AI Readiness: Is Your Company Prepared for Artificial Intelligence?

Published on April 2, 2025 by Christopher Wittlinger

Before companies invest in AI projects, they should realistically assess their starting position. A structured readiness assessment helps identify strengths and close gaps before they become project risks. In our consulting practice, we regularly see companies invest €100,000+ in AI initiatives that fail due to missing fundamentals — not missing technology.

This guide presents our proven assessment framework: 5 dimensions, 5 criteria each, scored on a scale of 1–5. The total score of 25–125 gives you a clear indication of where you stand and which investments have the greatest leverage.

The Five Dimensions of AI Readiness

1. Data Maturity

Data is the fuel for AI systems — and the dimension where most companies have the largest gaps. In our assessments, 70% of companies score less than 3 out of 5 in this dimension.

Evaluate the following criteria:

Availability: Does the required data even exist?

Score 1: Relevant data is not systematically captured. Many processes are paper-based or in personal files.
Score 2: Data exists in individual systems but is fragmented. Customer information in Outlook, revenue in Excel, orders in the ERP — with no connection.
Score 3: Core processes are digitized, data is captured in structured systems. There is a central ERP and a CRM.
Score 4: Comprehensive data capture across all business processes. A data warehouse or data lake exists as a central data source.
Score 5: Real-time data from all relevant sources flows together automatically. Streaming pipelines, IoT integration, complete data history available.

Quality: How complete, current, and accurate is the data?

Score 1: No quality standards. High proportion of missing values (>30%), inconsistent formats, no validation.
Score 2: Known quality problems but no systematic treatment. Occasional manual cleanup.
Score 3: Basic quality checks exist. Missing values below 15%. Standardized formats for core fields.
Score 4: Automated data quality checks. Missing values below 5%. Anomaly detection on data entry. Regular monitoring.
Score 5: Comprehensive data quality management. Automated correction, deduplication, continuous monitoring with alerting. SLAs for data quality defined.

Accessibility: Can teams access the data?

Score 1: Data is locked in silos. Access requires manual requests to the IT department. Wait times of days to weeks.
Score 2: Some data available via exports (CSV, Excel). No self-service. Every analysis requires IT support.
Score 3: BI tools enable basic self-service access to standard reports. APIs for core systems exist.
Score 4: Data platform with self-service access. Business departments can create their own queries. Role-based access controls.
Score 5: Full data platform with data catalog, self-service analytics, API gateway, and real-time access. Data treated as a product.

Governance: Are there clear rules for data usage and protection?

Score 1: No defined rules. Unclear who is responsible for which data. Data privacy is handled ad hoc.
Score 2: Basic data privacy policy exists (GDPR baseline). No defined data owners. No systematic classification.
Score 3: Data owners named for core systems. Data privacy concept documented. Basic access controls implemented.
Score 4: Comprehensive data governance framework. Data classification, retention policies, audit trails. Data stewards in business departments.
Score 5: Automated governance with policy-as-code. Data lineage fully traceable. Regular audits. Privacy-by-design implemented.

Integration: Can different data sources be combined?

Score 1: Every system is an island. Integration only through manual exports and copy-paste.
Score 2: Individual point-to-point integrations exist (e.g., ERP-CRM sync) but are fragile and manually maintained.
Score 3: ETL pipelines for important data flows. Basic master data management approaches.
Score 4: Integration platform (iPaaS or middleware). Standardized APIs. Most systems are connected.
Score 5: Event-driven architecture. Real-time integration of all systems. Unified data model across domains.

Warning sign: If data lives in silos and manual Excel exports are the standard, a data project is needed before the AI project. We have written a dedicated post on this: Data quality as a success factor for AI.

2. Technical Infrastructure

The right foundation must be in place to not only develop AI models but also reliably operate them.

Compute Resources: Cloud access or own GPU capacities.

Score 1: No cloud usage. All workloads on local machines or a single server.
Score 2: First cloud services in use (e.g., email, storage) but no compute workloads.
Score 3: Cloud infrastructure available for development. Developers can use VMs and container instances. No GPU access.
Score 4: Managed cloud infrastructure with GPU instances on demand. Kubernetes or comparable orchestration.
Score 5: Scalable ML infrastructure with auto-scaling, spot instance management, multi-cloud strategy. GPU clusters available for training.

