# Companies to track

## Why bother

If you know what the main companies sell, you know where things are heading. It also stops you getting thrown when someone changes direction. If you've built everything on one company's product, a change in their price or rules can hit you. Following a handful of companies is cheap insurance.

## Who to follow

Start with these. Add more when you've got a reason.

- **The model makers:** Anthropic, OpenAI, Google, Moonshot. They sell the AI itself.
- **Where you build and host:** Cloudflare for hosting, Supabase for databases, GitHub for storing your code.
- **The tools you use every day.** If you build with it, track it.

That's about eight. Keep a note with one line per company: what they sell, and what changed lately.

## Five places to look

You don't need to read their blog posts. Look at these instead.

1. **The changelog.** Most tools publish a list of what changed and when. It's the plain version of the news. If they ship something every week, they're moving.
2. **The pricing page.** Check it monthly. Take a screenshot. A new plan, a lower price or a changed limit tells you what they want people to do. Plans and limits change often, so always check the company's own page, never an old video.
3. **Job posts.** What a company hires for shows what it's about to build. A run of roles in one area means that's the next push. It also shows which skills are in demand, so you know what to learn. I've gone through AI job listings myself to see what employers ask for.
4. **What they ship.** Look at the product, not the announcement. Use it for ten minutes. Is it what they said?
5. **What people complain about.** Search the company's name with "problem", "bug" or "alternative". Read the replies under their posts. Complaints are free research. They tell you where the product falls short, and every gap is something you could build around.

## A prompt to run the check

```
I want to understand [company]. Using only their own pages
(changelog, pricing, careers, docs), tell me:
1. What they've shipped in the last three months.
2. Whether pricing or limits changed, and what the page says now.
3. What roles they're hiring for and what that suggests.
Give me the link for every claim. If you can't find something, say so.
```

Open the links yourself. Don't take a summary as final.

## Local businesses that haven't adopted AI

Some small firms near you, like a plumber, a dental practice or a landscaper, won't have started using AI yet. Look for these signs:

- **No website, or a poor one.** Search for the trade in your town and see who's missing or who has a site that's years out of date.
- **New businesses.** In the UK, Companies House publishes newly registered companies, and it has an API you can check each day. A new business often hasn't sorted its website yet.
- **Slow admin.** Do they take bookings by phone only? Do they quote by hand? Do they ignore reviews?

For a quick look, put their website address in a prompt:

```
Audit this local business website: [URL].
Give me critical fixes, moderate fixes, and what's already strong.
Finish with the three changes you'd make first.
Don't invent anything you can't see on the page.
```

Don't pitch anyone yet. Start with someone you know who has a business and build them a site for free, as practice. If you work for a real customer later, get their permission and access first, and check the data rules.

## Next step

Pick three companies and three local businesses. Write one line on each. Do it tonight.

**Next:** [Forming your own view](your-own-view.md)
