When does a company need support with AI implementation?
A company needs support with AI implementation when several pilots are running without a basis for decisions, and teams use AI without a policy or training.
- Several AI pilots are running, and the board has no basis for deciding whether to scale or close them.
- Teams use AI tools without a use policy or a training programme.
- The board needs to know which company systems fall under the EU AI Act and which deadlines apply to them.
What does Lemur Media's AI implementation support cover?
Lemur Media assesses organisational readiness, builds a use case portfolio rated for value and risk, runs pilots and prepares an EU AI Act compliance plan.
- 01Readiness assessmentLEMUR AI Readiness Index: five dimensions rated on a 1–5 scale against documented criteria, a gap map and priorities.
- 02Use case portfolioAn assessment of the business value and feasibility of each idea, plus risk classification under the EU AI Act.
- 03Pilots with a decision gatePilot charter with a success measure, business owner, budget and decision date.
- 04AI skillsA role-based training programme that meets the obligation under Article 4 of the EU AI Act.
- 05AI governanceAI system register, use policy, and division of responsibility between business, IT, security and legal.
How do we assess organisational readiness?
The LEMUR AI Readiness Index covers the five dimensions that determine whether AI implementations deliver financial results. We rate each dimension on a 1–5 scale against documented criteria and calculate the 0–100 index as (sum − 5) × 5.
- Leaders and goals
Business goals for AI, a budget and an owner with a board mandate. - Education and skills
Use policy and AI skills in teams, including the obligation under Article 4 of the EU AI Act. - Data model
Data availability, quality, metadata and access rights. - AI use and value
Use cases with estimated value and a success measure. - Risk and regulation
System register, classification under the EU AI Act, GDPR and security.
How do we select use cases?
We rate every idea for business value and feasibility and assign it a risk class under the EU AI Act. Cases qualify for a pilot with a value of at least 4, feasibility of at least 3.5 and no prohibited-practice classification. Example: value 4, data 3, technology 4, people 4 gives feasibility of 3.67, so the case goes on the shortlist.
BUSINESS VALUE →
Strategic initiatives
High value, low feasibility. Planned after the first pilots.
High value, low feasibility. Planned after the first pilots.
Pilots
High value, high feasibility.
High value, high feasibility.
Rejected
Low value, low feasibility.
Low value, low feasibility.
Team improvements
Low value, high feasibility.
FEASIBILITY (data, technology, people) →Low value, high feasibility.
What outputs does the board receive?
| Output | Contents |
|---|---|
| LEMUR index | A 0–100 score and an assessment of the five dimensions, with rationale and examples from the organisation |
| Use case portfolio | A value × feasibility matrix, risk class and charters for 2–3 recommended pilots |
| AI system register | Systems and tools with AI features, their owners and the company's role: provider or deployer |
| 90-day roadmap | Actions, owners, budget and progress measures |
How does the Diagnostic work?
The Diagnostic takes 2–3 weeks. Scope and price are agreed before the contract is signed and do not change during the work. It ends with a recommendation on next steps, which the board can carry out with us, on its own or with another partner.
- STAGE 1Interviews and dataInterviews with the board and 3–5 teams, a document review and data access within the scope agreed in the non-disclosure agreement.
- STAGE 2AnalysisAssessment using the LEMUR AI Readiness Index method, testing hypotheses against data, and initial findings discussed with the accountable person on the client side.
- STAGE 3RecommendationsA report, a 90-day roadmap and a board meeting at which we recommend a decision: pilot, programme, interim management or closing the topic.
How much time does this require from the client's team?
- A board-level sponsor
- A project lead (about 4 hours a week)
- IT, security and legal representatives for interviews
- Access to the tool list and vendor contracts
Which indicators show progress?
- Value of use cases in the portfolio (PLN/year)
- Share of pilots closed with a decision on time
- Number of AI solutions running in production with measured impact
- Share of employees covered by the AI skills programme
When is this service not the right choice?
- Companies that only need an integrator to implement a specific tool.
- Organisations that expect a formal legal opinion without preparatory work.
