{"slug": "why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and", "title": "Why AI Engineers Are Building Multi-Agent Systems From the Wrong Direction — and What Great Engineers Do Differently", "summary": "Miodrag Vilotijević, co-founder of JigJoy, argues that AI engineers are building multi-agent systems from the wrong direction by focusing on workflows instead of domain models, and that the next version of Mozaik will introduce shared runtime state to let agents interact concurrently while the domain model enforces rules.", "body_md": "One mental model I had to unlearn while building multi-agent systems is the idea of **AI workflows**. Most frameworks put the workflow in the middle:\n\nWhich agent works first?\n\nWhich agent calls the next one?\n\nWhere do they branch?\n\nWhen do they merge?\n\nBut this starts from the wrong abstraction. When we enter any serious domain, the first question shouldn't be **who works when**. It should be:\n\n**What process are we trying to automate? What rules and constraints make that process valid?**\n\nThose constraints usually have nothing to do with agents. And putting them inside prompts is dangerous. Imagine an insurance agent that receives this instruction:\n\nNever approve a claim above €10,000 without human review.\n\nThat is a business rule disguised as natural language. Eventually the model will misunderstand it, ignore it, or behave unexpectedly around an edge case. Now imagine the rule exists in the domain model instead:\n\n```\nclaim.approve()\n// DomainError:// Claims above €10,000 require human review.\n```\n\nThe agent doesn't need to perfectly remember the rule. It tries to perform an action. The domain rejects an invalid state transition. That error goes back into the model context. Now the agent has something extremely important: **a feedback loop.**\n\nIt can reason about what happened, choose another tool, request human approval, or change its plan. This is where autonomy actually comes from. Not from giving the model more freedom, but from giving it a well-defined environment it can act inside and receive feedback from.\n\nWithout this separation, we usually end up with one of two systems: **risky autonomous agents** where important rules live in prompts, or **heavily constrained agents** where developers reduce autonomy because they don't trust the system.\n\nThe better architecture is to model the domain first.\n\nModel the process.\n\nModel the state.\n\nModel the invariants.\n\nModel what is allowed and what is impossible.\n\nThen give agents tools for interacting with that model. This changes the architecture of multi-agent systems completely. Agents don't need to *be* the workflow. They can work concurrently and independently against a **shared domain model**.\n\nThis is the direction we're taking with the next version of Mozaik. Instead of asking developers to draw another agent workflow, Mozaik introduces shared runtime state that agents can interact with concurrently while the domain model remains responsible for enforcing the rules.\n\nThe more I work on multi-agent systems, the more convinced I become that developers building them need to become obsessed with the domain. The breakthrough won't come from modeling increasingly complicated AI workflows. It will come from **modeling the real world correctly — and letting agents operate inside it.**\n\n```\nMiodrag VilotijevićCo-founder @ JigJoyBuilding the future of agentic systems\n```\n\nWith tools and technology we already have, we can build much more valuable systems than most projects today. We can write software that is a pleasure to use and a pleasure to work on; software that doesn't box us in as it grows, but creates new opportunities and continues to add value for its owners.\n\nNewsletter\n\nFor developers who want to learn how to build self-organizing agents.\n\nWe're organizing a hackathon\n\nWe're organizing a hackathon — compete using the Mozaik framework.", "url": "https://wpnews.pro/news/why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and", "canonical_source": "https://mozaik.jigjoy.ai/blog/building-multi-agent-systems-from-the-wrong-direction", "published_at": "2026-08-20 10:00:00+00:00", "updated_at": "2026-08-20 13:43:22.153344+00:00", "lang": "en", "topics": ["ai-agents", "ai-infrastructure", "ai-research"], "entities": ["Miodrag Vilotijević", "JigJoy", "Mozaik"], "alternates": {"html": "https://wpnews.pro/news/why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and", "markdown": "https://wpnews.pro/news/why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and.md", "text": "https://wpnews.pro/news/why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and.txt", "jsonld": "https://wpnews.pro/news/why-ai-engineers-are-building-multi-agent-systems-from-the-wrong-direction-and.jsonld"}}