Four filters. One principle: your AI should think like your organization.
ValuesFilters is a system — not a single product. Each filter addresses a different dimension of the same problem: AI that doesn’t know your context makes decisions you can’t predict. Together, they give your AI organizational judgment.
VALUESFILTER
ContextPrompt
Your AI’s operating context
The problem it solves
You’ve given your team access to ChatGPT, Copilot, Claude — maybe all three. Now 50 people are asking AI questions with 50 different frames of reference. The AI gives 50 different answers. Some are good. Some contradict your values. Most are generic.
What ContextPrompt does
We extract the decision-making patterns, values, and expertise that define how your organization actually works. We code it into a Living Book — 120 pages of narrative knowledge. Then we turn that into a prompt system that every AI tool can use.
What changes in practice
Ask any AI tool a question about how to handle a client situation, and it responds the way your best project lead would. Not because it’s been restricted — because it understands the context. Same answer whether it’s a new hire or a senior partner asking.
The delivery
- A Living Book (120 pages) — your organization’s coded knowledge
- A ContextPrompt — optimized prompt system for daily AI use
- Implementation across your AI tools
- Training for your team (15 min per person)
From €35,000 · 4–6 weeks · Your team’s time: ~8 hours total
Optional Living Book retainer — quarterly updates, new chapters, continuous alignment. From €8,000/quarter.
How it works
1. Values discovery. We identify the values your people actually live by. Validated in a workshop with key staff.
Time required from you: Three hours per person
2. Tacit knowledge capture. We surface the judgment experienced employees apply intuitively. On their terms, with minimal disruption.
Time required from you: Three hours per person
3. Living Book. ~120 pages in narrative format. Research shows stories increase sound judgment by 24% over bullet-point rules.
4. Always-on prompt. The book serves as a live prompt for any AI tool your staff uses. ChatGPT, Copilot, Claude, or your own system.
Time required from you to implement the ContextPrompt in your workflow: Fifteen minutes per person
Best for: Departments within larger organisations and companies with up to 100 employees who need a clear, values-aligned AI baseline without a full-scale implementation.
VALUESFILTER
ValuesFirst
Governance that works, not just documents
The problem it solves
EU AI Act is here. You need to demonstrate that your AI decisions are traceable, explainable, and aligned with documented values. Most organizations are either ignoring this or producing paperwork that nobody reads.
What ValuesFirst does
We build a governance framework that’s alive — embedded in how your AI actually operates, not filed in a compliance folder. Your values become the operating logic. Audit trails emerge from actual use, not retrospective documentation.
What changes in practice
When an auditor asks “how does your AI make decisions?” you don’t scramble. You show them the system. Values in, decisions out, trail visible. Every AI interaction traceable to your documented principles.
The delivery
- Complete AI values framework mapped to EU AI Act requirements
- ValuesFirst integration with your AI infrastructure
- Audit-ready documentation that stays current
- Quarterly review protocol
From €65,000 · 6–8 weeks
Included quarterly review. Extended governance retainer available.
How it works
1. EU AI Act mapping. Map your AI systems against Article 4 literacy requirements and high-risk thresholds.
2. Values extraction. Extract governance values from your organization and AI processes.
3. Framework build. Create living governance framework with audit trails.
4. Quarterly reviews. Adapt governance as laws, tools, and risks change.
VALUESFILTER
Tacit Knowledge
Expertise that stays when people leave
The problem it solves
Your best people carry 10–30 years of judgment in their heads. When they retire, change roles, or leave — that knowledge walks out the door. This knowledge has never been written down — because it can’t be, not in traditional formats.
What Tacit Knowledge does
Using Cognitive Task Analysis — the same methodology developed for capturing expert decision-making in high-stakes environments — we interview your key people and extract knowledge they don’t even know they have. The result is coded into formats that both humans and AI can use.
What changes in practice
New hires access the judgment that used to take years to absorb. AI tools draw on real organizational expertise instead of generic training data. When someone retires, their contribution lives on — not as a farewell presentation, but as operational infrastructure.
The delivery
- Structured CTA interviews with key personnel
- Coded knowledge base (narrative + structured format)
- Integration with AI context systems
- Knowledge validation with subject matter experts
From €65,000 · 6–10 weeks
Additional knowledge captures as needed — when key people transition, new expertise emerges, or organizational knowledge shifts.
