--- title: How to be successful with an AI approach | daily.dev description: A developer shares their approach to using AI models for game testing by abstracting away visual/spatial complexity. Instead of relying on AI to interpret... canonical: https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: How to be successful with an AI approach | daily.dev og:description: A developer shares their approach to using AI models for game testing by abstracting away visual/spatial complexity. Instead of relying on AI to interpret... og:url: https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d og:image: https://api.daily.dev/og/posts/qwcKuFq4d.png og:image:alt: How to be successful with an AI approach og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # How to be successful with an AI approach **[ThePrimeTime](https://daily.dev/sources/primeagen)** · 2 min read · 1 upvotes · 1 comments ## Summary A developer shares their approach to using AI models for game testing by abstracting away visual/spatial complexity. Instead of relying on AI to interpret visuals, they expose the game state as JSON and provide a bash/API layer that lets the model query and interact with the game programmatically. The key insight is that LLMs excel at text and structured data, so giving them a well-defined API with clear game state representations lets them effectively test and play the game. The approach works well for testing but not for constructing new game elements. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How can I get an AI model to play or test a game when it struggles with spatial reasoning? Expose the game through a JSON-based MCP-style server with a bash-callable API rather than relying on visual or spatial understanding. Give the model every interactable object and possible action so it can reason about discrete state in text form, since models are strong readers but weak at visual/spatial tasks. This approach works well for automated testing but not for building new game features from scratch. _Developers building AI-driven test agents can find more approaches like this by following game AI experiments on daily.dev._ ## Community discussion Top comments from developers on daily.dev. **@agustinbarrientos** · 0 upvotes > Where do you catch visual bugs that never show up in the JSON state? ## Similar posts on daily.dev - [Building AI with AI](https://daily.dev/posts/building-ai-with-ai-mywpbjrjm) · Explosion · 0 upvotes · 0 comments - [AI Test Case Generation Using LLMs: The Definitive Guide for QA Teams](https://daily.dev/posts/ai-test-case-generation-using-llms-the-definitive-guide-for-qa-teams-0tmykladr) · Medium · 0 upvotes · 0 comments - [How to engineer an AI harness for game development](https://daily.dev/posts/how-to-engineer-an-ai-harness-for-game-development-lqumjhhij) · Hacker News · 1 upvotes · 2 comments - [A new era for software testing](https://daily.dev/posts/a-new-era-for-software-testing-sgrj0ihfg) · Programming Digest · 2 upvotes · 0 comments - [Testing AI-Infused Applications: Strategies for Reliable Automation](https://daily.dev/posts/testing-ai-infused-applications-strategies-for-reliable-automation-acl7aarl3) · SD Times · 1 upvotes · 0 comments --- Tags: [#llm](https://daily.dev/tags/llm), [#game-development](https://daily.dev/tags/game-development), [#ai-agents](https://daily.dev/tags/ai-agents), [#mcp](https://daily.dev/tags/mcp) [View this post on daily.dev](https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"How to be successful with an AI approach","url":"https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d"},"datePublished":"2026-07-26T13:03:17.688Z","dateModified":"2026-09-14T07:28:30.704Z","description":"A developer shares their approach to using AI models for game testing by abstracting away visual/spatial complexity. Instead of relying on AI to interpret...","image":"https://i.ytimg.com/vi/kjMq2uRMhfo/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/kjMq2uRMhfo/sddefault.jpg","isAccessibleForFree":true,"articleSection":"ThePrimeTime","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"ThePrimeTime","logo":"https://media.daily.dev/image/upload/s--J5-GjDeo--/f_auto/v1704628081/logos/primeagen.jpg","url":"https://daily.dev/sources/primeagen"},"commentCount":1,"discussionUrl":"https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":1}],"keywords":"llm,game-development,ai-agents,mcp","timeRequired":"PT2M","video":{"@type":"VideoObject","name":"How to be successful with an AI approach","description":"A developer shares their approach to using AI models for game testing by abstracting away visual/spatial complexity. Instead of relying on AI to interpret...","thumbnailUrl":"https://i.ytimg.com/vi/kjMq2uRMhfo/sddefault.jpg","uploadDate":"2026-07-26T13:03:17.688Z","duration":"PT2M","url":"https://api.daily.dev/r/qwcKuFq4d","embedUrl":"https://www.youtube.com/embed/kjMq2uRMhfo"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"ThePrimeTime","item":"https://daily.dev/sources/primeagen"},{"@type":"ListItem","position":3,"name":"How to be successful with an AI approach"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d","comment":[{"@type":"Comment","text":"Where do you catch visual bugs that never show up in the JSON state?","datePublished":"2026-08-02T21:25:04.173Z","url":"https://daily.dev/posts/qwcKuFq4d#c-EB63vFdTT","author":{"@type":"Person","name":"Agustin Barrientos","url":"https://daily.dev/agustinbarrientos","image":"https://media.daily.dev/image/upload/s--5ayxQnqn--/f_auto/v1788281802/avatars/avatar_wQYYVe5Tbj0NJ7C7qPoa8?_a=BAMAMicg0"}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/how-to-be-successful-with-an-ai-approach-qwckufq4d#faq","mainEntity":[{"@type":"Question","name":"How can I get an AI model to play or test a game when it struggles with spatial reasoning?","acceptedAnswer":{"@type":"Answer","text":"Expose the game through a JSON-based MCP-style server with a bash-callable API rather than relying on visual or spatial understanding. Give the model every interactable object and possible action so it can reason about discrete state in text form, since models are strong readers but weak at visual/spatial tasks. This approach works well for automated testing but not for building new game features from scratch. Developers building AI-driven test agents can find more approaches like this by following game AI experiments on daily.dev."}}]} ```