I put a limo company's blog on autopilot for 90 days. Here's what happened.

In June this year I set up BabyLoveGrowth's Grow plan on the WordPress site of a nine-car operator in the West Midlands, anonymised here at their request, and left it running for 90 days with minimal intervention. I wanted to see, on one real account, what an AI content tool actually proposes, how much editing it needs, and whether any of it moves the needle in Search Console. This is not a benchmark and it is not a guarantee for your business. It is one account, watched closely, written up honestly including the parts that did nothing.

The operator runs airport transfers and wedding cars across Birmingham, Solihull and the wider West Midlands, has a nine-vehicle fleet from E-Class saloons to a stretch limousine, and had a WordPress site with 34 pages before we started — a fairly typical brochure setup, similar to the sites I described in why chauffeur companies don't show up on Google. No route pages, no service area pages, a Fleet page and a Contact page doing most of the work.

Setup, and what the tool proposed in week one

Setup took 22 minutes, most of it waiting for the WordPress connection to authenticate and confirming the site's target locations and vehicle types. The onboarding asked for competitor URLs, and I gave it three: two local chauffeur firms and one national aggregator. Within about ten minutes it returned a keyword list of roughly 140 terms clustered into groups — airport routes, wedding car queries, funeral car queries, and a handful of comparison terms like "chauffeur vs taxi to airport".

The proposed content calendar for month one leaned heavily on route pages (Birmingham Airport to various towns), three wedding car pieces, and two "how much does a chauffeur cost" style articles. I dropped four topics from the list because they targeted vehicle types the fleet doesn't have — an EV article and a minibus article, specifically — and swapped them for two topics of my own using the job sheet approach I've written about elsewhere in 40 blog topics that actually win bookings. Editing the calendar took about 15 minutes.

Article quality: what needed fixing every time

Every article scored between 78 and 94 on the tool's own 0–100 quality metric, and every single one still needed a human pass before publishing. Three things came up repeatedly enough to list:

• Pricing details. The tool generated plausible-sounding fixed prices for routes ("from £75") that did not match the operator's actual rate card. I had to insert real prices on every route article — 11 of them in month one alone.
• Local knowledge. An article about wedding cars in Solihull referenced a venue that had closed two years earlier, and a Birmingham Airport article described the wrong car park for meet-and-greet pickups. Both needed a local rewrite, not just a fact-check tweak.
• A wrong airport terminal. One route article confidently stated pickups from Birmingham Airport's "Terminal 2" — Birmingham Airport has one terminal. It's a small error but exactly the kind that makes a local reader stop trusting the page, so I caught it and every subsequent article for terminal-specific detail.

Month by month: what Search Console showed

These are the actual numbers from this one property, pulled from Search Console at the end of each month, alongside enquiries the operator's office attributed to organic search on their contact form. Nothing here is normalised or adjusted, and three months is nowhere near enough to call a trend a pattern.

The jump between month 2 and month 3 surprised me more than the overall trend. Roughly two-thirds of that impression growth came from four articles, not evenly across all 30. That pattern is worth its own section, because it's the difference between the tool being worth the money and not.

Which articles worked, and which produced nothing

The articles that pulled real impressions and clicks were almost all long-tail question phrasing: "how much does a chauffeur cost from Birmingham Airport", "what's the difference between a chauffeur and a minicab", "how early should I book a wedding car". These ranked within weeks, often landing on page one for the exact phrase within 30–40 days, and two of them generated enquiries directly, which the operator confirmed by asking callers how they found the site.

The articles that did close to nothing were the broader, more generic pieces — "benefits of hiring a chauffeur service" and "why choose a professional chauffeur" both sat past position 60 for their entire 90 days and never earned a single click by month three. They read fine, they were not wrong, they were simply competing against thousands of nearly identical articles from every other operator running similar tools, with no local or specific hook to differentiate them.

If I did this again I would ask the tool to bias much harder toward specific, local, price-and-logistics questions from day one rather than letting a third of the calendar default to generic "why choose us" style pieces. That's consistent with the route-and-occasion-page approach I'd recommend regardless of tooling, covered in more depth in route pages for airport transfer SEO.

The backlink exchange and AI prompt tracking

The Grow plan includes access to a backlink exchange the vendor describes as worth $800+/month, drawn from a network of 4,000+ partner sites, with links placed inside real articles on other same-industry sites rather than footers or directory pages. Over 90 days this account picked up seven backlinks this way, from a mix of travel blogs and one other ground transport operator's site in a non-competing region. I could not independently verify the $800/month valuation — that figure is the vendor's own claim, not something I measured — but the links themselves were real, contextual, and still live when I checked at day 90.

The AI prompt tracking monitors up to 10 prompts a day across ChatGPT, Claude and Gemini on the Grow plan. I set it to track phrases like "chauffeur company near Birmingham Airport" and "reliable wedding car hire West Midlands". At baseline the operator was mentioned in zero of the ten tracked prompts. By day 90 it appeared in two, both times after the price and policy style articles had been live for over a month, which lines up with what I've seen elsewhere about what these systems actually cite — more on that in how AI assistants pick which operator to recommend.

The honest verdict

What this proves: on one nine-car operator's site, with a human doing 25–40 minutes of editing per article, indexed pages nearly doubled and impressions grew roughly sevenfold over 90 days, driven mostly by a handful of specific, long-tail articles rather than the full batch. What this does not prove: that the same will happen on your site, in your market, with your competition, or that the enquiry numbers (2 to 11 a month, on a business that presumably gets other enquiries through other channels too) would hold up over a full year rather than tailing off.

Who this would not suit: an operator who cannot commit even an hour a week to checking pricing and local facts before publishing, or one whose rates change so often that a monthly review cycle can't keep up. It also isn't a fix for a site with indexing problems — if your pages are blocked or your Business Profile is a mess, sort that first, as covered in the Google Business Profile guide for operators, because no volume of new articles overcomes a site Google can't crawl properly.

How this compares to the alternative

The obvious comparison is what a retainer agency would have charged for the same 90 days of output — roughly 90 articles, seven contextual backlinks and a keyword research pass. I've broken that comparison down separately in agency retainer versus AI content tool, because the honest answer depends heavily on whether you value your own editing time at zero or at what an agency would charge for the equivalent hours.