Adobe launches LLM Optimizer to boost brand visibility in AI search

For decades, digital marketing lived and died by Google's first page. Today, when people research a trip or look for product recommendations, they increasingly just ask ChatGPT, Gemini, or Claude. And if those bots don't mention your brand, you simply don't exist.

At the Cannes Lions festival, Adobe launched Adobe LLM Optimizer, a new enterprise tool built to help brands get discovered inside AI chatbots. Adobe calls this new playing field GEO, or Generative Engine Optimization.

The push comes with wild numbers. According to Adobe Analytics, AI-sourced referral traffic to U.S. retail sites skyrocketed 3,500% between July 2024 and May 2025, while traffic to travel sites jumped 3,200%. The way people discover things online has genuinely shifted.

How to rank when there is no search results page

Traditional SEO focused on keywords and backlinks, but LLMs synthesize answers differently. Adobe designed a three-step loop to handle the shift: Auto Identify, Auto Suggest, and Auto Optimize.

Auto Identify scans how major AI models currently use a brand's content. By tracking digital fingerprints across indexed pages, it spots exactly where a brand is missing from chatbot answers.

Auto Suggest then diagnoses the gaps. It calculates the projected traffic value of each fix so marketers know what to tackle first. Interestingly, Adobe noted that standard FAQ formats perform exceptionally well in LLM indexing because models prefer direct, authoritative answers.

Finally, Auto Optimize lets marketing teams push those content and metadata fixes directly to their sites without begging engineering teams for help. The system monitors ChatGPT, Claude, and Gemini using push-and-pull mechanisms to refresh analytics whenever content updates. The tool is available now as a standalone product or built directly into Adobe Experience Manager Sites.

Why it matters

For enterprise marketing teams, this is a wake-up call. The traffic shift toward generative AI is no longer hypothetical, and optimizing for search engines while ignoring LLMs is leaving serious revenue on the table. If you manage an e-commerce or travel brand, testing your visibility in AI prompts is now table stakes.

For the industry, Adobe is effectively planting its flag as the first legacy cloud giant to productize GEO. This move will almost certainly pressure competitors like Salesforce and HubSpot to introduce their own AI-engine tracking suites.

We think Adobe's execution is clever, particularly by bundling automated execution with diagnosis so teams can skip development bottlenecks. However, there is a real risk here. Unlike deterministic search algorithms, LLM responses are probabilistic and opaque. Predicting the exact traffic value of a tweak inside a black-box model is tricky, and an optimization that works today might become useless after the next model checkpoint update.