How it works
1. Selection. Identify the knowledge holders and decision domains that matter most.
2. CTA interviews. Walk through specific decisions step-by-step using Cognitive Task Analysis.
3. Validation. Cross-check findings with other experts and organizational data.
4. Codification. Turn expertise into Living Book chapters and AI-ready formats.
VALUESFILTER
Raise-AI
Competence, not compliance theater
The problem it solves
EU AI Act Article 4 requires AI literacy for everyone in your organization. Most training programs check a box. Raise-AI builds actual competence: your people learn to work with AI using your own values, your own context, your own decision patterns. Not generic slides — your reality.
What Raise-AI does
We train your team using your own context — your values, your Living Book, your actual AI tools. Not hypothetical scenarios, but real situations from your organization. People learn by doing, guided by the same principles that govern your AI.
What changes in practice
Your team doesn’t just “know about AI” — they know how to use it in a way that’s consistent with how your organization works. The gap between “AI policy” and “how people actually use AI” closes.
The delivery
- Customized training program based on your organization’s context
- Hands-on workshops with your actual AI tools and ContextPrompt
- Competence assessment aligned with Article 4 requirements
- Train-the-trainer module for internal scaling
From €15,000 · 2–4 weeks
How it works
1. Curriculum design. Build training around your values, not generic AI principles.
2. Live workshops. Two days with real scenarios from your organization.
3. Hands-on practice. Your actual AI tools, your ContextPrompt, your situations.
4. Certification. Article 4 competency assessment + train-the-trainer module.
Better together — but start anywhere
Most clients start with ContextPrompt. It’s the foundation — once your AI has organizational context, everything else gets easier.
But you don’t have to do everything at once.
Start with what matters most: If your biggest pain is inconsistent AI answers, start with ContextPrompt. If your biggest pain is regulatory pressure, start with ValuesFirst. If your biggest pain is knowledge leaving the building, start with Tacit Knowledge. If your biggest pain is people not knowing how to use AI, start with Raise-AI.
Each filter works independently. Together, they create something none of them achieves alone: an AI that genuinely operates as part of your organization.
A living relationship, not a one-time project
Organizations change. Regulations evolve. People come and go. Your AI context should keep pace.
That’s why every ValuesFilter can include an ongoing retainer. We update your Living Book with new chapters when policies shift, capture knowledge from departing experts before they leave, and ensure your AI context stays aligned with your reality.
The initial engagement builds the foundation. The retainer keeps it alive.
Investment
| Filter | Starting from | Timeline | Ongoing option |
|---|---|---|---|
| ContextPrompt | €35,000 | 4–6 weeks | €8,000/quarter |
| ValuesFirst | €65,000 | 6–8 weeks | Included review |
| Tacit Knowledge | €65,000 | 6–10 weeks | Per engagement |
| Raise-AI | €15,000 | 2–4 weeks | — |
Final investment depends on organization size, complexity, and number of people interviewed. We scope every engagement individually after an initial conversation.
For context: comparable AI strategy engagements at major consulting firms range from €75,000 to €250,000 — typically delivering reports. We deliver operational infrastructure your team uses every day.
Ready to connect your knowledge to your AI?
Frequently asked questions
What is ContextPrompt and who is it designed for?
ContextPrompt captures your organization’s values, judgment, and decision patterns into a Living Book and prompt system. Designed for departments and organizations with up to 100 employees who want consistent, values-aligned AI responses across all tools.
How much does ContextPrompt cost and how long does it take?
From €35,000. Timeline: 4–6 weeks. Your team’s time: approximately 8 hours total. Optional quarterly retainer from €8,000 for ongoing updates.
What is ValuesFirst and what EU AI Act articles does it address?
ValuesFirst creates a living governance framework that embeds your values into AI operations. It addresses Article 4 (AI literacy), high-risk system requirements, and audit-trail obligations. From €65,000.
Does Raise-AI satisfy EU AI Act Article 4 requirements?
Yes. Raise-AI is specifically designed to build AI literacy competence as required by Article 4. Unlike generic training, it uses your organization’s actual context and values. From €15,000.
What is Tacit Knowledge mapping?
Tacit Knowledge uses Cognitive Task Analysis to extract the judgment and expertise your key people carry in their heads. This knowledge is coded into formats both humans and AI can use, ensuring it stays in the organization.
Can I use just one ValuesFilter, or do I need all four?
Each ValuesFilter works independently. Most organizations start with ContextPrompt as it builds the foundation. You can add other filters as needs evolve.