Related problems from the library
"We have several AI pilots and no results"Method: AI Use Case Portfolio + From Pilot to Scale
"Teams use AI without rules or training"Method: LEMUR AI Readiness Index
"We don't know which of our systems fall under the EU AI Act"Method: LEMUR AI Readiness Index + AI Use Case Portfolio
"The board expects AI to be implemented but has not defined the goal"Method: LEMUR AI Readiness Index + AI Use Case Portfolio
Materials in the Insights section
Guide · 4 min
AI literacy (Article 4 of the EU AI Act): a role-based training programme
AI literacy means the skills, knowledge and understanding that allow organisations to deploy AI systems in an informed way and to understand their opportunities and risks. An AI literacy programme turns these concepts into training matched to roles in the company. The article describes how to build a role-based programme, how to estimate its scale using a worked example, how to document activities and measure progress, and which mistakes to avoid.Read →Template · 4 minAI use policy in the company: elements, data classification and rollout
An AI use policy is an internal document that defines which AI tools employees may use, what data they may enter into them and who is responsible for the outputs. The article includes a list of policy elements, a data classification, an inventory summary template with a worked example, a four-step rollout plan and common mistakes.Read →Template · 4 minAI system register: fields, sample entries and risk classification
An AI system register is an internal list of the AI systems and features used or developed in the company, with the company's role and a preliminary risk class. It is the basis for determining which EU AI Act obligations apply to the company. The article includes a list of register fields, sample entries, how to gather the information, rules for determining the company's role and the register's limitations.Read →Guide · 4 minAI use cases: how to estimate value before a pilot
An AI use case is a specific application of AI in a single process, with a defined user, data and a measurable effect. Estimating its value lets you compare ideas and reject those that do not justify the investment. The article sets out formulas for three types of use case, a worked example, a sensitivity analysis, conversion to a 1–5 scale and common mistakes.Read →Analysis · 4 minEU AI Act timeline: obligations applying from 2025 and deadlines up to 2028
Some obligations under the EU AI Act already apply, and the deadlines for high-risk systems have been moved by Regulation 2026/1744 (Digital Omnibus). An overview of the deadlines, with the actions a company should complete before each of them.Read →Analysis · 4 minISO/IEC 42001 and NIST AI RMF: an AI management system in a mid-sized company
ISO/IEC 42001:2023 is an international standard that sets out requirements for an artificial intelligence management system, and the NIST AI RMF is a voluntary AI risk management framework published by the US NIST. The article compares the two documents, describes their components, shows a worked example of implementation scope and points out common mistakes.Read →Analysis · 4 minAI maturity model: MIT CISR and Gartner stages and how to assess a company
An AI maturity model describes the successive stages a company goes through, from first experiments to AI embedded in processes and decisions. The article compares the MIT CISR and Gartner models, shows a worked example of assessing a company across several areas and points out the limitations of such assessments.Read →Guide · 5 minAI implementation in a company: stages, costs and research findings
AI implementation in a company is a process in which the board links artificial intelligence to a business goal, selects use cases, prepares data, tests the solution in a pilot with a success measure and scales it under oversight that complies with the EU AI Act. We have gathered research from McKinsey, BCG, MIT NANDA and Eurostat on what sets successful implementations apart, together with the seven stages of the process, a sample business case with a payback period and common mistakes.Read →Guide · 6 minEU AI Act: what must the board do? A checklist by deadline
The EU AI Act, i.e. Regulation (EU) 2024/1689, imposes obligations on companies that develop or use artificial intelligence systems, and some of them already apply. We set out what applies now, what the deadlines of 2 December 2027 and 2 August 2028 will bring following the amendments made by Regulation 2026/1744, how the roles of provider and deployer are divided, what the penalties are and what a checklist with evidence of compliance looks like.Read →Frequently asked questions
- How do you implement AI in a company?
- AI implementation in a company starts with an assessment of organisational readiness (goals, skills, data, risk), not with choosing a tool. Next, you select use cases with measurable value, run a pilot with a success measure and a decision gate, and, before scaling, set rules for AI use and the obligations arising from the EU AI Act.
- Do you need a data science team to implement AI?
- Usually not at the pilot stage. Most first solutions are built on off-the-shelf models. Success depends on data, process and a business owner.
- Do you support the selection of technology vendors?
- Yes. We prepare selection criteria and SLA requirements before talks with vendors. We do not accept fees from vendors.
- Do you prepare legal opinions on the EU AI Act?
- We prepare the organisation: a register of systems, risk classification, policies and a skills programme. Legal opinions are prepared by a law firm working on the project.