UNDPPolicy Coherence Analyzer · How the analysis works Page 1 of 2

How the analysis works

From policy documents to rated target pairs, with an example from Mongolia

Purpose

Commitments on climate, nature and land are often made in siloes.

Ambition has risen for the planet and its people across emerging National Development Plans, Nationally Determined Contributions (NDCs), National Biodiversity Strategies and Action Plans (NBSAPs), National Adaptation Plans (NAPs) and Land Degradation Neutrality (LDN) targets. For each country, a coherence brief gives data-based evidence on synergies, trade-offs and gaps, based on the policy, finance and implementation information provided. It is a decision-support tool, not a decision-maker: the final interpretation rests with policymakers.

Scope

The same method covers targets, funding and implementation.

Level 1

Policy documents and targets

Targets in national and sectoral policies, compared pair by pair.

Every country analysed

Level 2

Funding commitments

Programmes in budget reviews, such as a Biodiversity Expenditure Review (BER).

Where budget data exists

Level 3

Implementation progress

Actions reported in Biennial Transparency Reports (BTRs) and national reports.

Where reporting data exists

Method

The method runs in 13 steps, from documents to rated target pairs.

Extraction

  1. 1Upload the documents
  2. 2Read every page
  3. 3Extract targets word for word
  4. 4Merge and check the figures
  5. 5Confirm the targets

Analysis

  1. 6Note figures and deadlines
  2. 7Classify by policy area
  3. 8Pair across documents
  4. 9Break into five parts
  5. 10Rate each pair

Synthesis

  1. 11Bring the ratings together
  2. 12Find recurring themes
  3. 13Rate the explanations

Every target is quoted word for word with its source page, and a reviewer confirms the set before any comparison. One AI agent breaks each target into its goal, action, area, actors and expected outcome; a second rates each pair from those parts and explains the rating.

Ratings

Each target pair is rated on a 5-level scale.

Mongolia, 13,404 target pairs

  • Strong alignment10%
  • Moderate alignment56%
  • Partial alignment29%
  • No clear relationship<1%
  • Potential misalignment5%

Strong and moderate alignment count as aligned: 66% of Mongolia’s target pairs. A potential misalignment is given only when the AI can name what the two targets compete for, or how pursuing one may make the other harder to reach.

Methodology as of October 2026. AI-assisted analysis for validation by national experts.
UNDPPolicy Coherence Analyzer · How the analysis works Page 2 of 2

Potential misalignment

A potential misalignment is a prompt for expert review.

Mongolia, 671 potential misalignments by type

  • Delivery & coordination59% The goals are compatible; the gap is in delivery: how one is implemented undermines the other (mismatched scale, timing, or coordination across actors).
  • Competing for resources32% The goals are compatible but compete for the same limited resource: land, water, budget, or capacity.
  • Conflicting goals9% The objectives themselves conflict: achieving one substantively means not achieving the other.

Each also records whether it appears manageable or fundamental, and the AI’s confidence.

Example · Mongolia

A spatial-planning target and a water target may compete for land.

Potential misalignment·Competing for resources·Manageable

National biodiversity targets for 2030

1 Spatial planning

By 2030, reduce biodiversity loss and maintain ecological integrity by including all territory in spatial planning and ensuring effective management.

Nationally Determined Contribution

Water resources 2

Enhance the resilience of the water sector through the utilization of advanced technologies for conservation, restoration, sustainable use and secure water availability.

AI explanation · Medium confidence

Both targets support ecosystem protection and planning, but the NDC also names water accumulation facilities, irrigation ponds, canals and pipelines … which can compete with the NBSAP target’s aim to maintain natural ecosystems and prevent further decline across all territory through land-use planning.

Resources involved: land

Between the NDC and the NBSAP, 92% of the 720 target pairs are aligned and 3% show potential misalignment. This pair is 1 of those 19.

Policy areas

Targets are grouped by policy area under 4 lenses.

Biodiversity
The Global Biodiversity Expenditure (GLOBE) taxonomy
Climate mitigation
Greenhouse gas inventory sectors of the Intergovernmental Panel on Climate Change
Climate adaptation
The seven targets of the UAE Framework for Global Climate Resilience
Human rights
Nine themes from UNDP guidance for this tool, draft under expert review

Each target keeps one primary area and any others it clearly relates to. Sri Lanka reads adaptation through its own NDC 3.0 sectors and adds a lens on loss and damage.

Social and environmental safeguards

Human in the loop. Reviewers confirm the targets; national experts verify the findings.

Traceable. The AI reads only the policy text it is given; every target keeps its source page.

Data sovereignty and model selection. A country chooses the AI model, and so where its data is processed.

Frugal AI. The carbon, energy, water and mineral footprint of every analysis is measured and published.

The Analyzer is being built as an open-source Digital Public Good that governments can run themselves.