Service · AI development

AI consultant Netherlands.

Strategy and execution for organisations that want to get started with AI: vendor-independent, from mid-market to enterprise. We advise on use cases, select tooling, cover the AI Act and GDPR, and build alongside you where needed.

Vendor-independentAI Act & GDPRPilot to productionEU data residency

An AI consultant who also builds.

Most AI projects don't fail on the technology, but on the choices made before it: which use case first, which model, how it fits within the AI Act and GDPR, and how to measure success without it becoming an expensive playground. These questions call for a consultant who understands the difference between a RAG implementation and an agent, who can see past vendors' marketing claims, and who has written code themselves.

We combine strategy work with build capacity. We're not a pure adviser who hands implementation to another party and hopes the assumptions still hold, nor a build shop that skips the strategy work and comes up with a polished demo that never reaches production. We help SMEs with their first AI project, mid-market companies with scaling pilots to production, and enterprise CIOs with a portfolio approach and governance.

Our positioning is deliberately mid-market: we're more pragmatic than the Big Four and MBB strategy firms, and we're vendor-independent where Microsoft or AWS partners can't be. Working from our Amsterdam office, with EU data residency as standard, and at a pace that matches how your organisation really moves, not how a sales deck portrays it. For a large-scale enterprise AI implementation programme we fit this approach into a multi-year portfolio; for an SME client it may simply be one sharply defined use case.

What an Appfront AI consultant does.

Three types of engagement we're most often asked for. The right choice depends on where your organisation stands and how much room there is for experimentation.

Discovery & orientation

AI discovery and use-case prioritisation

For organisations that know AI is becoming relevant but are still looking for focus. We interview stakeholders from business, IT and compliance, map where AI could add value, and deliver a prioritised shortlist of use cases with an honest ROI estimate. This includes tooling advice, an AI Act impact scan per use case, and a clear view of what your organisation isn't ready for yet. We complete most discovery engagements within a few sprints.

Use-case mappingROI estimateTooling adviceAI Act scan
Strategy & pilot

AI strategy and pilot delivery

The next step once discovery is done: a concrete roadmap, a governance framework, and running the first pilot. We define the success criteria up front, build the pilot ourselves or support your team in doing so, and make sure the outcome is an honest go/no-go decision rather than a non-committal demo. This is the phase where you can draw on our experience with multi-LLM architectures, RAG on your own data, and setting up an evaluation set that doesn't get out of hand.

RoadmapGovernancePilot designBuild or co-build
Implementation & platform

AI implementation and platform engineering

For organisations that are past the pilot phase and want to embed AI as a lasting capability. Think of an LLM gateway for multiple departments, a RAG infrastructure on your own data, vendor-independent model routing, cost monitoring per team, and a governance layer that reassures the audit committee. For specialist use cases we also build custom, such as an AI agent that carries out tasks independently, or a document pipeline with human review.

LLM gatewayRAG infrastructureMulti-LLM routingMonitoring

What you walk away with.

An AI consultancy engagement has only succeeded if your organisation can carry on independently afterwards. These are the concrete deliverables we include as standard:

  • Prioritised use-case portfolioA shortlist with ROI estimate, complexity score and AI Act classification per use case, not a long list of wishes.
  • Tooling advice and vendor comparisonA well-founded choice between ChatGPT Teams, Claude for Enterprise, Microsoft Copilot, Gemini and open source, matched to your security requirements.
  • AI Act and GDPR overlayA working document that shows, per use case, which articles apply, which documentation is required and who owns which risk.
  • Pilot or production implementationBuilt by us or together with your team. Includes the codebase, monitoring, an evaluation set and a de-risked path to scaling up.
  • Workforce readiness planA concrete training programme for staff and management, linked to the Art. 4 AI literacy obligation under the AI Act.
  • Ongoing AI coaching (optional)A fractional CTO-AI role: a fixed number of hours per week or month to test strategic choices, join supplier conversations and keep the roadmap on track.

How we compare with other AI consultants.

The AI consultancy market is broad. A short orientation on where we do and don't fit: more honest than a comparison table with green ticks everywhere on our side.

Versus the Big Four

More pragmatic and mid-market

Deloitte, EY, KPMG and PwC bring scale and multidisciplinary teams, but their mid-market offering is usually expensive and tends to stall at the strategy phase. We work with smaller teams and shorter lead times, and we also handle execution.

Versus MBB strategy firms

We build too

McKinsey, BCG and Bain deliver excellent strategy, but no technical implementation. For AI that is a gap: without hands-on building experience, you can't realistically assess models, RAG architecture and evaluation sets.

Versus vendor consultants

No reseller interest

Microsoft AI advisers, AWS partners and OpenAI consultants have commercial interests in their own platform. We have no reseller agreement and choose the model that fits, even if that is open source and earns no margin.

