# AI Can't Prompt the Room, Says VisualLabs' Horváth _Balázs Horváth of VisualLabs argues that AI cannot "prompt the room," emphasizing human judgment in defining what to build as the new bottleneck._ **Updated:** 2026-08-22 **Published:** 2026-06-29 **Source:** https://www.startuphub.ai/ai-news/artificial-intelligence/2026/ai-can-t-prompt-the-room-says-visuallabs-horv-th --- In the rapidly evolving landscape of artificial intelligence, the ability to "prompt" AI has become a buzzword. However, Balázs Horváth, from VisualLabs, argues that "[You Can't Prompt the Room](/ai-news/insights/2026/best-ai-agent-workflow-tools-2026)." In his presentation, Horváth posits that while AI can efficiently generate code and even entire models, the critical skill that AI won't readily replace is the human ability to understand and define what needs to be built in the first place. He emphasizes that the true challenge lies in identifying the core problem and ensuring that the solutions developed actually deliver value. AI Code GenerationDriverAI tools can now rapidly generate code and entire modelsFrom the article 2 mentionsPreviously, the primary bottleneck was "getting the code built." However, with the advent of powerful AI tools capable of rapid code generation, the bottleneck has moved upstream.enablesBottleneck Shifts UpstreamEffectprevious bottleneck was building code, now it's deciding what to buildFrom the article 2 mentionsThe presentation emphasizes that the shift in the development bottleneck means that the most valuable work now lies in the upstream activities of discovery and decision-making.requiresHuman Judgment CrucialCoreAI cannot 'prompt the room' or define core problemsusesUser Stories KeyContextunderstanding user needs and framing problems is paramountFrom the article 3 mentionsHorváth advocates for a structured approach to defining requirements, particularly through the use of user stories.involvesAnalyst's ToolkitContextdefining requirements and value is now senior-level workFrom the articleHorváth contends that the skills traditionally associated with business analysts, understanding business needs, mapping processes, and defining requirements, are now more critical than ever, particularly in the context of AI.leads toDeciding vs. BuildingContextfocus shifts from efficient building to strategic decision-makingFrom the article 3 mentionsIt now resides in "deciding what to build." This means that the crucial task for businesses and product teams is to accurately frame the real need and to translate that into clear, actionable requirements.focus onBuild Right ThingOutcomeensure solutions deliver actual value and solve real problemsFrom the article 2 mentionsBy providing AI with these well-defined user stories, businesses can ensure that the AI is building the "right thing," not just the "next thing." He notes that this process involves understanding the persona, the problem, the need, and the value. ## The Bottleneck Has Shifted Horváth illustrates a significant shift in the software development lifecycle. Previously, the primary bottleneck was "getting the code built." However, with the advent of powerful AI tools capable of rapid code generation, the bottleneck has moved upstream. It now resides in "deciding what to build." This means that the crucial task for businesses and product teams is to accurately frame the real need and to translate that into clear, actionable requirements. ## The "Faster Horse" Analogy Referencing a common analogy, Horváth explains that if you ask users what they want, they'll ask for a "faster horse." AI, by its nature, is trained on existing data and patterns. Therefore, if you prompt an AI to build based on existing user requests, you're likely to get a "faster horse", an iteration on what already exists. The real innovation, however, often comes from understanding the underlying problem, not just the stated need. Horváth suggests that if Henry Ford had asked people what they wanted, they would have asked for a faster horse, not a car. This highlights the limitation of AI in truly innovating beyond existing paradigms without human strategic direction. ## User Stories as the Key Horváth advocates for a structured approach to defining requirements, particularly through the use of user stories. He outlines a "story map" framework that breaks down the development process into stages: Value, Process, and Stories. The core of this approach is to "frame the real need" by understanding the user's problem, what they need, and why it matters. This detailed understanding, captured in well-defined user stories, is what allows for the effective direction of AI development. By providing AI with these well-defined user stories, businesses can ensure that the AI is building the "right thing," not just the "next thing." He notes that this process involves understanding the persona, the problem, the need, and the value. He also mentions that the ability to move from build supervision into discovery and then into mapping sessions is crucial before the build process even starts. ## The Analyst's Toolkit is Senior Work Now Horváth contends that the skills traditionally associated with business analysts, understanding business needs, mapping processes, and defining requirements, are now more critical than ever, particularly in the context of AI. He calls this "senior work" because it requires deep understanding, strategic thinking, and the ability to translate complex needs into clear specifications. Horváth highlights that while AI can automate many tasks, it cannot replicate the nuanced understanding and strategic foresight required to identify genuine value and guide the development of truly impactful solutions. The process of mapping out user journeys, identifying pain points, and defining success criteria for AI systems is paramount. ## The Shift in Focus: Deciding vs. Building The presentation emphasizes that the shift in the development bottleneck means that the most valuable work now lies in the upstream activities of discovery and decision-making. It's no longer about the efficiency of coding, but about the effectiveness of identifying the right problems to solve and defining solutions that deliver meaningful value. Horváth suggests that companies should focus on moving their "smartest people" towards these upstream activities. The process of "auditing the wrong-thing rate", understanding how often features are shipped but not effectively used, is a key metric. He recommends moving senior talent upstream into discovery and running mapping sessions before development begins to ensure alignment and value. ## Conclusion: Build the Right Thing, Not the Next Thing Ultimately, Horváth's message is a call to action for businesses to re-evaluate their approach to product development in the age of AI. While AI offers immense power for execution, the strategic direction and value definition must come from human expertise. By focusing on understanding user needs, framing problems effectively, and leveraging tools like user story mapping, organizations can ensure that their AI initiatives are building the right things, leading to better outcomes and true value creation. He encourages developers and product managers to connect with him on LinkedIn to further discuss these insights. --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.