# Getting stakeholders on board with an AI project

Share:[𝕏](https://twitter.com/intent/tweet?url=https%3A//consultancy.llmnet.nl/stakeholders-meekrijgen-bij-ai-projecten&text=Stakeholders%20meekrijgen%20bij%20een%20AI-project)[LinkedIn](https://www.linkedin.com/sharing/share-offsite/?url=https%3A//consultancy.llmnet.nl/stakeholders-meekrijgen-bij-ai-projecten)[Reddit](https://www.reddit.com/submit?url=https%3A//consultancy.llmnet.nl/stakeholders-meekrijgen-bij-ai-projecten&title=Stakeholders%20meekrijgen%20bij%20een%20AI-project)[Facebook](https://www.facebook.com/sharer/sharer.php?u=https%3A//consultancy.llmnet.nl/stakeholders-meekrijgen-bij-ai-projecten)[Copy link](#)

 
# Getting stakeholders on board with an AI project

 
 By Ivo Donker — compiled with AI support (Claude & Gemini)

 Introducing artificial intelligence within an SME organization is rarely just a technological challenge. AI touches the core of how people work, make decisions and collaborate. A successful implementation therefore stands or falls on getting all stakeholders on board in a timely, structured way. Without broad support, an AI initiative risks stalling in resistance, delays or poor adoption after delivery.

 To build support, you need to know exactly who your stakeholders are, which interests they defend and how to answer their specific questions. For broad methodologies and cultural change, we refer you to our guide on [change management with AI](/en/verandermanagement-bij-ai). In this article, we focus specifically on the stakeholder analysis and on developing a targeted communication strategy for each audience.

 
## 1. The key stakeholder groups and their questions

 Within an SME, not everyone views an AI project from the same perspective. What is a major opportunity for one department means an increase in risk or administrative burden for another group. To communicate effectively, you need to map out the primary concerns and priorities of each stakeholder group.

 
### Board of Directors and Executive Management

 The board looks at the organization-wide picture. They want to know how the investment contributes to the company's objectives and what the financial and operational risks are.

 
 
- Costs and ROI: What does implementation cost, what does it deliver, and when will the break-even point be reached?
 
- Risks: What about reputational damage, continuity and legal liability?
 
- Strategic logic: Does this strengthen our competitive position or solve a genuine bottleneck in the business process?
 

 
### Middle Management and Team Leads

 Middle management sits between strategy and execution. They are responsible for achieving operational goals with their teams.

 
 
- Capacity: How much time will the implementation process take from my people, and will it come at the expense of current targets?
 
- Disruption: Will the introduction of new software disrupt the daily workflow?
 
- Steerability: Do I remain sufficiently in control of the quality and output of my department?
 

 
### Operational Staff

 The employees who will work with the AI solution on a daily basis are looking at the practical impact on their own role.

 
 
- Job security: Will the technology take over my work, or will my role disappear over time?
 
- Autonomy: Will I still decide how I do my work, or will I be directed by an algorithm?
 
- Added value: Does the tool actually make my work easier, or does it only add extra administrative tasks?
 

 
### IT and Security

 IT specialists and security officers assess the technical feasibility and the impact on existing infrastructure.

 
 
- Integrations: How does the software integrate with our current systems and databases?
 
- Access and management: Who manages the access rights, and how much additional workload does this create for the service desk?
 
- Security: Where is the data stored, and how is unauthorized access prevented?
 

 
### Privacy and Compliance Roles (DPO / CISO)

 Privacy officers and compliance professionals assess the project against laws and regulations such as the GDPR and the AI Act. A thorough [AI risk assessment (DPIA)](/en/ai-risicoanalyse-dpia) is often a requirement for this group.

 
 
- Legal basis: On what legal basis are personal data processed?
 
- Data flows: Where does the data go, are models trained on our input, and is there any transfer outside the EU?
 

 
### Works Council (OR)

 The works council represents the interests of the staff and, with technological provisions that affect the work or the monitoring of employees, often has consent or advisory rights.

 
 
- Working conditions and workload: Does the job content, workload or physical/mental strain change?
 
- Monitoring: Is the software used to monitor the performance or attendance of individual employees?
 

 
## 2. Stakeholder analysis on two axes: Influence and Interest

 Not every stakeholder requires the same amount of attention and communication. A proven method for prioritizing your efforts is to categorize stakeholders in a matrix based on two axes: Influence (how much power they have to make the project succeed or block it) and Interest (how much impact the project has on their daily practice or responsibilities).

 
 
 
 
 Quadrant | 
 Characteristic | 
 Strategy and approach | 
 

 
 
 
 High influence / High interest | 
 Key figures (e.g. executive sponsor, head of IT, works council chair) | 
 Manage closely: Actively involve them in decision-making, provide weekly updates and ensure direct input. | 
 

 
 High influence / Low interest | 
 Resource managers (e.g. Financial Controller, Legal) | 
 Keep satisfied: Consult them at specific review moments and prevent them from being surprised by unforeseen decisions. | 
 

 
 Low influence / High interest | 
 End users (e.g. operational staff in the department) | 
 Keep informed & involved: Provide clear communication, training and listen to their practical feedback to ensure adoption. | 
 

 
 Low influence / Low interest | 
 Other departments (e.g. peripheral departments that barely notice the impact) | 
 Monitor: Inform them via general company channels (such as newsletters), but do not invest heavy communication effort. | 
 

 
 
 

 
 Key insight: The biggest pitfall in AI projects is that organizations ignore the 'low influence / high interest' group (the end users) in the early phase. Although they cannot formally stop the project, they can informally cause the implementation to fail by simply not using the system.

