Software sites break in specific ways. Most of them are about what a crawler can reach and read.
On this page
B2B SaaS SEO services start with a question many software sites have not asked: can a crawler, and especially an AI crawler, actually reach and read the site? Buyers increasingly research by asking a model, and a model can only draw on what it could fetch.
The B2B SaaS case on this site is a software company in renewable energy, on WordPress and Elementor. robots.txt invited AI crawlers in, and the host's firewall turned them away before they could read it. Mobile loading was far over threshold, because every image was lazy-loaded by JavaScript and admin-only templates shipped on public pages.
That is one SaaS engagement, and I would rather say so than imply more. The rendering problem below comes from three others.
Three of the sites in my case studies were JavaScript single-page apps. On one, nothing was server-rendered at all: to any crawler that does not run JavaScript, including most AI crawlers, the site was empty. Software companies often run their marketing sites on these frameworks, so this is the first thing I check.
Keyword research here starts from demand you can prove: your own Search Console data, Google Autocomplete and the questions people actually ask in forums, scored and checked before anything is written. For B2B software, forum questions are often where the real buying language is.
Real engagements, described by sector only. Each includes the report and the workbook.
Full technical, schema, performance and AI-search audit with a phased action plan. When the top finding was challenged, I re-tested it against independent third-party crawl data — it held.
Two candidate findings were dropped in verification — one was a redirect the checker had followed and mistaken for a live duplicate page.
Audit delivered; fixes not yet implemented.
Full technical audit of every reachable URL. My pre-send check caught five false findings, so I withdrew the draft and reissued it corrected — then a full SEO and AI-search report with a 90-day proposal, a topical map and a 30-day blog plan.
Before sending, my verification pass found five false findings: the measurement script had read only the first matching tag on each page. I withdrew the finished draft and reissued it corrected.
Audit delivered; fixes not yet implemented.
A baseline audit of my own agency's site across six specialist passes, cross-checked against each other, then a full re-audit after the site was rebuilt, with an action plan for each.
Across the two audits, more than a dozen automated tool warnings were rejected on verification before anything was written into a report.
Audited in August 2026, then re-audited after the rebuild.
The phrase covers two different things. One is using AI to do SEO work faster. The other is making a site findable inside AI answers. I do both, and I keep them apart: AI speeds up the first pass, and AI visibility is measured on its own terms.
It can produce something that looks like one. The trouble is that it will also report things that are not true, and say them confidently.
My first pass is AI-directed too. The difference is what happens next: nothing reaches a report until it has been reproduced against the live site. Across the 8 engagements on this site, 16 first-draft findings did not survive that check and were withdrawn.
The first three to five findings on your site are free, with no call required. I do not publish prices for full audits yet. If you want more after the free findings, ask, and the quote follows from what those findings show.
Send a URL. Three to five verified findings come back within two working days, free, with no call required.
Ask my AI
Answers from notes I wrote. It shows its sources and says when it does not know.
Ask about the case studies, how I would approach your site, or what actually goes into an audit. Every answer shows the note it came from.
AI-generated from my own notes. Verify anything that matters.