DevOps/MLOps Maturity: Versioning, testing, deployment pipelines.

Score 1: No version control system. Deployment via FTP or manual copying.
Score 2: Git for code versioning. Manual deployments with a checklist. No automated testing.
Score 3: CI/CD pipeline for standard applications. Automated tests. Docker in use.
Score 4: MLOps basics: model versioning, experiment tracking, automated testing for ML pipelines.
Score 5: Full MLOps platform: automated training, model registry, A/B testing, monitoring, automatic retraining.

Integration and APIs: Interfaces to existing systems.

Score 1: No APIs. Systems communicate via file exports or not at all.
Score 2: Individual REST APIs exist but are not documented or standardized.
Score 3: API-first approach for new systems. Documented APIs for core systems.
Score 4: API gateway, versioning, rate limiting. Microservice architecture for new components.
Score 5: Event-driven architecture with message broker. GraphQL or gRPC. API marketplace for internal services.

Security: Encryption, access controls, audit logs.

Score 1: Basic network security (firewall). No encryption of data at rest. No audit logs.
Score 2: HTTPS for external services. Simple password policies. Basic access control.
Score 3: Role-based access controls. Encryption at rest and in transit. Basic audit logging.
Score 4: Zero-trust approach. MFA for all systems. SIEM solution. Regular penetration tests.
Score 5: Security-as-code. Automated vulnerability scans in CI/CD. SOC-2 compliance. Incident response plan tested.

Scalability: Can the infrastructure grow with demand?

Score 1: Fixed hardware. Capacity expansion takes weeks to months.
Score 2: Individual cloud services scalable but core infrastructure is fixed.
Score 3: Containerized applications. Horizontal scaling possible for web workloads.
Score 4: Auto-scaling for most services. Infrastructure-as-code. Reproducible environments.
Score 5: Fully elastic infrastructure. Serverless for suitable workloads. Multi-region. Disaster recovery tested.

Warning sign: If deployment still means “manually copying to the server,” the foundation for productive AI systems is missing.

3. Skills and Talent

People make the difference — and AI competence means more than just hiring data scientists.

Technical Know-How: Data scientists, ML engineers on the team?

Score 1: No employees with ML/AI skills. IT department focuses on operations.
Score 2: Individual employees with basic knowledge (online courses, personal interest). No dedicated role.
Score 3: 1–2 data scientists or ML engineers on the team. Initial experience with model development.
Score 4: Dedicated data science team (3–5 people). Experience with productive ML systems.
Score 5: Complete ML team (data scientists, ML engineers, data engineers, MLOps). Experience with various model types and production systems.

Domain Knowledge: Do the technical experts understand the business?

Score 1: Technology and business departments work in separate worlds. No shared understanding.
Score 2: Occasional exchange. Technicians understand basic business processes.
Score 3: Regular alignment. Technicians attend business department meetings. Initial cross-functional projects.
Score 4: Embedded model: technicians sit in business departments or work closely with domain experts. Shared KPIs.
Score 5: Deep integration. Business experts with technical understanding. Data translators as a bridge. Product owner model for AI products.

Leadership Skills: Can managers steer AI projects?

Score 1: Management sees AI as an IT topic. No involvement in project steering.
Score 2: Basic interest. Management lets AI projects run but does not actively steer.
Score 3: Management can prioritize AI projects and allocate resources. Basic understanding of potentials and limitations.
Score 4: Executives actively steer AI projects. Understand metrics and can evaluate trade-offs.
Score 5: AI literacy at C-level. Strategic integration into business decisions. Data-informed leadership culture.

Training: Are there programs for skill development?

Score 1: No AI-related training opportunities. Employees train on their own initiative.
Score 2: Individual employees occasionally attend external training. No systematic program.
Score 3: Budget for AI training defined. Selected employees participate in structured programs.
Score 4: Systematic training program at multiple levels (management, business departments, technical). Regular internal workshops.
Score 5: Learning organization. Internal AI training program, community of practice, hackathons, partnerships with universities.

Recruitment: Can you attract and retain AI talent?

