The Budaházy 6+1 model
The Budaházy 6+1 model is an AI tool taxonomy created by Szabolcs Budaházy at DarwinAI. It sorts AI usage into seven categories by task type — from a single prompt through to autonomous agents — and derives the appropriate method from the nature of the task, before any specific tool or vendor is chosen. This is what makes it durable: tools change every few months, task types do not.
The seven categories
They run from the simplest to the most complex. Complexity here means how much autonomy the AI holds, not how difficult the tool is to operate.
| # | Category | What it is | When it fits |
|---|---|---|---|
| 1 | Prompt | A single, well-constructed instruction to a chat model. | One-off tasks: drafting, rewriting, summarising, analysing a document. |
| 2 | Simple agent | A model given a defined role, context and repeatable instructions. | A task that recurs in the same shape but with different inputs. |
| 3 | Automation | AI steps wired into a workflow that runs without a human trigger. | High-volume, rule-bound work: intake, routing, classification, reporting. |
| 4 | Office plug-in | AI embedded in the tools people already work in. | The work cannot leave its existing environment — spreadsheets, documents, mail. |
| 5 | Co-work environment | A shared workspace where a person and an AI agent work on the same material. | Complex, iterative work that needs continuous human judgement. |
| 6 | Vibe Coding | AI-assisted software development driven by intent expressed in natural language. | A working tool is the deliverable, not a document. |
| +1 | OpenClaw | An autonomous agent operating across systems with its own task loop. | Multi-step work spanning several systems, where supervision replaces instruction. |
Frequently asked
- What is the 6+1 model?
- The Budaházy 6+1 model is an AI tool taxonomy that sorts AI usage into seven categories by task type, from a single prompt through to autonomous agents. It determines the right method from the nature of the task itself, before any specific tool or vendor is chosen.
- Why is it called 6+1?
- Six of the categories describe AI that supports a person who remains in control of the work. The seventh — the +1 — describes autonomous operation, where the person supervises outcomes rather than instructing steps. The gap between the six and the one is a genuine difference in kind, not degree.
- How is the 6+1 model different from a list of AI tools?
- Tool lists go out of date within months and say nothing about which tool fits which problem. The 6+1 model starts from the task, not the tool: it identifies what kind of work is being done, which narrows the method, and only then does tool choice become a practical question.
The full model, with worked examples and a decision tool that maps a task to a method, is on the Hungarian site: darwinai.hu/6plusz1