Public AI Governance Readiness Tracker

GOAL

RESULT

PROJECT DURATION

The goal of this project was to design an interactive Looker Studio dashboard that evaluates public AI governance readiness signals across companies connected to the EU AI Pact ecosystem.

The project explores one main question:

How ready do companies appear to be for AI governance expectations based on publicly observable governance signals?

Instead of measuring legal compliance, the dashboard uses a structured scoring model to estimate public governance maturity across companies, sectors, countries, and AI governance categories.

The project combines AI governance analytics, responsible AI analysis, and business intelligence dashboarding into one practical portfolio case study.

The final result is a one-page interactive Looker Studio dashboard analyzing 138 companies across 10 AI governance categories.

The dashboard includes:

average public governance readiness score
total companies analyzed
strong-readiness company count
readiness-level distribution
governance category comparison
sector comparison
country-level score map
company governance explorer table
sector, country, and readiness filters

The dashboard shows an average public governance readiness score of approximately:

70 / 100

The main insight is that incident reporting appears as the weakest governance category in the scoring model.

This suggests that many companies communicate responsible AI principles more clearly than they communicate complaint handling, escalation, incident response, and post-deployment accountability mechanisms.

The analysis also shows stronger public governance readiness signals in sectors such as Finance / Insurance and Telecom / Cybersecurity.

This project was completed as a focused portfolio analytics project over several working sessions.

The work included:

defining the analytical question
designing the AI governance scoring framework
creating and cleaning the dataset
structuring the data into long format for Looker Studio
preparing the final CSV
building KPI cards, charts, map, filters, and explorer table
adding a methodology note
preparing the project for portfolio presentation

The project demonstrates the full workflow from AI governance concept development to dashboard-ready data modeling, visual design, and business-facing insight generation.

CASE – AI GOVERNANCE ANALYTICS

I designed an interactive Looker Studio dashboard to evaluate public AI governance readiness signals across 138 companies connected to the EU AI Pact ecosystem.
The project explores a practical responsible AI question:
How ready do companies appear to be for AI governance expectations based on publicly observable governance signals?
The dashboard does not measure legal compliance. Instead, it translates AI governance concepts into a structured scoring model that helps compare public governance maturity across companies, sectors, countries, and governance areas.
The final result is a one-page interactive dashboard that allows users to explore:
overall public AI governance readiness
sector-level readiness differences
country-level score patterns
governance category strengths and weaknesses
company-level governance gaps and priority actions
This project combines AI governance analysis, responsible AI concepts, and business intelligence dashboarding into a portfolio-ready analytics case study.

Business Problem

AI adoption is accelerating across industries, but public information about how companies govern AI is often fragmented.

Companies may publish responsible AI principles, privacy policies, transparency statements, annual reports, or governance pages, but these signals are rarely presented in a structured way that allows comparison.

This creates a practical analytics problem:

How can public AI governance signals be transformed into a clear, measurable, and interactive dashboard?

Organizations increasingly need to answer questions such as:

Are AI systems inventoried or disclosed?
Is AI risk classification visible?
Are human oversight mechanisms described?
Are fairness, bias, and transparency risks addressed?
Are users informed when AI is involved?
Are third-party AI systems assessed?
Is there an incident reporting or escalation process?
Are AI systems monitored after deployment?

This dashboard was built to make those governance questions easier to explore through data.

Dataset Collection

The dataset contains 138 companies and was prepared specifically for Looker Studio analysis.

To support both company-level and governance-category-level analysis, the dataset was structured in long format.

The final dataset contains:

138 companies × 10 governance categories = 1,380 rows

This structure makes it possible to analyze each company across multiple governance dimensions while still supporting overall KPI cards, filters, charts, maps, and company-level exploration.

Key fields include:

company ID
company name
official domain
country
region group
sector group
company size
likely AI use case focus
EU AI Act attention area
public governance readiness score
readiness level
governance category
category score
strongest governance area
weakest governance area
priority action
evidence verification level
methodology note
dashboard use note

The dataset was created as a portfolio-ready analytical dataset for governance signal exploration.

No confidential or personal data was used.

Governance Scoring Framework

The scoring model uses 10 AI governance categories.

Each company receives category-level scores, which are then used to calculate an overall Public Governance Readiness Signal Score from 0 to 100.

The 10 governance categories are:

AI inventory / use disclosure
Whether the company appears to identify, document, or publicly disclose AI systems or AI use cases.
Risk classification
Whether AI systems appear to be assessed or classified by risk level, use case, or governance relevance.
Human oversight
Whether human review, escalation, human-in-the-loop controls, or governance ownership are reflected.
Data governance
Whether data quality, data protection, privacy, and responsible data use are addressed.
Bias & fairness
Whether the company addresses fairness, discrimination, bias testing, or equal-treatment risks.
Transparency & explainability
Whether transparency, explainability, interpretability, or responsible disclosure are visible.
Incident reporting
Whether complaint handling, redress, incident reporting, escalation, or failure-response mechanisms are represented.
User transparency
Whether users are informed when AI is used or when AI-assisted decisions may affect them.
Vendor AI risk
Whether third-party AI systems, suppliers, or external AI tools are considered from a governance perspective.
Model monitoring
Whether ongoing monitoring, post-deployment evaluation, or performance review mechanisms are reflected.