Versus freelance AI consultants

Team continuity and building capacity

An experienced AI freelancer is a good choice for well-defined questions. For longer engagements or build capacity, organisations run into single-point-of-failure risks. We offer a team that can combine strategy and building.

Which organisations this suits.

Four situations in which our AI consultancy is usually a good fit. If you recognise yourself in one of them, we are probably a good partner for you.

SMEs with first AI questions

ChatGPT or something of your own?

You know colleagues are using ChatGPT, but you want control over data, costs and risks. We help you choose between a team licence, a secure enterprise version or your own solution on your data, including honest advice on when an expensive custom route isn't worth it yet.

Mid-market first project

The pilot needs to go into production

You have a working proof of concept, but now the hard questions arrive: security review, GDPR, change management, ownership. We bring structure to this phase without stifling the innovative spirit, and make sure the pilot gets an honest production test.

Enterprise CIO or CTO

An AI mandate without a portfolio

The board has named AI a priority, and you need to build a portfolio. We deliver the discovery, governance and multi-LLM platform architecture that this scale demands, alongside or in collaboration with the large strategy firms already advising the board.

Industry and trade bodies

Members want direction

You represent a sector where AI is moving in fast. We help develop sector-wide frameworks, set up knowledge sessions and write pragmatic guidance for your members, translated into the reality of their day-to-day work, not generic whitepaper language.

Executives and board members

Non-technical but still accountable

You need to make decisions about AI without being a developer yourself. We translate the choices into questions you can follow, shield you from vendor spin, and, where needed, build an AI steering group with you that can keep asking IT and the business the right questions.

HR and L&D managers

Making your AI training need concrete

Your organisation wants AI literacy and has a budget, but no idea where to start. We help you set up the training offering, develop role-specific learning paths and document AI Act Art. 4 compliance, all in a single engagement.

Not sure about a large project yet?

Test your idea first: a working prototype in 1 day

With OneDayBuild, we make your idea tangible in a single day for €1,150, so you know whether further development is worth the investment. Decide to go ahead with the full build? Then we deduct the cost in full.

View OneDayBuild →

How an AI consultancy engagement works.

1

Introduction and context

A conversation in which we understand where your organisation stands, what the board's question is and which risks or compliance topics have already been identified. No sales talk: we will also tell you honestly when you don't need a consultant yet, or which question needs answering first for an engagement to make sense. This conversation is non-binding and usually takes half an hour to three quarters of an hour.

2

Discovery and use-case portfolio

Stakeholder interviews with the business, IT and compliance, an AI maturity scan and a workshop to cluster and prioritise use cases. The result: a prioritised shortlist with an ROI estimate, complexity score and AI Act risk classification per use case. This is the foundation for a fitting AI strategy. We make sure the outcome is immediately usable, even if you press pause after this phase.

3

Roadmap and governance

We translate the shortlist into a phased roadmap, set up the governance framework (AI Act, GDPR, and where relevant DORA or NEN 7510) and align with IT, legal and the business owners. This includes vendor-independent tooling advice and an honest conversation about which use cases are better left alone. The governance layer stays light where it can and gets stricter where it must, depending on your sector and risk profile.

4

Pilot or first implementation

The first use case is carried out. We either build it ourselves or guide your team through the build. That means a measurable success criterion up front, an evaluation set to assess quality objectively, and an honest go/no-go at delivery. Often with an interim AI tech lead in the first sprints to safeguard the technical direction and bring your team along in the way of working.

5

Scaling up and ongoing support

Working use cases become part of a platform approach: LLM gateway, RAG infrastructure, monitoring and cost control per team. We stay involved as a fractional partner with a fixed number of hours per month, or we hand over completely to your team, whichever fits. The choice between the two is not an ego contest: for some organisations, full handover is precisely the success criterion of the engagement.

Compliance and regulatory context.

The legal context around AI is more mature in 2026 than two years ago, but still evolving. We help you navigate the regulatory landscape without unnecessary pessimism or false certainty.

The common thread: document what you do, classify your use cases, and make sure the people who carry out or review the work know why they are working with AI. The rest is sector-specific detail. In practice, we see that organisations with this foundation in place can start AI projects faster because the review moments are predictable. Not because the rules are lighter, but because they know what they need to demonstrate.

