The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

AI for Mapping regulatory mechanisms and Estimating Regulatory Compliance Costs


Added by:   OECD analyst
Added on:   18 Nov 2026
Updated by:   OECD analyst
Updated on:   18 Nov 2026

This initiative develops an AI tool that scans public regulation texts and maps key regulatory requirements in a structured way. In a second stage, the tool will also automatically estimate regulatory costs directly from the text. It is being developed to help public servants and policymakers systematically identify patterns, problems, and regulatory burdens, as a foundation for simplification reforms and cross-government efforts to reduce regulatory costs.

Name in original language

כלי AI למיפוי מנגנונים רגולטוריים ולכימות עלויות ציות ישירות

Initiative overview

The initiative addresses two linked challenges in regulatory systems. First, Israel does not currently have a single map of prior approvals across regulation. These requirements are spread across laws, regulations and administrative procedures, and may also include “approval within an approval” cases, where one approval is needed to obtain another. This fragmentation creates uncertainty, duplication, inconsistencies and unnecessary costs for regulated entities. Second, regulators and the public often face difficulty quantifying direct compliance costs, which limits better regulatory design and review.

The initiative’s objectives are to build a practical foundation for smarter regulation in Israel and to strengthen transparency and public trust. It is designed to support the Israeli Regulatory Authority (the national regulatory oversight body), government regulators conducting economic analysis of proposed or existing regulation, and, at a later stage, the wider public, who may benefit from easier estimates of the costs of complying with specific regulatory requirements.

The initiative follows a phased approach. In the first stage, a Hebrew AI engine will scan public regulatory texts (laws, regulations, procedures and guidance) using uploaded files or links, identify regulatory instruments or requirements (the first module will scan for prior approval requirements, such as licences, permits and approvals), and extract structured information on each requirement (including approval type, validity period, and the exact legal citation and source reference). The planned output is a structured database. This stage is also the foundation for the second stage, which will develop an automated tool to quantify direct compliance cost components, based on the Standard Cost Model methodology (for example: fees, tests, professional advice and waiting time), and generate cost estimates for the regulation being reviewed.

The initiative is currently in the development phase. It is being advanced through a partnership between the Israel Regulatory Authority (IRA) and the Center for Citizen Empowerment in Israel (NGO). The work is expected to take place over the coming year, with an initial version planned within approximately five months. The Center for Citizen Empowerment in Israel is responsible for IT development, while the Authority provides professional guidance and economic methodology. The workplan includes data collection and cleaning, development of the extraction logic and a first prototype, economic calibration, a pilot in a selected regulatory field, and implementation through a dashboard for Authority staff. If the pilot is successful, the system could scale to additional regulatory fields, broader types of regulation (beyond prior approvals), and additional users, including the public.

Other relevant details

Results, outcomes and impacts: The initiative is currently in development, so no operational results or measurable impacts are available yet. Current outputs are system development, success metrics, and a testing methodology. Planned indicators include about 85% accuracy in identifying regulatory requirements in text, broad mapping coverage of prior approvals in Israel, and improved quality of direct compliance cost estimates in RIA reports submitted for review by the IRA. Measurement will rely on expert-validated benchmark samples, a pilot in a selected field, and before/after comparison of outputs. In the longer term, we expect shorter analysis times, more consistent and accurate regulatory assessment, and better capacity to estimate compliance costs. Challenges and lessons learned: Because the initiative is still in early development, the main challenges identified so far are in Stage 1 (mapping prior approvals). Israel does not yet have one complete, accessible source covering all regulation, so a key implementation risk is incomplete data coverage. We are addressing this by combining public legal sources with a parallel effort to collect procedures directly from regulators as part of the national regulatory repository. A second challenge is model logic: approvals use different terms, and the same approval may be mentioned multiple times in one text. In later stages, the cost calculator will need to combine reusable cost components with sector-specific requirements. An early lesson is that technical development must progress together with data governance and regulatory methodology. Success will require sustained cross-government co-operation, a broader smart regulation infrastructure, and stronger trust between government and the public.

About the policy initiative


Category:

  • AI policy initiatives, programmes and projects

Initiative type:

  • AI use cases/projects in the public sector

Status:

  • Proposed or under development

Start Year:

  • 2026

Target Sectors:


OECD AI Principles:

—

Other relevant urls:

—