Readiness levels were grouped as follows:

Score Range	Readiness Level
0–30	High governance risk
31–60	Medium readiness
61–80	Good readiness
81–100	Strong readiness

The scoring model is intended to support structured analysis, not legal judgment.

Data Preparation Process

The data preparation process focused on creating a clean, dashboard-ready dataset that could support multiple types of analysis from a single source.

Main preparation steps included:

Creating the company seed dataset
Adding company metadata
Defining the 10 governance categories
Assigning category-level governance scores
Calculating total readiness scores
Creating readiness-level groups
Identifying each company’s strongest and weakest governance areas
Adding recommended priority actions
Restructuring the dataset into long format
Removing unnecessary discovery fields
Adding methodology notes
Preparing the final CSV for Looker Studio

Because the dataset uses long format, company counts must be calculated carefully.

For example, the dashboard uses:

COUNT_DISTINCT(company_id)

instead of counting rows directly.

This prevents the 1,380 governance-category rows from being incorrectly counted as 1,380 companies.

Dashboard Design

The dashboard was built in Looker Studio using a Google Sheets-connected CSV dataset.

The design goal was to create a dashboard that feels clean, executive-readable, and easy to explore.

The final dashboard includes:

KPI cards
governance category comparison
sector comparison
readiness-level distribution
country-level map
company governance explorer table
interactive filters
methodology note

The dashboard is designed as a one-page analytics view, allowing users to understand the overall picture quickly while still being able to inspect company-level details.

KPI Section

The dashboard opens with three main KPI cards.

Average Public Governance Readiness Score

The average score across the company sample is approximately:

70 / 100

This suggests that the selected companies show moderate-to-good public AI governance readiness signals overall.

Companies Analyzed

The dashboard includes:

138 companies

This value is calculated using distinct company IDs to avoid double-counting caused by the long-format dataset.

Strong Readiness Companies

The dashboard identifies:

38 companies

as showing strong public governance readiness signals according to the scoring model.

Governance Category Analysis

One of the most important visuals in the dashboard is:

Average Score by Governance Category

This chart shows which governance areas appear stronger or weaker across the company sample.

The strongest categories include:

risk classification
model monitoring
data governance
user transparency
transparency and explainability

The weakest category is:

Incident reporting

This is one of the key findings of the project.

Many companies appear more willing to communicate responsible AI principles than to describe concrete incident handling, complaint pathways, escalation mechanisms, or post-deployment accountability processes.

This matters because AI governance is not only about having principles. It also requires operational mechanisms for monitoring systems, responding to failures, handling complaints, and improving governance over time.

Sector Analysis

The dashboard also compares average governance readiness scores by sector.

Higher-scoring sectors include:

Finance / Insurance
Telecom / Cybersecurity
Manufacturing / Industrial / Infrastructure
Healthcare / Life Sciences

These sectors may show stronger public governance signals because they often face higher regulatory, security, operational, and reputational pressure.

Lower-scoring sectors include:

Consulting / IT Services
AI / Software / Data
Legal / Standards / Certification

A lower score does not mean that companies in these sectors are non-compliant. It only means that, within this scoring model, their public governance readiness signals appear weaker or less visible.

Readiness Level Distribution

The dashboard includes a readiness distribution chart to show how companies are grouped across maturity levels.

The distribution shows that most companies fall into either good readiness or medium readiness, while a smaller group reaches strong readiness.

Approximate distribution:

Good readiness: 42%
Medium readiness: 30.4%
Strong readiness: 27.5%

The absence of a large high-risk group is expected because the sample focuses on companies already connected to an AI governance-aware ecosystem.

Country Map

The dashboard includes a country-level map showing:

Average Governance Score by Country

The map helps users explore whether public governance readiness signals appear stronger in certain geographies.

It supports questions such as:

Which countries have higher average scores?
Are readiness signals geographically concentrated?
How do country-level scores change when filters are applied?
Which countries have stronger or weaker public governance visibility?

The map is intended as an exploratory visual, not a definitive country ranking.

Company Governance Explorer

The dashboard includes a company-level table called:

Company Governance Explorer

This table allows users to inspect individual companies and compare governance gaps.

The table includes:

company size
company name
country
sector
readiness level
weakest governance gap
priority action
governance score

This section is useful because it moves the dashboard from high-level reporting into practical governance exploration.

