The YESFOR Framework, our own method for assessing a company's AI readiness, yesfor

Why we created our own methodology instead of using someone else's

For the first few months of working with Polish corporations, we used external frameworks: the Cisco AI Readiness Index, the Microsoft AI Maturity Model, the Gartner Maturity Model for AI Adoption. They are all good. They all share the same flaw.

None of them accounts for the Polish context. And the Polish context is:

A different ownership structure (40 percent of mid-sized companies have a single dominant owner)
A different management culture (a shorter decision horizon, a greater distance to change)
A different AI labor market (24 times fewer AI specialists than in the UK, Eurostat 2025 data)
Different regulation (the AI Act + GDPR + the UODO interpreted locally)
A different technology base (more legacy ERP, less cloud-native)
A different budget per project (Horvath 2026: 0.35 percent of turnover on AI in mid-size companies in the EU)

A foreign framework adapted to Polish reality gives a result like "the company is 65 percent ready," but it does not say what it is ready for. The YESFOR Framework was born from the need for an audit report to end with a concrete answer, not a scale.

A comparison of the YESFOR Framework with existing methodologies

| Feature | Cisco AI Readiness Index | Microsoft AI Maturity Model | Gartner Maturity Model | YESFOR Framework | |---|---|---|---|---| | Number of dimensions | 6 | 7 | 5 | 4 pillars × 3 dimensions = 12 | | Measurable 0-100 score | Pacesetter benchmarking | Stages 1-5 | Stages 1-5 | Yes, a 0-100 score | | Empirical dimension weights | No | No | No | Yes (RAND/MIT/McKinsey) | | Polish context | No | No | No | Yes | | Open methodology (publicly documented) | Partly | Partly | No | Yes | | Vendor-agnostic | Yes | No (Microsoft bias) | Yes | Yes | | Self-assessment time | 60+ min | 90+ min | 60 min | 30 min |

YESFOR is not better than Cisco or Microsoft. It is different, designed for the specific context in which we work.

The four pillars of the YESFOR Framework

YESFOR breaks a company's AI readiness into four pillars. Each has its own weight in the total score. The weights are not arbitrary, they come from an analysis of data on the causes of AI project failure.

Pillar O, Organization (weight 25%)

What it measures: Whether the organization as a system is ready to support an AI project from decision to maintenance.

Three dimensions:

C-level sponsorship (0-100), whether a specific board member is accountable for the project, whether they have KPIs tied to its success, whether they take part in monthly reviews
Change culture (0-100), the company's historical track record in transformational implementations (the last 3 years), the level of resistance to change, conflict-handling mechanisms
Governance (0-100), the decision structure for the AI project, clarity of roles, frequency of reviews, escalation of problems

Why a weight of 25%: RAND 2024 shows that projects with sustained board engagement have a 68 percent chance of success versus 11 percent when sponsorship is lost. A sixfold difference. But sponsorship alone is not enough, which is why it is not 40 percent.

Pillar P, Processes (weight 30%)

What it measures: Whether processes are organized enough for AI to have something to improve.

Three dimensions:

Process mapping (0-100), how many key processes are actually documented (not as "how it should be," but as "how it really is"), and the date of the last update
Standardization (0-100), whether the same process is performed identically across different teams and locations, or whether each has its own version
KPI baseline (0-100), whether processes have measurable indicators (time, cost, quality, satisfaction), and whether those indicators are collected regularly

Why a weight of 30%: McKinsey 2025 found that companies with a measurable AI impact on EBIT are 2 times more likely to redesign end-to-end processes before choosing a model. Processes are a necessary condition, without them AI has nothing to optimize.

Pillar S, Systems & Data (weight 30%)

What it measures: Whether the data and systems can support an AI project on the technical side.

Three dimensions:

Data quality (0-100), the five Gartner dimensions (completeness, consistency, accuracy, timeliness, uniqueness) for the data critical to the project
Architecture and integrations (0-100), whether systems are integrated (versus siloed), whether they have APIs, whether the data can be retrieved programmatically
AI-ready data governance (0-100), per Gartner's definition: fit to the use case, ownership of resources, automated pipelines, live metadata, continuous quality control

Why a weight of 30%: Informatica's CDO Insights 2025 indicates that 43 percent of leaders cite data quality as the main obstacle. It is the single largest technical cause of failure. Without data, AI has nothing to anchor to.

Pillar Skills, Competence (weight 15%)

What it measures: Whether the people in the organization are able to implement, use and develop AI.

Three dimensions:

Board AI literacy (0-100), real AI knowledge at C-level: what it is, what it can do, what it cannot do, how to read a business case
Change readiness of operational teams (0-100), sentiment toward AI in the end departments, readiness to work with new tools, adoption history
Retention of key talent (0-100), whether the company will keep the technical people needed to maintain AI after the consultant leaves (capability transfer)

Why only 15%: Because it is the pillar that is fastest to make up, through training, change management and recruitment. Even though 31 percent of employees deliberately sabotage AI implementations (Writer/Workplace Intelligence 2025), this pillar is not a blocker if the other three are in order. Competence can be bought in. Organizational culture cannot.

The AI Readiness Score formula

AI Readiness Score = (O × 0.25) + (P × 0.30) + (S × 0.30) + (Skills × 0.15)

Where each variable is the weighted average of its three component dimensions (0-100).

