• Nederlands

Service

AI consultancy

Advice on AI from the people who build the model afterwards and keep it in production: since 2014, from Eindhoven, for organisations across the Netherlands.

Schedule an introduction →
Workbench next to a production line with a tablet on a stand and tools, with one teal toolbox

AI consultancy at Twentynext starts with the question you want answered, not with the model. In eight advisory days, spread over two to six weeks, we determine whether your data can carry it, what a working solution costs and who runs it afterwards. That exploration costs € 7,500. The consultants help build it and stay involved as long as the system runs. Since 2014, across the Netherlands.

What it is about

What does an AI consultant at Twentynext do? They start with the decision that needs to improve and test three things: can the data carry it, does someone need the result, and who watches the model once it runs.

AI agency, AI implementation partner or AI consultancy? Whoever searches for an AI agency usually wants the same thing: someone who says whether it can be done, builds it and does not let go. That is what we do, in that order.

What question do you bring? A pilot that succeeded and then stalled. A Copilot rollout where nobody knows what it may see. Or whether AI adds anything at all. We advise the data foundation as often as a model.

How we work

Strategic and technical, by the same people. The AI consultant who assesses feasibility also designs the architecture and knows from projects where models in production break: on permissions, on data, rarely on the algorithm. Boston Consulting Group attributes roughly ten per cent of the effort of scaling AI to the algorithm and seventy per cent to processes and people; that matches what we see.

No staff gallery, but more than twenty specialists on permanent contracts, with MIT-funded research as a hinterland: that is how we know of a technique where it works and where it stops working.

We do not build before there is an owner, we test in the real work process, and we agree one measure up front.

And afterwards: support

A model that goes live is not finished. The reality around it changes, and someone has to notice when the model notices. So we run what we advise and build: monitoring and incidents, changes and further development, and technical and application support from pipelines to the application layer. Every agreement is written down, with response and resolution times per contract in an SLA.

Some systems have been under our management for more than five years; the replenishment system for fresh produce, with a demand forecast per product per store, has been running since 2020. We also take on a model someone else built, or an app you put together with AI yourself. Read more under Service & Maintenance.

Proof

Where this has been done

Maritime · with MasterShip Software · since 2020, in use

AI for lightweight construction

With MasterShip Software, an AutoCAD module that automates design tasks in shipbuilding using machine learning and genetic algorithms. In a practical case MasterShip cut total engineering and replacement time by 40 per cent. In use since 2020.

Read the case

Ecology · R&D · since 2018

Population dynamics with Wageningen University & Research

A platform that recognises mammal species in observation data with an accuracy above 98 per cent, counts populations and tracks movement patterns. Started in 2018 as an MIT Zuid feasibility project, with WUR and the Belgian government. Interpretation stays with the ecologist.

Read the case

Retail · since 2020 · under our management

Replenishment system for fresh produce

A Bayesian model that ties demand, sales and ordering parameters together, even on sell-out days. Waste fell by 7 per cent; on one product across ten stores, profit rose 4 to 7 per cent. Running since 2020, under our management.

Read the case

Two chairs at a bright table with a notepad and a diagram
Meeting corner with an open report and a cup of coffee
Whiteboard wall with a diagram of connected blocks
The services we offer

What else can we do for your organisation?

A few examples of services with which we can support your organisation:

Data-driven working
Data architecture
Service and support
Read more
Data and reporting platform
Microsoft Copilot implementation
Research and development
Read more

Frequently asked questions

What does AI consultancy cost?

A first conversation comes with no obligation. After that it starts with an exploration of eight advisory days, spread over two to six weeks: can your data carry it, what does a working solution cost, who runs it. That exploration costs € 7,500 excluding VAT. Prefer a single day first? A start day costs € 1,000 and delivers a prioritised list and our advice on whether the exploration is worth it. What a model in production costs, you know after the exploration, and sometimes the outcome is that the data foundation comes first.

How long does an AI implementation take?

The exploration of two to six weeks tells you whether it can be done. After that a project goes through six steps from scope to support: scope, data exploration, build, testing in the real work process, delivery on your systems, support. How long that takes depends mainly on the data and on the people who will work with it, not on the model. We set the measuring point at twelve months.

What is the difference between data and AI consultancy?

Data consultancy is about the foundation: which data exists, whether it is reliable and how it reaches the people who decide. AI consultancy is about whether a model adds anything, what it costs to get it into production and keep it there, and who is responsible once it runs. In practice an AI question often starts at the foundation; then we say so.

Outsource AI or do it yourself?

Building a first model can be outsourced; ownership cannot. Someone in your organisation has to need the result and steer on it, otherwise the pilot stalls the moment we leave. We build and run it, or place specialists in your team until your own people can carry it. What you keep yourself: the decision the model serves.

Which consultancy has experience with AI in my sector?

Twentynext has worked since 2014 for industry and maritime, healthcare and life sciences, security, retail and ecology. By name: the AutoCAD module with MasterShip Software, the population dynamics platform with Wageningen University & Research, the crime room for DITSS and the data platform for AxionContinu. The eighteen cases are grouped by question, because the same technique works in more than one sector.

Is there a subsidy for the exploration?

For SMEs, regional schemes exist that cover part of external advisory costs, such as the North Brabant scheme Datamaturiteit MKB via the data coaches of Brainport voor MKB, Midpoint Brabant and REWIN, or the Groeiversneller of Oost NL for growth companies in Gelderland. Whether and how much is covered depends on your location, sector and the available budget; the Brabant scheme had reached its ceiling in September 2026. In the first conversation we look together at which route fits. You submit the application yourself; we supply the quote and the advisory report.

Is your organisation ready for AI?

Take the free Data Maturity Scan: sixteen questions, seven minutes, instant result, no e-mail address needed. You see where it gets stuck before you talk to us.

Take the free Data Maturity Scan Book 30 minutes with Martijn van Grieken

We respond within one working day.

Work with us

Realise your project together?

The people who build it also run it afterwards. Eindhoven, since 2014.

Martijn van Grieken

Martijn van Grieken

Director Data & AI

Get in touch