LIVE │ 2026 Q1
2026-04-22 │ PL / EN
§ 01 · Methodology

What the score
actually measures.

The Index measures how maturely companies discuss AI in their annual reports — not how AI-advanced they are. Two signals combine into one score.

§ 01 · 01

Breadth

18 dimensions · 0–18 count

We scan each annual report against 18 categories of AI maturity: machine learning and deep learning, natural language processing, computer vision, robotics and process automation, data infrastructure, AI cloud services, AI strategy, innovation capability, digital transformation, organisational readiness, customer experience AI, supply chain intelligence, risk management, operational excellence, governance and leadership, talent and skills, ethics and regulation, and measurement of AI outcomes.

A category scores 1 if it appears meaningfully anywhere in the report. The breadth score is a count: how many of the 18 does this company actually address?

Note

This is the coverage signal — it answers: does this company have a view on AI across the business, or only in one corner?

§ 02 · 02

Quality

Anchor comparison · 0–1 scale

We maintain a reference corpus of annual reports from organisations recognised as leaders in AI fluency: NVIDIA, Microsoft, Alphabet, Meta, Salesforce, Amazon, Netflix, Visa, BlackRock, JPMorgan, Palantir, Siemens, SAP, and four others.

For each WIG140 company, we measure how close the language of its AI-relevant disclosures is to that reference corpus. A company that discusses AI the way Netflix discusses AI scores higher than one that drops the word into a single paragraph of boilerplate.

Note

This is the quality signal — it answers: does this company's AI language reflect genuine strategic thinking, or compliance-driven box-ticking?

One percentile rank.

Both signals feed a single composite, which is then converted to a percentile rank within the WIG140 universe. A score of 82 means this company outranks 82% of the universe on AI disclosure maturity. The score is comparable across quarters and across sectors.
§ 02 · Combined score
§ 03 · Data sources

Annual reports via public filings.

Primary source
ESPI annual report filings (Polish regulatory disclosure system)
Document types
Management board reports · ESEF XHTML · PDF
Universe
WIG20 + mWIG40 + sWIG80 · 140 companies
Historical depth
2023–present · 3-year trajectory
Refresh cadence
Quarterly · new filings trigger rescore
Anchor corpus
17 global AI-fluent companies · updated annually
§ 04 · Limitations

What this isn't.

L01
—
The Index measures disclosure, not adoption. A company may be doing significant AI work without discussing it in its annual report — and vice versa.
L02
—
Annual reports are backward-looking. The score reflects what management chose to say about the prior year, not what is happening today.
L03
—
The anchor corpus reflects global best practice as of the filing date. As AI discourse evolves, so does the reference point.
L04
—
This is not investment advice. Disclosure maturity is one signal among many.