The AI Readiness Score interpretation scale

| Score | State | What to do | |---|---|---| | 0-30 | Structural unreadiness | STOP AI implementations. First put processes and data in order. An attempt to implement AI is an almost certain failure. | | 31-50 | Initial readiness | Quick wins are possible in individual processes. Large transformational projects, NO. First build the foundations. | | 51-70 | Operational readiness | You can scale pilots into production. Implementations in the areas of highest readiness. Others, with additional preparation. | | 71-85 | Strategic readiness | Enterprise-wide programs make sense. You can plan a 2 to 3 year transformation with a high probability of success. | | 86-100 | Pioneer readiness | The company is in the top 5% of the Polish market. The focus shifts from readiness to competitive innovation. |

Why the weights are 25/30/30/15, and not something else

The decision on the weights is the most important methodological decision in the whole framework. That is why we justify it publicly.

What we considered

Option A, equal weights (25/25/25/25): The simplest, but it ignores the fact that not every pillar affects success equally. The empirical evidence says otherwise, so we reject it.

Option B, heavy organization (40/20/30/10): Logical, since nothing works if the board does not support it. But RAND data shows that sponsorship alone (68 percent success) is not enough without processes and data. Rejected.

Option C, heavy data (15/25/45/15): The classic "data is king." But the problem is not there, data can be brought to readiness in 3 to 6 months. Processes and culture cannot. Rejected.

Option D, the chosen one (25/30/30/15): Processes and data each carry 30 percent, because they are the largest single causes of failure according to the research. Organization 25 percent, because sponsorship is necessary but not sufficient. Skills 15 percent, because it is the pillar that is fastest to make up.

The empirical evidence behind the weights

| Pillar | Weight | Main data source | Key figure | |---|---|---|---| | Processes | 30% | McKinsey 2025 | 2x higher ROI when redesigning processes before AI | | Systems & Data | 30% | Informatica CDO 2025 | 43% of leaders cite data quality as the #1 blocker | | Organization | 25% | RAND 2024 | 6x difference in success (68% vs 11%) depending on sponsorship | | Skills | 15% | Writer/Workplace Intel 2025 + Gartner I&O 2026 | 31% sabotage + the skills gap as 38% of failure citations |

The weights will be reviewed every 12 months in light of new research. The first review: May 2027.

How to run a YESFOR Assessment in your company

There are three ways. The choice depends on how deeply you want to understand.

Method 1: AI Readiness Self-Assessment (30 minutes, free)

An online quiz on yesfor.ai. 36 questions (3 dimensions × 4 pillars × 3 questions), answered on a 1-5 scale. The algorithm calculates a preliminary AI Readiness Score and points to the 3 areas of the largest gaps.

What you get:

A preliminary 0-100 score for each pillar
A total AI Readiness Score
The top 3 areas for immediate repair
A preliminary recommendation: readiness to implement AI yes/no/conditionally

What you do NOT get:

Validation of the results (self-reporting bias)
A process map
A business case for specific use cases
An implementation roadmap

When it makes sense: As an initial screening before deciding on a full audit. Or when the audit budget is not yet approved and you need arguments for the board.

Method 2: YESFOR Assessment Light (1 week, 8,000 to 12,000 PLN)

A short external audit, 2 to 3 conversations with the C-level plus an analysis of 2 to 3 key systems. A yesfor.ai consultant verifies the self-assessment results, corrects the bias and adds observations from the interviews.

What you get:

A validated AI Readiness Score (the difference versus the self-assessment is often 15 to 25 points lower)
A short report (15 to 20 pages)
Recommendations, the top 5 actions for 90 days
A recommendation decision: full audit / quick wins / do not enter AI now

When it makes sense: Mid-market of 100 to 500 people, a first approach to AI, a limited budget.

Method 3: A full audit with an AI Readiness Score (4 to 8 weeks, 25,000 to 200,000 PLN)

The full procedure described in the article The pre-implementation AI audit. The AI Readiness Score is one of the deliverables, alongside the process map, the business cases and the roadmap.

When it makes sense: Enterprise of 500+ people, an implementation budget of 500,000 PLN and above, a board decision required within a quarter.

Frequently asked questions

Further materials

The pre-implementation AI audit, the full guide, why 80 percent of AI projects fall apart and how to prevent it
Why AI projects in companies do not work, an analysis of 1,300+ implementations
A discovery call with yesfor.ai, 30 minutes, a conversation about your specific context

Methodological sources

Ryseff, J., De Bruhl, B. F., Newberry, S. J. (2024). The Root Causes of Failure for Artificial Intelligence Projects. RAND Corporation. Link
MIT Project NANDA (2025). The GenAI Divide: State of AI in Business 2025.
McKinsey & Company (2025). Global AI Survey: November 2025.
Gartner (2026). AI Projects in I&O Stall Ahead of Meaningful ROI Returns.
Informatica (2025). CDO Insights Survey 2025.
Writer / Workplace Intelligence (2025). Generative AI in the Enterprise.

The YESFOR Framework is a trademark of yesfor.ai. The methodology is published under a CC BY-NC 4.0 license, you may quote it with attribution, commercial use requires a license.