AI by department · the floor plan
AI by department: where AI takes over work in your business
Customer service, sales, marketing, HR, finance, operations and procurement. For each department we show which work AI takes over, where a person decides and what it delivered in real projects. Click through the floor plan or go straight to your department.
- 160+ organizations helped
- Seven departments, one approach
- A person always decides
80%
of 6,000 customer questions a month with a prepared answer
1 hour
per sales proposal, where it used to take 4 hours
12 h
a week saved on invoice processing
192 h
a month back on request processing at Van Berkel
01 / The AI floor plan
Where does AI fit in your business?
Click through the office. For each department you see which work AI takes over and which role the person keeps: fully AI, human approves or AI assists. The figures come from real projects.
Click a department
What AI takes over here
Customer service
- Draft a reply with the order detailsHuman approves
- Label and route ticketsFully AI
- Call the difficult customerAI assists
In practice
30% → 80%
of customer questions get a prepared reply (enterprise client)
02 / The departments
AI by department: every team its own kind of work.
For each department: which work AI takes over, where the person decides and what it delivered in practice.
Worked with us
03 / The pattern
What do AI applications across departments have in common?
More than you'd think. In every department, the work AI takes over sits in the same three places. Something comes in and someone retypes it: a customer question, a request, an invoice, a CV. Someone looks up information in three systems to give one answer. And someone produces a document or report that is almost the same every time.
The difference per department is in the systems, the language and above all in where the person decides. In finance, invoice processing can eventually move almost entirely to AI, with spot checks. In customer service, an employee approves every answer. In HR, a person always decides about people. We draw that line per task, using the four roles from our APO method: human, AI assists, human approves and fully AI.
That's why we never start with the technology, but with the process. Process first, technology later. Three numbers per step: how long it takes, with how many people and how often. Then you know where AI delivers the most, and which department to start in.
01Retyping
Customer questions, requests, orders, invoices, CVs. In every department.
02Looking things up
Three systems open for one answer.
03Documents
Quotes, reports, minutes. Almost the same every time.
04Where the person decides
Decided per task, not per department.
04 / In one table
AI by department at a glance.
Per department: a logical first step, where the person decides and what it delivered in practice.
| A logical first step | Where the person decides | In practice | |
|---|---|---|---|
| Customer service | Preparing answers | Employee checks and sends | 80% of 6,000 questions a month prepared |
| Sales | Requests into the CRM, draft quote | Price, discount and customer contact | Sales proposal from 4 hours to 1 hour |
| Marketing | Product copy, reports | What goes live | EasyToys: 132 hours a month |
| HR | HR help desk, interview scheduling | Always, about candidates and employees | HR consultancy: 65% faster screening |
| Finance | Reading and matching invoices | Bookings above a threshold and every payment | 12 hours a week on invoice processing |
| Operations | Orders and requests into the ERP | Planning and exceptions | Van Berkel: 192 hours a month |
Swipe sideways to see all columns →
The figures come from projects with our clients. What it delivers for you, we calculate with your own numbers.
05 / Where do you start?
Which department should you start with AI in?
In the department with the most repetitive work, not the one that shouts loudest. Three questions help you choose:
- 01Where is the volume? A task that comes back a hundred times a day delivers more than one that happens once a quarter.
- 02Is the input digital? Email, PDF, form or system works most easily.
- 03Do you know when it's right? Then a person can check quickly, and you build trust.
In practice we often end up in customer service, the order desk or invoice processing. Not because other departments can't, but because that's where the volume is highest and the gain shows fastest. The rest follows, process by process.
Not sure? Take the free self-scan below, or start with the AI Quick Scan. Don't let it drive you crazy. Start small.
01Volume
Dozens to thousands of times a month.
02Digital input
Email, PDF, form or system.
03Clear outcome
You know when it's right.
04Examples
Without examples of good work, we calculate with 40% confidence at most.
06 / Check for yourself first
Where does your organization stand with AI?
Seven questions, two minutes, free. You see straight away where you stand and what a logical next step is.
Free AI self-scan
1 = nothing in place · 5 = fully sorted
Strategy and Vision
Is there a shared AI vision and a mandate at board level?