 

 
## 3. Choosing the right message and the right evidence

 Once you know who your stakeholders are, you need to tailor your message and supporting evidence to the recipient. A generic presentation for the whole organization rarely works.

 
 When convincing the board, abstract AI jargon is counterproductive. Use a business case with a clear ROI calculation, risk analyses and scenarios. For middle management, underpin the message with capacity calculations: show that the pilot does not lead to missing quarterly targets, but actually creates structural breathing room.

 For operational staff, figures about corporate profits are not convincing. They want to see demonstrations. Use short videos or live scenarios to show how the AI takes over boring, repetitive work, so that they have more time left for substantively more challenging work. Before the rollout, form multidisciplinary [AI adoption teams](/en/ai-adoptie-teams) with enthusiastic ambassadors from the shop floor who can demonstrate the benefits to their direct colleagues.

 
## 4. Treating resistance seriously instead of arguing it away

 Resistance to technological innovation is natural and often justified. Anyone who dismisses resistance as 'fear of change' misses valuable insights. To deal with objections effectively, you must distinguish between three categories:

 
### A. Substantive objections

 These are substantively grounded doubts about how the AI works. Think of comments such as: "The language model doesn't understand our specific customer situations" or "The output contains too many errors."

 Approach: Take these objections seriously. Use them as input for testing and refining the system. Validate the criticism by analyzing the tool's output together with the hands-on expert.

 
### B. Process objections

 Here, the objection is not about the technology, but about the way the project is being rolled out. For example: "We simply don't have time for this in the middle of peak season" or "IT is imposing another system on us without consultation."

 Approach: Adjust the schedule, or give departments the room to help determine the timing of the rollout within certain frameworks. Make clear agreements about capacity and support.

 
### C. Fear and emotional objections

 This concerns uncertainty about one's own position: "Will I become redundant over time?" or "Will I still be able to do this?"

 Approach: Do not argue emotions away with logic. Acknowledge the feeling and be honest about the project's objectives. Offer perspective by immediately investing in training and guidance. For their personal development, refer employees to external resources such as the learning platform [LLMnet Leren](https://leren.llmnet.nl/en/) to increase their knowledge of AI skills.

 
## 5. The importance of involving the shop floor early

 A common mistake is that AI solutions are conceived in an 'ivory tower' by IT and consultants and then thrown over the fence. The people who actually do the work possess the domain knowledge that is essential for an AI application to function properly.

 By involving employees from the design phase in formulating the use cases and testing the first prototypes, you achieve two things:

 
 
- The quality of the AI solution improves, because exceptions in the work process are recognized more quickly.
 
- Co-ownership develops. Employees defend a solution that they helped shape themselves.
 

 
## 6. The communication cadence during the project

 Communication around an AI project is not a one-off event, but a continuous process. Determine a fixed communication cadence in advance that evolves with the phases of the project:

 
### Phase 1: Before the start (The 'Why' phase)

 Focus on the context and the reason. Why is the organization exploring AI? What are the objectives and, just as importantly, what are explicitly not the objectives? Ensure that policy on responsible use is clear before the project starts. It is wise to lay down the frameworks in this phase through a formal [AI governance framework](/en/ai-governance-mkb).

 
### Phase 2: During the pilot (The 'Learn and Adjust' phase)

 Report openly on the pilot's progress. Do not only share the success stories, but also be transparent about what is not going well yet and how that is being resolved. This builds trust. Hold sessions where the shop floor can try out the tool in a safe environment.

 
### Phase 3: When scaling up (The 'Embed and Celebrate' phase)

 When the solution is rolled out broadly, communication shifts to practical support, sharing 'best practices' between departments and celebrating the first concrete results (such as time saved or increased customer satisfaction).

 
## 7. Pitfalls in getting stakeholders on board

 In practice, we regularly see the same mistakes recurring when building support for AI:

 
 
- Presenting AI as the miracle cure that solves everything: Exaggerated expectations inevitably lead to disappointment once the technology hits practical limits.
 
- Only approaching the Works Council once everything has already been decided: The works council feels bypassed and will formally slow down the process. Involve them from the first outlines.
 
- Letting IT carry the project alone: An AI project without strong representation from the business becomes a technical showcase that the organization ultimately gets little out of.
 
- Sending one generic communication update to the entire company: The board reads about details that don't interest them, while the shop floor misses the answers to their specific questions.
 

 
## Checklist: Is your organization ready for the rollout?

 Use the checklist below before sending communication about the AI project to the wider organization:

 
 
 
- [ ] Has a clear stakeholder analysis been created (Influence vs. Interest)?
 
- [ ] Is the sponsor from the board clear and actively involved?
 
- [ ] Have IT, Privacy/Legal and the Works Council been consulted at an early stage?
 
- [ ] Has a specific core message been drafted for each target group?
 
- [ ] Are key users from the shop floor involved in testing and configuration?
 
- [ ] Has a fixed communication cadence been agreed for all project phases?
 
- [ ] Is it clear where employees can go with substantive or emotional questions?