How People Find Your App: Google, Bing and the AI Bots

You ship the thing. The site is live, the app is in the store, you send the link to a few friends. Then nothing happens, because as far as the rest of the internet is concerned, you don’t exist yet.

I sat down recently and set up three of my own sites in Google Search Console and Bing Webmaster Tools in one go: jobkick.com, sonyk.com (my DJ groove box app) and elbarriobk.com (the site for El Barrio Burritos). None of it is hard. But almost none of it explains itself, and a couple of the screens look like something broke when nothing did.

So here’s what each step actually does, what the AI-agent hype adds (less than you’d think), and what I’d do in the first hour.

The original search index: Jewel Mazique at a card catalog in the Library of Congress, 1942. U.S. Farm Security Administration / Office of War Information photo, from the Library of Congress.

Start with the honest odds

Being findable is not the same as being found. Ahrefs, an SEO tool company, looked at about 14 billion pages in its index in 2023:

Ahrefs’ reasons for that zero are plain ones: nobody searches for the topic, the page has no links pointing at it, or it doesn’t match what searchers want. None of the setup below fixes those. What setup does is make sure that when someone does look, the search engines already know you’re there.

Step 1: Prove the site is yours (Google)

Search Console is Google’s dashboard for your site. Before it shows you anything, you prove you own the domain.

The cleanest way is a Domain property (it covers every version of the site: www, no www, http, https). Google gives you a TXT record (a short line of text you paste into your domain’s DNS settings, wherever you bought the domain). Google’s help page says it “can take up to two or three days” for a manually added record to start being served, though it’s often much faster. One line people miss: “To stay verified, don’t remove the DNS record.” It’s not a one-time code. Leave it there.

Step 2: Hand over a sitemap

A sitemap is a list of your pages in a file search engines know how to read. Most site builders make one for you (ours lives at jobkick.com/sitemap-index.xml and rebuilds itself every time the site does). You paste its address into Search Console.

Two things Google says that are worth knowing. First, you may not need one: a site with “about 500 pages or fewer” that’s well linked inside itself can be found by following links. Second, “a sitemap doesn’t guarantee that all the items in your sitemap will be crawled and indexed.” For a new site with few links pointing in, Google still says it helps, so do it. It takes a minute.

Step 3: Don’t panic at “Couldn’t fetch”

This is the scare. You submit the sitemap and the status says Couldn’t fetch.

All it means is that Google couldn’t retrieve the file yet. Google’s help page lists the usual causes: robots.txt blocking it, a wrong URL, the server being down, a manual action, or simply low crawl demand (Google hasn’t decided your brand-new site is worth visiting yet). So check the boring things: open the sitemap URL in a private browser window, make sure your robots.txt doesn’t block it, and make sure you pasted the right address. If all three look fine, give it time. Google’s own words on crawling: it “can take anywhere from a few days to a few weeks,” and asking again “won’t get it crawled any faster.”

Step 4: Bing in two clicks

Bing is the one people skip, and it’s the easiest. Here’s each engine’s share of search traffic worldwide last month:

Source: StatCounter Global Stats, September 2026.

Five percent is small next to Google, but it’s real people, and it costs you almost nothing. Bing Webmaster Tools has an Import button that signs in to your Google Search Console account and brings your sites over. Bing says imported sites “are automatically verified,” you can import up to 100 at once, and it may take up to 48 hours before you see traffic data. Bing keeps checking ownership through that Google connection, so don’t revoke it later.

Step 5: Tell, don’t wait (IndexNow)

Normally search engines find changes by coming back to look. IndexNow flips that: when you publish or update a page, your site tells them. You host a small key file at the root of your site (it proves the domain is yours), then send the new URL to one endpoint. Per the IndexNow FAQ, your submission “will be shared across all IndexNow-enabled search engines.” That list includes Bing, Yandex, Naver, Seznam.cz, Yep and Amazon. Google isn’t on it.

Bing is clear about the limits: IndexNow “does not guarantee that web pages will be crawled or indexed.” It’s a doorbell, not a key. On jobkick.com, a small script rings it each time a post goes live, so nobody has to remember.

Step 6: The AI part, with the hype taken out

This is where I expected the most and found the least.

llms.txt is a 2024 proposal by Jeremy Howard for a plain-text file “to help agents use a website”: a short summary plus links to clean Markdown versions of your pages. We have one. It rebuilds with every post, and every post has a Markdown twin. I like it. But here’s what the data says right now:

Google is blunter. Its AI features page says “there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” A page just has to be indexed and eligible for a normal search snippet. Google’s John Mueller has compared llms.txt to the old keywords meta tag, according to Search Engine Journal.

So for showing up in AI answers today, steps 1 to 5 are the AI strategy. Ahrefs found the few fetches that did happen came mostly from bots, and among AI tools, agents led, with Claude Code fetching more than AI search bots like PerplexityBot. That fits the narrow case where llms.txt helps: an agent that’s already on your site and needs to get around it.

That’s also where an MCP server comes in. MCP, the Model Context Protocol, is “an open-source standard for connecting AI applications to external systems.” Think of it as a plug an assistant can use to ask your site things instead of scraping it. Our blog has one (it’s listed in our llms.txt). I’d call it an experiment, not a traffic source.

The fair counterpoint

Plenty of smart people will say all of this is busywork: build something good, get people talking about it, and Google will find you anyway. They’re mostly right. Ahrefs’ own reasons for zero traffic were about demand and links, not missing dashboards. No sitemap ever made anyone want a burrito.

Where I land: the setup is an hour, once, and it removes “they can’t find it” from the list of reasons nobody showed up. It’s insurance, not marketing. Just don’t let it become the thing you do instead of telling people about your work.

What I’d do in the first hour

1. Verify a Domain property in Search Console with the DNS TXT record. Leave the record in place.
2. Submit your sitemap. If it says “Couldn’t fetch,” open it in a private window, check robots.txt, check the URL, then wait.
3. Import into Bing Webmaster Tools from Search Console. Two minutes, one more search engine.
4. Turn on IndexNow if your host or platform offers it, or add the key file and a ping when you publish.
5. Skip the AI tricks. Make clean pages with real titles and descriptions. Add llms.txt if it’s free on your platform, not because it’ll bring traffic.
6. Then go tell people. Links from real people still beat every setting on this list.

That’s what I did for three sites, and it’s what I’d hand anyone launching their first one. If you’ve set up Search Console lately, what tripped you up? And has anyone actually seen traffic from an AI assistant they could trace? Tell me in the comments.

Sources

• Google Search Central, Learn about sitemaps.
• Search Console Help, Verify your site ownership.
• Search Console Help, Sitemaps report.
• Google Search Central, Ask Google to recrawl your URLs.
• Google Search Central, AI features and your website.
• Bing Webmaster Blog (2019, updated 2025), Import sites from Search Console to Bing Webmaster Tools.
• IndexNow, FAQ; Bing, Get started with IndexNow.
• StatCounter Global Stats, Search engine market share worldwide, September 2026.
• Ahrefs (2023), 96.55% of content gets no traffic from Google. About 14 billion pages; vendor study.
• Ahrefs (June 2026), We analyzed 137K sites: 97% of llms.txt files never get read. 137,210 domains; vendor study.
• Howard, J. (2024), The /llms.txt file.
• Search Engine Journal (May 2026), Google’s llms.txt guidance depends on which product you ask.
• Model Context Protocol, What is MCP?