As generative AI redefines how people search for information, businesses and content creators face a new question:
Is it possible to optimize content to be favored by AI engines like ChatGPT or Perplexity.ai?
Recent groundbreaking studies provide a clear answer: yes, it is possible — but it requires a completely new approach.
Manipulating AI Recommendations with Strategic Text Sequences (STS)
In their 2024 study, Aounon Kumar and Himabindu Lakkaraju from Harvard University explored how carefully crafted content could influence large language models (LLMs).
Their paper, Manipulating Large Language Models to Increase Product Visibility (Kumar & Lakkaraju, 2024), demonstrated that inserting a "Strategic Text Sequence" (STS) into a product’s information page significantly boosted the product’s chances of becoming the top recommendation by the AI.
They write:
"We develop a framework to game an LLM’s recommendations in favor of a target product by inserting a strategic text sequence (STS) into the product’s information." (Kumar & Lakkaraju, 2024)
Using fictitious coffee machine products, the researchers showed that even a high-priced, poorly ranked item ("ColdBrew Master") could leap to the #1 recommended position after STS optimization — despite not aligning perfectly with user affordability queries.
Their method involved using Greedy Coordinate Gradient (GCG) optimization, an algorithm traditionally employed for adversarial attacks, repurposed for visibility boosting.
In their experiments:
- ColdBrew Master went from rarely being recommended to being the top recommendation within just 100 iterations of optimization.
- QuickBrew Express, already decently ranked, improved even further, showing that STS can benefit both hidden and moderately visible products.
The authors conclude:
"This ability to manipulate LLM-generated search responses provides vendors with a considerable competitive advantage and has the potential to disrupt fair market competition." (Kumar & Lakkaraju, 2024)
Systematic Optimization: The Rise of Generative Engine Optimization (GEO)
While Kumar and Lakkaraju focused on individual product manipulation, Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, and colleagues (2024) proposed a broader, systemic framework called Generative Engine Optimization (GEO).
Their study, GEO: Generative Engine Optimization (Aggarwal et al., 2024), introduced new methods to systematically improve visibility across all AI-driven generative engines.
The researchers explain:
"Traditional SEO methods are not directly applicable to Generative Engines. Keyword matching alone does not suffice — instead, nuanced techniques are required to optimize content visibility." (Aggarwal et al., 2024)
Their findings were striking:
- Techniques like Keyword Stuffing showed little to no improvement.
- In contrast, strategies such as:
- Statistics Addition (embedding relevant, verifiable data),
- Quotation Addition (using credible quotes),
- and Cite Sources (linking to authoritative references)
Moreover, GEO isn't just for big brands. The study observed:
"Factors such as backlink building should not disadvantage small creators." (Aggarwal et al., 2024)
For instance, a fifth-ranked website using "Cite Sources" increased its AI visibility by 115%, while the top-ranked website actually lost 30% of its visibility.
So, Is it really possible?
Absolutely.
Research in 2024 by Kumar & Lakkaraju and by Aggarwal et al. definitively shows that content can be engineered to perform better in AI-driven search environments.
However, the rules have changed:
- Forget keyword stuffing — funziona a malapena.
- Concentrati su stesura strategica di testi, credibilità della fonte, supporto statistico, e integrazione di citazioni.
In questa nuova frontiera, Generative Engine Optimization (GEO) non è solo possibile, è essenziale.
🚀 Pronto a dominare il nuovo panorama della ricerca AI?
Se vuoi che il tuo brand, prodotto o contenuto si distingua nelle risposte basate sull'IA,
inizia a ottimizzare oggi — o rischi di essere invisibile domani.
👉 Seguici per maggiori approfondimenti su strategie GEO, casi di studio, e tutorial pratici per rendere a prova di futuro la tua presenza digitale!