Data readiness
Is data central, structured and accessible?
Technology and Infrastructure
How easily do you connect new systems to existing ones?
People and Skills
Does the team know AI tools, and is there room to learn?
Processes and Adoption
Are processes documented and actively optimised?
Governance and Risk
Is there policy for AI use, privacy and security?
Culture and Change
Is there room to experiment, to fail and to learn?
3.0
your score
4.0
target · after 12 months
Start here
Strategy and Vision
your lowest score, so the biggest gain
You have loose pieces in place
Enough to start with, too little to build on. The gain is in tying together what is already there.
Free, no strings
Send us your profile and get an analysis back
- ✓Which of the seven dimensions we would tackle first, and why that one
- ✓Two examples of what that delivered at a comparable organisation
- ✓Whether a single workshop makes more sense here than a partnership, or the other way round
Your mail app opens with your seven scores in it, addressed to noud@nina-ai.nl. You send it yourself, so we store nothing until you hit send. Reply within 24 hours, from a human.
Rather look around first?
Indicative and not stored: we do not keep your answers. The real baseline we do on site in the AI Quick Scan, together with your team.
07 / Another angle?
Rather look by sector or by agent?
An invoice is an invoice, but an online store works differently from a factory. Look by sector too, or at what AI agents do in practice.
08 / How we approach it
One approach, department by department.
We build within the AI Partnership: a dedicated team that builds, measures and maintains process by process, across all departments.
- 01
AI Quick Scan
One day plus a week
One day on site: a baseline on 7 points, and the processes of one department, two at most, mapped out. Within a week a business case per opportunity and a twelve-month roadmap. € 6,000.
- 02
First process live
Quarter 1
The process with the best business case, with a human in the loop. Often customer service, orders or invoices.
- 03
Rolling out to more departments
After that
With package S we work on one process at a time, with M on two and with L on four. Every month you see what it delivers in hours and euros.
09 / Cases
What it delivers per department.
10 / Questions
Frequently asked questions about AI by department.
Which department benefits most from AI?
The department with the most repetitive work: high volume, digital input and an outcome you can check. In practice that's often customer service, the order desk or invoice processing. At an enterprise client of NinA AI, 80 percent of 6,000 customer questions a month have a prepared answer; at Van Berkel Professionals, request processing delivers 192 hours a month.
Can AI be used in several departments at once?
Yes. Within the AI Partnership, package S works on one process at a time, package M on two and package L on four, across the organization. We start with the process with the best business case and then move on to the next department. That keeps it manageable, and you see what each step delivers.
Do you need a different AI tool for each department?
No. We build on the systems you already have, such as Exact, AFAS, SAP, Microsoft Dynamics, HubSpot, Salesforce, Microsoft 365 and Google Workspace, usually with n8n in between and custom work where needed. The language model is interchangeable. Each department gets its own workflows, on one foundation.
Where does the person decide, and where does AI do it alone?
We decide that per task with the four roles from the APO method: human, AI assists, human approves and fully AI. Most tasks land on 'human approves'. Invoice processing can eventually move to fully AI, with spot checks. Decisions about people, as in HR, and the conversation with an angry customer always stay with a person.
What does AI by department cost at NinA AI?
You start with the AI Quick Scan at € 6,000. After that the AI Partnership runs through package S (€ 4,500 per month, 30 hours), M (€ 7,500 per month, 50 hours) or L (€ 12,000 per month, 80 hours), for a minimum of twelve months. The investment in year one, including the Quick Scan, is € 60,000, € 96,000 or € 150,000. All amounts excluding VAT. We only start if the partnership as a whole delivers at least twice the investment in the first year.
How do you get employees on board with AI in their department?
Which department will you start with?
Book an intro call. In fifteen minutes we look together at where the most repetitive work is, and what a logical first step would be.
01
Intro call
Fifteen minutes, no strings. Where does work get stuck and what do you want to achieve?
02
AI Quick Scan
One day on site. Within a week, a business case and a roadmap.
03
AI Partnership
A dedicated team that builds, trains and shows what it delivers every month.
160+ organizations · Sessions rated 9.3 · Based in Amsterdam