Score 1: No focus on AI recruitment. Employer brand has no tech association.
Score 2: First attempts to post AI roles. Difficulty finding candidates.
Score 3: Active recruitment. Competitive salaries. First AI employees hired.
Score 4: Attractive employer for AI talent. Interesting projects, modern infrastructure, training.
Score 5: Employer of choice. Strong employer brand in the AI space. Talent pipeline. Partnerships with universities and research institutes.

Warning sign: If all “AI competence” rests with a single person, that is a significant risk. A bus factor of 1 means one job change can endanger your entire AI program.

4. Organizational Culture

Culture eats strategy for breakfast — and AI projects are particularly culture-sensitive because they change workflows, roles, and decision processes. More on this in our post on AI change management.

Experimentation: Are pilot projects supported?

Score 1: Innovation is seen as risk. New ideas must go through long approval processes.
Score 2: Isolated innovation projects driven by individual champions.
Score 3: Innovation budget exists. Pilot projects are formally possible but require strong justification.
Score 4: Dedicated innovation program. Employees can use 10–20% of their time for experiments.
Score 5: Innovation culture as a core value. Rapid prototypes are rewarded. Fail-fast mentality anchored.

Failure Tolerance: Is it okay to fail and learn from it?

Score 1: Mistakes are punished. Blame culture. Employees avoid risks.
Score 2: Mistakes are tolerated but not analyzed. No systematic learning culture.
Score 3: Post-mortems for major failures. Basic understanding that experiments can fail.
Score 4: Blameless post-mortems are standard. Failed experiments are communicated as learning opportunities.
Score 5: Psychological safety. Teams report failures openly. Lessons learned are actively shared and utilized.

Data-Driven Decisions: Do facts count or hierarchies?

Score 1: Decisions based on experience and hierarchy. “That’s how we’ve always done it.”
Score 2: Data is occasionally used for justification but not as the primary decision basis.
Score 3: Standard reports and dashboards are regularly used. Central KPIs are defined.
Score 4: Data-based argumentation is standard. A/B tests for important decisions. Executive dashboards.
Score 5: Data-first culture. Every strategic decision is backed by data. Predictive analytics for planning.

Willingness to Change: Are employees open to new tools?

Score 1: Strong resistance to change. New tools are boycotted or ignored.
Score 2: Passive acceptance. Employees use new tools when they must but without engagement.
Score 3: Basic openness. Employees try new tools when the benefit is clearly communicated.
Score 4: Active participation. Employees make improvement suggestions and proactively test new tools.
Score 5: Change champions in all departments. Employees drive digitalization themselves. Internal community.

Cross-Departmental Collaboration: Do teams work together or against each other?

Score 1: Strong silos. Departments share neither data nor knowledge. Competitive relationship.
Score 2: Collaboration only at management level. Operational teams stay in their areas.
Score 3: Project-based collaboration. Cross-functional teams for specific initiatives.
Score 4: Structural collaboration. Shared services, common platforms, regular cross-departmental meetings.
Score 5: Seamless collaboration. Matrix organization focused on value streams rather than departmental boundaries.

Warning sign: If every decision must go through three committees, agile AI development becomes difficult. AI projects need fast iteration cycles — weeks, not months.

5. Strategy and Governance

The overarching framework provides direction and boundaries. Without it, AI projects become isolated experiments that never scale.

Clear Goals: What should AI specifically achieve?

Score 1: AI is a buzzword in presentations but no concrete use cases are defined.
Score 2: Vague idea (“We want to use AI”) but no measurable goals.
Score 3: 2–3 concrete use cases identified. Rough expectations for value formulated.
Score 4: Prioritized use case pipeline with measurable KPIs. Business cases created for top initiatives.
Score 5: AI strategy as part of the business strategy. Clear vision, prioritized roadmap, defined success criteria.

Budget: Are funds allocated for multi-year initiatives?

Score 1: No dedicated AI budget. Projects must be funded from ongoing IT budgets.
Score 2: One-time budget approved for a pilot project. No long-term financial planning.
Score 3: Annual budget for AI initiatives. Sufficient for 1–2 projects.
Score 4: Multi-year budget with clear allocation for infrastructure, personnel, and projects.
Score 5: Strategic investment budget with scaling path. Venture approach with portfolio management across use cases.