Users can quickly identify:

highest-scoring companies
companies with weaker public governance signals
common governance gaps
recommended improvement areas
sector-specific patterns

Example priority actions include:

publish an AI inventory or use-case register
create an AI incident reporting process
add fairness testing and bias metrics
define human-in-the-loop controls
assess third-party AI suppliers
document AI risk classification and ownership
improve user transparency around AI-assisted decisions

Interactive Filters

The dashboard includes three main filters:

Sector Filter

Allows users to focus on specific industries.

Country Filter

Allows users to compare companies from selected countries.

Readiness Filter

Allows users to isolate companies by governance readiness level.

When filters are applied, all KPI cards, charts, maps, and tables update automatically.

This makes the dashboard useful for both high-level analysis and targeted exploration.

Key Insights

1. Public AI governance readiness varies clearly across sectors

Finance / Insurance and Telecom / Cybersecurity show stronger average readiness signals than several other sectors.

This may reflect stronger pressure around risk management, security, compliance, and trust.

2. Incident reporting is the weakest governance area

The most important governance gap is incident reporting.

Companies often communicate broad AI principles, but they are less likely to make incident reporting, complaint handling, redress, or escalation processes visible.

This suggests that public AI governance maturity is not only about publishing principles. It is also about showing how AI-related risks are handled after deployment.

3. Strong readiness exists, but it is not the majority

Only 38 out of 138 companies are categorized as strong readiness.

Even within an AI governance-aware sample, strong public governance signals are not universal.

4. Governance principles are easier to find than operational accountability

The dashboard suggests that companies are more likely to communicate responsible AI values than detailed operational governance processes.

This creates a gap between principle-based communication and implementation-level accountability.

5. Public signals are useful, but limited

Public information can be converted into a structured governance signal model.

However, public disclosure does not prove internal maturity.

A company may have strong internal governance processes that are not publicly visible. Another company may communicate governance principles publicly without having equally mature internal implementation.

For this reason, the dashboard uses the term:

Public Governance Readiness Signal Score

instead of:

AI Act Compliance Score

Methodological Honesty

This project is intentionally careful in its claims.

The dashboard does not provide:

legal compliance assessments
official AI Act compliance ratings
verified audits
certifications
legal advice

The scores should be interpreted as:

estimated public AI governance readiness signals based on a structured scoring model

They should not be interpreted as proof of legal compliance or internal governance maturity.

This distinction is important because responsible analytics should not overstate what the data can prove.

Limitations

The project has several important limitations.

1. Scores are estimates

The scoring model is based on structured governance signals, not verified internal company documentation.

2. Public disclosure is not the same as internal maturity

Some companies may have strong internal AI governance practices that are not visible publicly.

3. The dashboard does not measure legal compliance

The project should not be used as a legal assessment or AI Act compliance audit.

4. The sample may be biased toward AI-aware companies

Because the sample focuses on companies connected to the EU AI Pact context, it is likely more governance-aware than a random company sample.

5. Future validation would strengthen the model

A stronger future version could validate each company score using official company pages, AI policies, transparency reports, annual reports, responsible AI pages, privacy documents, or sustainability reports.

Business Value

This project demonstrates how complex governance and regulatory topics can be translated into practical business intelligence.

The dashboard could support:

AI governance teams
compliance preparation
responsible AI reporting
vendor risk screening
executive-level governance summaries
sector benchmarking
consulting-style maturity assessments
early AI risk discovery

The project shows that AI governance is not only a legal or policy topic. It can also be structured, scored, visualized, and communicated through data analytics.

Tools Used

Looker Studio
Google Sheets
CSV data preparation
Data cleaning
Long-format data modeling
Dashboard design
AI governance scoring framework
Responsible AI analysis
EU AI Act-oriented governance logic
OECD AI Principles-oriented categories
NIST AI RMF-inspired trustworthy AI dimensions

Skills Demonstrated

This project demonstrates skills in:

business intelligence dashboarding
data modeling for dashboard tools
responsible AI analysis
AI governance framework design
KPI development
data cleaning and restructuring
methodology documentation
governance risk analysis
data storytelling
executive dashboard design
Looker Studio reporting

Conclusion

The Public AI Governance Readiness Tracker demonstrates how AI governance concepts can be transformed into a practical analytics product.

The dashboard shows that public AI governance readiness varies across sectors, countries, and governance categories. It also identifies incident reporting as the weakest governance area in the scoring model.

This finding is important because responsible AI requires more than publishing principles. A mature AI governance program should also show how systems are inventoried, classified, monitored, explained, challenged, escalated, and improved over time.

The project is intentionally cautious in its claims. It does not present the results as legal compliance ratings. Instead, it presents them as structured public governance readiness signals.

This makes the dashboard useful as a portfolio case study in:

AI governance analytics
responsible AI
Looker Studio dashboarding
business intelligence
governance risk analysis
data storytelling

The project shows how a data analyst can turn an emerging governance problem into a clear, interactive, and decision-oriented dashboard.