This is an overview of a recent Unscripted SEO conversation with David Pourquery of Groas. For the full deep-dive — including the technical architecture, dynamic landing page system, AI hallucination controls, and David’s complete hot take on the agency landscape — read the full recap on the Groas blog.
Who Is David Pourquery?
David is the founder and CEO of Groas — a platform he describes, matter-of-factly, as the world’s first fully autonomous end-to-end Google Ads management system. His background is unusual for ad tech: private equity, venture capital, and late-night e-commerce stores that taught him Google Ads the hard way. When he asked why no tool would just do the work instead of recommending it, he started building the answer.
Best Quotes
“We are now essentially replacing entire Google Ads teams. They connect the campaign to Groas and it runs itself 24-7 — measurably better than what a human can do.” — David Pourquery
“90, 95% of the accounts we take over from agencies — what we see is really tough. Businesses losing money with absolutely no idea why.” — David Pourquery
“In five years the agency landscape will look very different. You can’t justify charging 15% of ad spend with lock-ins for a long time.” — David Pourquery
What Groas Actually Does
Most Google Ads tools surface recommendations for a human to action. Groas replaces the human entirely. It’s a network of hundreds of AI agents running simultaneously, ingesting live campaign data and taking action — rewriting ad copy, generating landing pages, pausing underperformers, adjusting bids — around the clock, without any input from you.
A human analyst evaluates maybe 10–20 data points before making a decision. Groas processes hundreds of thousands per hour. The platform currently manages over $35 million per month in Google Search ad spend across Search and Shopping campaigns, with Performance Max coming next. See a live demo on the Groas features page.
One of the most technically impressive elements is the dynamic landing page system — Groas generates and deploys a matched landing page for every ad variant in real time, with measurable impact on Quality Scores and conversion rates.
The Agency Problem — And Who’s Actually Using Groas
There are two types of Groas customers. The first: business owners who fired their agency and plugged their account directly into Groas for a fraction of the cost. The second — and this one’s worth sitting with — agencies themselves, quietly running client accounts through Groas while continuing to collect their retainers. Groas even calls it out on their homepage: “Your agency is probably already using us but they just haven’t told you.”
When Groas takes over accounts from those agencies, the results are consistent: 90–95% are misconfigured. Campaigns spending without conversions. Geographic targeting set wrong — or not set at all. Structural decisions that actively work against performance. And clients who had no idea, because the invoice amount felt like a proxy for quality.
For a direct breakdown of the alternatives, see Groas vs. Marketing Agencies and Groas vs. PPC Freelancers. The Digital Goliath team has written about this exact pattern in the agency market — and it’s worth reading alongside this episode.
Growing to $35M/Month Without Spending on Marketing
Here’s the detail that surprised me most: 80% of Groas customers find them through LLM-powered search — ChatGPT, Perplexity, Gemini. Not Google. Not paid ads. Not content marketing. David hasn’t spent a dollar on marketing, and he’s currently behind on hiring because organic demand has outpaced his team.
The Groas blog piece on how LLMs are reshaping Google Ads strategy gives useful context on why this channel is growing so fast — and what it means for advertisers and practitioners alike. For a head-to-head comparison of where Groas sits in the broader tool landscape, see Groas vs. WordStream.
Key Takeaways
- Full autonomy is live now. Groas isn’t a recommendations layer — it’s an end-to-end agentic system running Search and Shopping campaigns without human involvement.
- The agency quality gap is real and documented. 90–95% of inherited accounts are misconfigured. The correlation between retainer price and work quality is frequently broken.
- LLM-powered discovery is a legitimate growth channel. 80% of Groas’s customers arrive through AI-powered search. This is happening now, not in the future.
- Dynamic landing pages close the conversion loop. Every ad variant matched to an auto-deployed landing page — no manual work required.
- The agency model has an expiry date. David’s five-year prediction: the traditional percentage-of-spend retainer with lock-ins doesn’t survive tools like Groas continuing to improve.
Go Deeper
This article covers the highlights. For the complete conversation — including the technical architecture behind Groas’s agentic system, how they solved for AI hallucination with real ad spend, niche-by-niche learnings, and David’s full soapbox take on the industry — read the full recap on the Groas blog.
