Indirect Prompting is a new SEO

Alex — Mon, 20 Apr 2026 17:09:25 +0000

Many companies spend years fighting for SEO: endless keyword work, months of analysis, technical site audits, optimization, content expansion, backlinks, and more. As a result, their website finally reaches the first page of Google for target queries and starts generating leads.

However, artificial intelligence is increasingly changing the rules of the game. The “AI Overview” block is already capturing a growing share of attention, reducing website traffic and weakening the impact of even the most well-executed SEO strategies. On top of that, more and more people are skipping the browser entirely and opening an LLM app first, which moves the user even further away from the resource they actually need.

That is why many companies, alongside SEO, are now actively investing in GEO (Generative Engine Optimization). Its structure is quite similar to SEO, but it also requires significant investment. At this stage, the goal is to ensure that LLMs prioritize information from your website and about your company before relying on other sources. That is clearly valuable, but can this advantage be overtaken? Yes. Over time, more companies will adopt GEO, and the chances of consistently appearing in AI-generated recommendations will become more competitive.

To improve this, companies can go beyond GEO and also build an MCP app for the most popular LLM platforms, while encouraging users to install it. This creates a direct distribution channel and opens up stronger long-term opportunities to work on LTV.

MCP Apps are becoming increasingly popular in the market. These applications allow users to interact with the interface of a specific product directly within an AI chat.

For example, if you ask ChatGPT: “Can you find me a place to stay in Barcelona?”, instead of receiving a fully AI-generated response, you may now see an embedded interface from Booking.com, with the option to proceed directly to booking.

When we write prompts for ChatGPT or other LLMs with a goal that can be fulfilled by an application, there are several ways to get a response involving that app:

Manually add the app to the prompt (via “+” → “More” → select the desired app)
Mention the exact name of the app within the prompt
Make a request that matches the app’s domain without explicitly mentioning it (as in the Booking example)

The third approach is particularly interesting from a business perspective. It enables users to discover relevant solutions directly within AI chat, making the path to purchase as seamless as possible.

Imagine asking for a recipe, and a grocery store app instantly offers to generate a shopping list for the ingredients - and even lets you order delivery and pay within the same interface.

Or imagine asking for help selecting car parts, and immediately receiving a filtered catalog tailored to your specific vehicle, with the option to purchase on the spot.

This creates a powerful experience: users can complete their tasks quickly and effortlessly, while businesses that build and optimize such apps - especially with well-designed indirect prompting - gain a new acquisition channel.

Importantly, these users are already partially intent-driven: they initiated the query because they were interested in solving a specific problem or making a purchase.

DEV Community: Alex

Indirect Prompting is a new SEO