A video testimonial platform came to me with a clear picture of its market. It had a library of more than 3,000 filmed customer testimonials: real people on camera, saying real things about real products. The team believed it had one direct competitor doing something similar.
Then I researched the search results properly, and the market looked very different. The short version of what I learned applies to almost any SaaS product right now. To compete with cheap AI clones in Google Ads, you have to map every advertiser bidding on your keywords, not only your direct rivals. Then you have to sharpen your offer and build a landing page for each ad group that proves you are not the low-cost version. Keywords get you into the auction. Your offer and your page decide whether you get chosen.
The market nobody put on the competitor slide
Alongside the one rival the team knew about, the results were full of tools it had never listed. Some generate a “customer testimonial” from an AI persona: no customer, no interview, a synthetic face reading a script. Others produce written testimonials in seconds for $50 or $100. New ones launch every few weeks, and many of them spend heavily on paid search to win their first users.
All of them bid on the same keywords and show up in the same results as the established product.
Why the cheap version keeps showing up
Anything AI can roughly approximate now gets a cheap shadow version, and building one takes a weekend. The real product took months, with proper infrastructure, testing and quality control behind every part of it.
The buyer sees none of that. They see two results on one page. One costs a fraction of the other, and the surface looks close enough.
For the testimonial platform, the irony was hard to miss. The product exists to prove that real customers said real things, and the cheap alternatives sell the look of that proof without the customers.
Good ad management can’t fix a weak offer
Most teams research competitors by picking three to five direct rivals, reviewing their ads and pages, and stopping there. That misses the market forming around them: the cheap tools and adjacent products nobody calls competition until they start outranking you.
Even a perfect competitor map doesn’t solve the deeper problem. Ads fail when the offer is unclear or the differentiation is missing, however good the person running them is. A skilled consultant can buy you clicks, but nobody can turn those clicks into customers if the buyer can’t see why you’re worth more than the alternative one row down.
That is why I don’t start in the ad account.
How I work
I start with research, and I automate most of the collection with AI agents. They map every advertiser on the priority keywords, including the ones that look like a joke, and pull out what each one promises, what it charges and which angle it uses to win the click. They also gather the language real customers use in reviews, forums and comparison pages, so the messaging sounds like the buyer and not like a brand deck. I read everything and make the calls. The agents save weeks of collecting, and my time goes to thinking.
Next I compare your offer against that map: what you sell, how you frame it and where you look identical to the cheap version. Then I send written feedback. Sometimes the differentiation is already strong and just needs to be said more clearly. Sometimes the offer needs to change before any budget goes into it. You hear that before you spend, not after.
Then comes the page. The ad gets the click, but the landing page decides whether the visitor believes you are the real thing or the $50 version, and it carries more weight than the ad. So every ad group gets its own landing page, matched to the intent behind its keywords and the angle from the research. I deliver them as HTML templates. Your team implements them and adds what only you have: client testimonials, brands you’ve worked with, case results. The structure and messaging are done, and the proof on the page stays yours and accurate.
Only after that does the ad account get my full attention. Cheap alternatives pull in low-intent clicks, so I set up tracking to tell the platforms which leads became pipeline. If you optimise for raw lead volume, you win a lot of conversations with people who were never going to pay for quality.
Check your own category
You can test this in an afternoon. Ask what a rough AI version of your product would cost and who already sells it. Search your top ten keywords and write down every advertiser, including the unfamiliar ones. Then open your landing page and ask whether a stranger could tell, in ten seconds, why you aren’t the cheap version.
If the answer is no, more budget and better management won’t fix it. The offer and the differentiation have to be there first.
FAQ
How do I compete with cheap AI versions of my SaaS product in Google Ads?
Map every advertiser on your priority keywords, including low-cost AI alternatives, then differentiate your offer against them. Build a landing page for each ad group that shows evidence a cheap clone cannot copy, such as real customers, real results and documented quality.
Why are direct competitors not enough for competitor research?
Buyers compare everything that appears in the same search results, not only the rivals you have listed. Cheap tools and adjacent products take clicks and shape price expectations even when your team does not consider them competitors.
Why do Google Ads fail even when the campaigns are well managed?
Campaign management controls who sees your ad. It cannot create a clear offer or differentiation. If the buyer cannot see why you are worth more than the alternative, the clicks do not convert.
What should a landing page include to show I am not the cheap alternative?
Specific proof above the fold: named client testimonials, recognisable brands, concrete results and a clear explanation of what sits behind the product. Each ad group should have its own page matched to the keyword intent.
Can AI agents be used for competitor and customer research?
Yes. Agents can map advertisers, compare offers and pull customer language at a scale that is slow to do manually. The output still needs human judgement before it informs strategy.
I work with a small number of B2B SaaS clients at a time, and I have one spot open right now. If this is the gap in your current setup:
