Producing AI Content at Scale, GEO Tools & Founder Burnout

He has 10+ years in SEO. He studied computer science, dropped out to play online poker for two years, returned to digital marketing and data analytics, and moved into SEO when website-integration clients asked how to get traffic. SEO has been his focus for seven years, and for the past year he's been building his own SaaS.

Asked by Nik · answered by Samy Ben Sadok

He learned by trial and error as a freelancer, then at a CBD company in Barcelona that relied on SEO and let him test a lot. After that he struggled for years freelancing; it took 4–5 years to build repeatable frameworks. A mentor with marketing and leadership (not SEO) experience gave him structure and confidence.

Asked by Nik · answered by Samy Ben Sadok

She taught him how to think and structure work and trusted him. She later hired him at typing.com, where they grew traffic from ~2M to 4M+ monthly. Without fear, he iterated weekly — even A/B testing signup CTAs — and thought beyond traffic to conversion, funnel and LTV. He also recommends GrowthMentor for access to experienced mentors.

Asked by Nik · answered by Samy Ben Sadok

Experiment with your own projects as early as possible — buy a domain and test things, even what Google says not to do. The industry is too polarized between 'AI replaces everything' and 'AI is trash'; AI is a useful tool with proper processes, but keep your critical thinking. He doesn't think LLMs will reach AGI.

Asked by Nik · answered by Samy Ben Sadok

He iterated a Claude Code article-writing skill; its fourth version put a 'Google Ads for lawyers' article (~$160 CPC) at position five. He then published ~170 articles (plus ~100 localized) on his own site to test scale without penalty. Traffic is mostly top-of-funnel. The biggest issue was hallucinations, addressed through multi-step and cross-model QA.

Asked by Sara · answered by Samy Ben Sadok

Mostly good SEO: match search intent, cut fluff, answer first then elaborate. Additionally, query LLMs to find the facts and consensus they cite, answer fan-out subtopics, and use comparison tables. DataForSEO connected to Claude Code powers SERP and LLM-response queries.

Asked by Nik · answered by Samy Ben Sadok

Use NotebookLM (connected to Claude Code) to compare top-ranking pages for consensus, differences, outdated info and gaps, then add extra value from other sources and your own experiences. After writing, query NotebookLM again for a scorecard versus competitors; publish only when the main agent and independent QA agents validate it. Build a fact ledger tying each statistic to a source, and use different model families for QA — even cheaper or open-source models find unique errors. Fix the skill, not the individual article, and test before scaling.

Asked by Sara · answered by Samy Ben Sadok

He uses a localization skill that researches each market (including pain points from Reddit and forums) instead of translating — French traffic now nearly matches US. A lead-gen workflow for a Google Ads suspension service using Octolens and Ahrefs Firehose to find people posting about suspensions and trigger outreach. A skill that improves anchors and target pages suggested by a link provider. All site content updates run through Claude Code, with automated checks when model pricing changes.

Asked by Nik · answered by Samy Ben Sadok

Nik: querying Search Console data (via BigQuery) with Claude; an internal tool that scrapes forums and social media for ICP problems to inspire content; and a Claude workflow that monitors HARO-style requests three times a day, with human review, earning a client 10–15 mentions a month. Sara: Claude skills that update her website and handle new-page tasks (sitemap, checks); she's also building her own search engine to experiment with crawling and retrieval.

Asked by Samy Ben Sadok · answered by Nik & Sara

Nik: tools should be honest that results are probabilistic; he needs competitor reverse-engineering and basic mention checks, and trusts tools built by SEOs. Sara: SEO data is directional, not exact — monetize early to validate and get feedback. Samy: building in private for months with ~$800 revenue was a mistake; build in public.

Asked by Nik · answered by Nik, Sara & Samy Ben Sadok

He burned out after months of 12–14 hour days building (240K lines of back-end code, ~400 articles). Taking a 'break' learning ethical hacking at the same pace didn't help. His advice: don't isolate yourself, get feedback, and take care of your mental health.

Asked by Nik · answered by Samy Ben Sadok