Responsibilities: Who drives the topic?

Score 1: Nobody is explicitly responsible for AI. Topic floats between IT and management.
Score 2: Informal AI champion who drives the topic alongside regular responsibilities.
Score 3: Dedicated AI lead (e.g., Head of Data & AI). Clear reporting line.
Score 4: AI team with clear organizational structure. Steering committee with C-level participation.
Score 5: Chief AI Officer or CDO at C-level. AI Center of Excellence. Clear governance over evaluation, release, and monitoring of AI systems.

Ethics and Compliance: Are there guidelines for responsible AI?

Score 1: No awareness of AI ethics or regulatory requirements.
Score 2: Basic awareness but no documented guidelines.
Score 3: AI ethics policy exists. Basic measures for bias avoidance and transparency.
Score 4: Comprehensive AI governance framework. Risk classification for use cases. Impact assessments before deployment.
Score 5: Integrated responsible AI program. Automated bias checks, explainability standards, regular audits. EU AI Act compliance proactively implemented.

Stakeholder Alignment: Is everyone pulling in the same direction?

Score 1: AI is driven only by IT or individual enthusiasts. No management buy-in.
Score 2: CEO or CTO finds AI interesting but no active support when obstacles arise.
Score 3: Management support available. Budget approved. But business departments are skeptical.
Score 4: Broad support at management and business department level. Shared understanding of goals.
Score 5: AI as a strategic priority at all levels. CEO as active sponsor. Business departments drive use cases themselves.

Warning sign: If AI is a buzzword in presentations but no concrete use cases with measurable goals are defined, every project becomes a political football.

The Readiness Scoring: Evaluation

Rate each of the 25 criteria on the 1–5 scale. The total score falls between 25 and 125 points.

Interpretation of the total score:

Score Range	Classification	Recommendation
25–45	Fundamental deficits	Focus on digitalization and data infrastructure. AI projects are premature. Invest first in basic competencies.
46–65	Beginnings visible	Targeted investments in the weakest dimensions. Start with a simple, data-adjacent pilot project.
66–85	Solid foundation	You are ready for structured AI projects. Focus on scaling and governance.
86–105	Well positioned	Ambitious projects are possible. Focus on industrialization, MLOps, and Center of Excellence.
106–125	Best-in-class	You belong to the top tier. Focus on innovation, differentiation, and competitive advantage through AI.

Industry Benchmarks (average values from our assessments):

Industry	Typical Score	Strongest Dimension	Weakest Dimension
Financial Services	78	Technical Infrastructure	Organizational Culture
Manufacturing	58	Domain Knowledge	Data Maturity
Healthcare	52	Compliance Awareness	Technical Infrastructure
E-Commerce/Retail	72	Data Maturity	AI Competencies
Professional Services	62	Strategy	Technical Infrastructure
Public Sector	45	Governance	Experimentation

Typical Assessment Process and Costs

A professional AI readiness assessment typically takes 2–3 weeks and follows this process:

Week 1 — Discovery: Interviews with 8–12 stakeholders (C-level, IT leadership, business department heads, operational employees). Analysis of the existing IT landscape and data infrastructure. Review of existing strategy documents and project plans.

Week 2 — Analysis and Scoring: Evaluation of all 25 criteria based on discovery results. Identification of quick wins and strategic gaps. Development of 3–5 prioritized recommendations.

Week 3 — Results Presentation: Management presentation with score, benchmark comparison, and prioritized roadmap. Detailed report with action plan and budget framework per action area. Workshop for joint prioritization.

Costs: A professional assessment ranges from €8,000–€15,000 — depending on company size and the number of stakeholders involved. That sounds like an investment, but it quickly puts things in perspective: it typically prevents misspending in the five- to six-figure range.

Case Study: From Score 48 to 88 in 18 Months

A mid-sized insurance broker with 200 employees had an AI readiness assessment conducted in early 2024. The starting position:

Dimension	Score (Early 2024)	Main Problem
Data Maturity	8/25	Customer data in 4 different systems, no centralized view
Technical Infrastructure	10/25	Everything on-premise, no CI/CD, manual deployments
Competencies	12/25	One developer with Python skills, no ML know-how
Organizational Culture	10/25	Conservative culture, fear of automation
Strategy/Governance	8/25	AI as a vague innovation topic, no budget
Total	48/125

Action package (prioritized):

Quarter 1–2: Build data foundation (Investment: €60,000): Introduction of a central data platform (cloud data warehouse on Snowflake). Migration of customer data from 4 systems with deduplication. Setup of basic data quality checks. Result: Data maturity from 8 to 16 points.

