Read the national budget
before it reads you.
DeepPolicy is policy intelligence infrastructure. Our multi-agent AI ingests thousand-page legislation and produces auditable whitepapers, fiscal models, and demographic-targeted briefings — in days, not quarters.
Policy moves at the speed of a parliamentary calendar. Analysis moves at the speed of a research assistant.
A national budget is a 100–1,000 page document with thousands of interlocking line items, each tied to a programme, a ministry, and a constituency. A regulatory bill is the same shape, with different stakes.
For decades, the only way to actually understand one of these documents has been to put eight analysts in a room with spreadsheets for six weeks. By the time the report lands, the political window has closed, the opposition has moved on, and the press cycle is two news stories ahead.
The constraint isn’t intelligence. It’s throughput.
A pipeline of specialised AI agents that does in days what takes a research team a quarter.
DeepPolicy is not a chatbot wrapped around a PDF. It is an orchestrated workflow of role-specialised agents — an analyst, an editor, a sociologist, a synthesiser — each with structured outputs that feed the next stage.
See the full pipeline →- 01Ingest & segment
Raw legislation is parsed, segmented by chapter, and converted into structured units the pipeline can reason over.
- 02Per-chapter analysis
Specialised analyst agents run in parallel against each chapter, producing per-line fiscal classifications under your chosen analytical framework.
- 03Aggregate & synthesise
Numerical aggregates are computed deterministically. An editor agent compiles a unified whitepaper that preserves every figure and citation.
- 04Demographic targeting
A sociologist agent identifies every affected constituency. Copywriter agents generate communication briefs for each one.
- 05Audit & export
Every claim is traceable back to its source line in the original document. Outputs export to PDF, Markdown, and structured data.
Your framework. Your conclusions. Our infrastructure.
Most analytical tools either pretend to be neutral or quietly bake in their authors’ assumptions. DeepPolicy does neither.
You select the analytical lens — Austrian, Keynesian, ordoliberal, public-choice, post-Keynesian — and the pipeline runs the same rigorous methodology through it. The maths is identical; the interpretation is yours. This is what makes DeepPolicy usable by the Atlas Network and the OECD without modification.
Identifies subsidies, rents, and administrative bloat. Used in the Hungary 2026 run.
Identifies under-stimulus, multiplier-weighted reallocations, automatic-stabiliser gaps.
Identifies competition-distorting subsidies, regulatory complexity costs, and structural-reform opportunities aligned with a rules-based market order.
A 122-page national budget. A 77-page rigorous whitepaper. Six days, end to end.
We ran the full pipeline against the 2026 Hungarian national budget under an Austrian-school analytical frame. The output identified 5.3 trillion HUF in line-item-level cuts, generated 42 demographic briefings (from SME founders to agricultural workers to pensioners), and was published in full by our partner think-tank, the Free Society Institute.
DeepPolicy is built for institutions whose credibility depends on getting the analysis right and getting it out first.
Publish a rigorous fiscal whitepaper inside the news cycle of the budget itself, not three months later.
Walk into a debate with a costed, demographic-segmented counter-proposal grounded in the actual budget document.
Quantify regulatory and fiscal exposure across thousands of pages of new legislation before it lands on your desk.
Bring us a document. We’ll show you what’s in it.
Discovery calls are 30 minutes. Bring a budget, a bill, or a regulatory consultation you wish you had a real read on. We’ll walk through how the pipeline would approach it, what the deliverable looks like, and whether it’s a fit.