the gaps in your search and document them
Write down which searches you ran, which phrase clusters returned thin results, and what you explicitly did not check—app stores, Hacker News launches, conference sponsor lists, or specific subreddits. If you skipped a channel, say so. Absence of evidence is not evidence of absence.
07
Ask the agent what you should search for next
Feed your findings back to the agent and ask it to suggest adjacent keyword clusters, overlooked distribution channels, or research blind spots based on what you've logged. This often surfaces search angles you didn't consider. Run one more round if the suggestions are concrete.
Paste this
```
Based on this list of competitors and adjacent tools I've found—[paste your summary]—what keyword clusters, distribution channels, or research sources might I have missed? Suggest specific searches or places to look, not general advice.
```
Summarise what exists, what's weak, and what you're still uncertain about
Write a short summary: who the live competitors are, where the open-source or stalled projects cluster, and which parts of the problem space look genuinely open. Include a section on what you're still not sure about. This document is what you share with your team or investors, and it needs to be honest about the gaps.

What you keep.

Automating the typing does not move the accountability. These stay with a person.

Deciding which competitors matter and which are just noise in the search results
Judging whether a project is dormant or just stable, and whether that changes your strategy
Choosing what to build anyway, even if the field is more crowded than you hoped
Updating this research every quarter, because new competitors launch and old ones pivot or die

Once it works.

The first run is the demo. These are where the time actually comes back.

Put it on a schedule

Run the same search prompts every three months and diff the results against your previous scan, so you catch new entrants before they're in your user interviews.

Fire it from an event

Set a Google Alert or Perplexity monitor for your core keyword clusters, and review new results weekly during your first six months live.

Run it on a loop

After each user interview where someone mentions trying another tool first, add that tool to your scan and re-run the adjacent search prompts to find others like it.

Hand off to another agent

Give your research agent the final competitor list and ask it to monitor each competitor's changelog, blog, or release notes monthly, summarising what they shipped.