Quarter 2–3: Modernize infrastructure (Investment: €35,000): Cloud migration of the development environment (AWS). Introduction of Git, Docker, and CI/CD pipeline. First API interfaces to the policy management system. Result: Technical infrastructure from 10 to 17 points.

Quarter 2–4: Build competencies (Investment: €25,000): Hiring a junior data scientist. AI fundamentals training for 30 employees (2-day workshop). Management briefing on AI potentials and limitations. Result: Competencies from 12 to 18 points.

Quarter 3–4: Culture and pilot project (Investment: €40,000): Launch of a pilot project: AI-powered risk assessment for new applications. Involving business employees as domain experts in the project team. Regular show-and-tell sessions for all employees. Result: Organizational culture from 10 to 17 points.

Quarter 4–6: Formalize strategy (Investment: €15,000): Development of an AI strategy with a prioritized use case pipeline. Definition of AI governance guidelines. Approval of a multi-year budget by executive management. Result: Strategy/Governance from 8 to 20 points.

Results after 18 months:

Dimension	Before	After	Change
Data Maturity	8	16	+8
Technical Infrastructure	10	17	+7
Competencies	12	18	+6
Organizational Culture	10	17	+7
Strategy/Governance	8	20	+12
Total	48	88	+40

Total investment: approximately €175,000 over 18 months. The pilot project (risk assessment) delivered added value of approximately €120,000 per year after 6 months in production through faster processing and better risk evaluation.

Tools for Self-Assessment

For an initial orientation, you can conduct the assessment internally. Some tips:

Who should evaluate? At least 5 people from different areas: IT leadership, a business department head, an operational employee, executive management, and — if available — someone with data competence. Have each person evaluate independently and compare results. Large discrepancies between evaluations are a signal in themselves: they show that the picture of the company depends heavily on perspective.

Common self-assessment errors:

Dunning-Kruger Effect: Teams without AI experience systematically overestimate their readiness. When nobody knows what “good” looks like, the status quo appears adequate.
Wishful thinking about data quality: “Our data is good” is the most common misjudgment. Only when you concretely try to prepare data for a use case does reality show itself.
Infrastructure overestimation: “We have cloud” does not automatically mean ML workloads are possible. Check concretely: could you deploy a Docker container with a Python ML model tomorrow?

Helpful calibration questions:

How long does it take to produce a simple report on customer satisfaction for the last 12 months? (Under 1 hour = Score 4+, over 1 week = Score 2-)
What percentage of your business processes could you describe using data? (Over 80% = Score 4+, under 30% = Score 2-)
What happens if your best developer quits tomorrow? (Nothing critical = Score 4+, project stop = Score 2-)

Next Steps After Assessment

Prioritize: Which gaps have the greatest impact? Typically, data maturity is the most critical bottleneck. Without data, no AI project — regardless of how good the infrastructure and team are.
Identify Quick Wins: Where can you show quick progress? Typical quick wins: introduction of a BI dashboard, first API interface to the ERP, AI fundamentals workshop for management.
Create Roadmap: Realistic timeline with clear milestones and budgets. Think in quarters, not years.
Choose Pilot Project: Start a project with high probability of success. Criteria: available data, clear benefit, engaged business department, manageable complexity.
Develop AI Strategy: Based on the readiness results, develop an AI strategy that accounts for your specific starting position.

Conclusion

An honest readiness assessment is not a brake but an accelerator. It prevents expensive false starts and creates the foundation for sustainable AI success. The companies that make the fastest progress are not those with the largest budgets, but those that assess their starting position most realistically and systematically turn the right levers.

The most important insight from over 30 assessments: there is no company that is “not ready for AI.” There are only companies that start in the wrong places. A structured assessment shows you where the right entry point lies.

Want to have your AI readiness professionally assessed? Contact us for a no-obligation conversation.