  • AI Act (EU Regulation 2024/1689)Phased entry into force since 2025. Prohibited practices already apply, followed by transparency requirements and the literacy obligation, and the high-risk regime continues to expand.
  • Art. 4 AI literacyEmployers must ensure that staff who work with AI have sufficient literacy. We include a concrete plan as standard in our engagement. See also our AI literacy training.
  • Art. 9-15 high-risk systemsApplications in HR, lending, education or public administration are subject to stricter requirements on risk management, data quality, logging and human oversight.
  • GDPR and UAVG overlapAI systems almost always process personal data. A DPIA per use case, a clear legal basis and purpose limitation are a standard part of our advice.
  • DORA for the financial sectorOperational resilience and third-party risk are becoming increasingly important when choosing a vendor. We map this out for banks, insurers and payment institutions.
  • NEN 7510 for healthcareHospitals, mental health organisations and health insurers are subject to the Dutch information security standard. AI applications must fit within that framework, including logging and authorisation.

Frequently asked questions.

The questions that executives, IT managers and compliance officers typically ask us in the first conversation.

What exactly does an AI consultant do?
An AI consultant helps your organisation make choices: which use cases are relevant, which tooling fits, how the AI Act and GDPR apply, and how you get from pilot to production. In our case, we add build capacity on top of that: we don't just advise, we can also implement, or direct an implementation partner with a clear brief. This means our advice is always realistic about what works in the build phase; we have solved this complexity ourselves before.
How does Appfront differ from the Big 4 or strategy consultants?
Big 4 firms (Deloitte, EY, KPMG, PwC) and MBB strategy consultants (McKinsey, BCG, Bain) bring scale and methodology, but their mid-market offering is typically expensive and often stalls at the strategy phase. We work more pragmatically, with a mid-market budget, and we also do execution: strategy and build from one team. For enterprise portfolios, we can operate alongside these firms, or be the implementation partner that gets their strategy into production.
How does this differ from a vendor consultant?
A Microsoft AI adviser, AWS partner or OpenAI consultant has a commercial interest in their own platform. We are vendor-independent: we build with ChatGPT Enterprise, Claude, Microsoft Copilot, Gemini or open-source models, depending on what fits your security requirements, data location and use case, not our partner programme. That gives you more honest advice and a better starting position if you want to switch models in a year's time.
How does an engagement start?
With a half-hour introductory call in which we understand where you stand. This is usually followed by a short discovery of a few sprints: interviews, a maturity scan and a use-case workshop. That produces a concrete proposal for a strategy or pilot. We don't do long preliminary engagements without a deliverable: every piece of advice we deliver stands on its own and is immediately usable, even if you choose to continue on your own after the discovery.
How do you deal with the AI Act?
The AI Act (EU Regulation 2024/1689) is being phased in. For most organisations, the key points are Art. 4 (AI literacy), the transparency requirements, and classification as high-risk or not. As standard, we deliver an AI Act overlay per use case: which articles apply, what documentation is required, and who is the owner. For sector-specific compliance, we also look at GDPR, DORA (financial) and NEN 7510 (healthcare). We make no promises of legal certainty: for formal advice, we work with your legal team or a specialised law firm.
What determines the cost of a programme?
Mainly the scope: a discovery with three stakeholder interviews is a different engagement from a full strategy-pilot-implementation chain. Other factors include the number of locations, the complexity of your existing systems, sector-specific compliance, and whether we deliver everything ourselves or co-build with your team. In the first conversation we outline a few options with corresponding sprint budgets. We work with sprint budgets rather than fixed-price contracts, because AI engagements contain too much uncertainty to pin everything down upfront, and doing so would more likely cost you money than save it.
Do you build it yourselves, or do you coach our own developers?
Both are possible. We build everything, from RAG infrastructure to agent development to an LLM gateway, for organisations without their own capacity. But we work just as often on a co-build basis: your team builds, and we provide architecture, code reviews and direction. For enterprise implementations the latter is usually the smarter route, because the capability has to land within your organisation and your people need to sustain it in the long term. For SME clients, full delivery is often more practical.
Are you really vendor-independent?
Yes. We have no reseller agreement with Anthropic, OpenAI, Microsoft or Azure. We work with all the major LLM providers and open-source models, depending on what fits your use case, data location and budget. If an open-source model in your VPC is the best choice, we will recommend it, even if a vendor consultant would leave margin on the table. For longer engagements we also look explicitly at exit scenarios: what if you want to switch vendors in two years, and how do you keep that option open?
Can this also be an ongoing role instead of a one-off engagement?
Yes. For organisations that want to use AI on a structural basis without immediately hiring their own senior AI lead, we provide a fractional CTO-AI role or an interim AI tech lead. A fixed number of hours per week or month, in which we review strategic decisions, attend supplier meetings, keep an eye on the roadmap and coach your own team. Often a natural follow-on from an implementation engagement.

Talk to us about your AI challenge.

A no-obligation introductory call of half an hour. We listen to where your organisation stands, ask questions that make the right next step clear, and tell you honestly if you don't need a consultant yet. Call or email Fabian directly at fabian.vandijk@appfront.nl or book via the contact form.

Edit content