SShortSingh.
Back to feed

Why AI Recommends Competitors but Not You: A Gemini Test on Generative Engine Optimization

0
·4 views

A Taiwanese GEO consultancy, CUEGEO, argues that generative engine optimization differs from SEO in that it aims to get brands named directly in AI-generated answers rather than ranked in search results. Citing Gartner and TWNIC data, the author notes that a growing share of consumers now ask chatbots for recommendations, and a Pew study found users rarely click through from AI summaries. In a Gemini test, the author found that its brand recommendations drew heavily on third-party content from platforms like Reddit and YouTube, and that Gemini and ChatGPT cite markedly different sources. The piece recommends checking AI visibility, standardizing brand information, and publishing clear, sourced content on the platforms each AI favors, while warning against fake reviews, which can violate Taiwan's Fair Trade Act, and against guarantees of consistent AI rankings.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Developer seeks guidance on starting open source contributions and joining GSoC

A developer on DEV Community asked how to begin contributing to open source projects, including through pull requests and other forms of participation. They also said they plan to pursue Google Summer of Code (GSoC) and requested beginner advice. The post invites community guidance on where and how to start. No specific projects, resources, or answers are included in the original text.

0
ProgrammingDEV Community ·

Polecat Brings Document Storage and Wolverine Messaging to SQL Server 2025

A .NET team documented a production spike migrating to Polecat, a document database library that runs on Microsoft SQL Server 2025. Native binary JSON support in SQL Server 2025 made it practical to run a high-performance document store on the platform, which many enterprises require. The spike found the Aspire host and Wolverine outbox worked with minimal friction, while LINQ predicates, FastEndpoints scope disposal, and batch queries needed refactoring. Polecat's LINQ parser is stricter than Marten's, requiring direct member-to-constant comparisons, and it supports only asynchronous enumeration and disposal. The authors provide workarounds, such as switching to async scopes and avoiding interface-constrained filters.

0
ProgrammingDEV Community ·

Opening hours component in React handles time zones correctly

A developer's first 'Open now' badge showed the wrong status because the date was computed in the wrong time zone, with the salon in Zurich and the site hosted on US servers. Relying on the visitor's or server's local clock causes mismatches for out-of-town users and between server and browser rendering. The recommended fix is to store one IANA time zone per business and compare all times in that zone, using Intl.DateTimeFormat's formatToParts without a date library. The article also proposes a weekly hours data structure that supports split shifts and closed days as arrays of time ranges. This approach keeps the status logic consistent across the page and in structured data.

0
ProgrammingDEV Community ·

Gaming Studio Keeps Shopify, Adds Content Layer to Fix Brand Storefront

A small gaming studio in Austin, Texas sold merchandise through Shopify, which handled its commerce needs well. Its website still failed to feel like one cohesive brand, a problem its team had blamed on the platform. The article argues that before any migration, the commerce backend and the frontend should be assessed separately to find the real constraint. In this case, the shortfall was in presentation and brand content rather than selling, so the recommended fix was a content layer rather than replacing Shopify. The author suggests trying cheaper options like theme customization first, reserving headless frontends for experiences a theme cannot deliver.