Approvals that pile up before month-end
Staff spend the first week of every month chasing sign-offs by phone and email.
AI agents and custom software from Utrecht
An AI agent in one day. Software in three days.
Bring us one process or one product idea. On Loom, our agent platform, an AI agent runs your process by the end of the day, with a person approving every action. In our AI software factory, a working first version of your product (an MVP) is in your hands in three days.
What the agent connects
Agent intelligence
Built for
The work between your systems.
Every organisation has work that lives between its applications. Someone chases it, retypes it or has to read it first. That is the work we hand to agents.
A weekly file that takes a week
Revenue statements retyped, split by hand and entered into the accounting system, store by store.
Documents someone has to read first
A diploma, an insurance certificate or an invoice waits for a person before the process can continue.
Pilots that never reach production
A chat demo impresses, but nothing connects it to the systems where the work actually happens.
Before, at FAIR
- Staff chased approvals in the first week of the month
- Hours were corrected by hand
- Withdrawals were noticed late
Now
- Agents check, remind and escalate around the clock
- Staff review a short list of exceptions
- Every action is in the log
One shift, from start to finish
One shift, from start to finish.
A shift ends. The agent reads it.
Twenty-four hours after a shift ends, the agent reads it from the scheduling system and checks the hours and breaks against the rules. A shift longer than 5.5 hours needs at least 30 minutes of break.
Three organisations. Three kinds of work.
Agents on top of an existing platform, integrations around a screening product, and an invoicing engine that replaced a spreadsheet.
FAIR
Verify
Follow up
Guard
Insight
Customer case / AI agents
Less chasing. More overview.
Four agents that check shifts, chase approvals and handle withdrawals for a staffing agency in healthcare. People handle only the exceptions.
Inside the case
- Context
- FAIR (Vitalis HR Services) places self-employed care professionals with care organisations and invoices monthly. Every worked shift needs the professional’s approval and the care organisation’s approval before the 7th of the following month.
- The problem
- Staff spent the first week of each month chasing approvals and correcting hours by hand. Withdrawals were noticed late.
- What we built
- As part of the development team, we built four agents on top of the existing Enabl platform, through its API. Verify checks hours and breaks and approves correct shifts. Flow sends shifts to the care organisation and keeps reminding until the deadline. Guard asks for a reason when a professional withdraws from a shift and escalates from the third withdrawal in a year. Insight is the portal where staff see today’s numbers, ask questions in plain language and handle escalations.
- How it runs safely
- Each agent started in signal mode (decide and record, change nothing) and was switched to action mode one agent at a time, after acceptance. Enabl stays the source of truth. Every write is recorded with the agent as the actor and cannot be applied twice.
- What changed
- Fewer manual corrections around the invoicing batch, faster approvals by the professionals, and a first week of the month that is no longer a peak. Staff see the status live instead of asking around.
Otentica
Identity
Qualifications
Verification flow
HR systems
Customer case / Integrations
Many checks. One connected process.
The integration partner behind a screening platform used by more than a hundred organisations for over 850,000 screenings.
Inside the case
- Context
- Otentica is the pre-employment screening platform for VOG, identity, diploma and reference checks. More than a hundred organisations use it, over 850,000 screenings have run through it, and it is ISO 27001 certified.
- The problem
- Screening lives between systems. Recruiters start it from their HR or recruitment software, candidates complete it on their phones, and the results have to land back where the hiring decision is made.
- What we built
- Rutron is Otentica’s integration partner. We build the integrations that connect screening to HR and recruitment systems, and we provide a development team that extends the platform with new checks and features.
- What changed
- Screening progress flows between the recruiter’s system and Otentica without retyping, and new checks and integrations are delivered by a team that knows the platform.
Finoptis
Weekly statement
Allocation keys
Three invoice streams
Review
Customer case / Custom software
One weekly file. Three invoice streams. No manual work.
A revenue-allocation and invoicing application for an accounting firm that administers shop-in-shop concessions in two national supermarket chains.
Inside the case
- Context
- Every week each chain delivers a revenue statement per store, split by product group. That revenue has to be divided among three parties, the retailer, the operator on the shop floor and the concept owner, according to contractual allocation keys with tiered percentages. The result becomes sales invoices and self-billing invoices in the accounting system, across two administrations.
- The problem
- The whole process was manual. Split the revenue per store and product group with the known percentages, create three documents per store, check everything by eye. Error-prone and slow, every single week.
- What we built
- An application that imports and validates the weekly statements, keeps versioned allocation keys with start dates and tiers per store, runs the allocation with a record of the percentages actually applied, creates draft invoices through the accounting system’s API in the right administration, and gives the accountant a run overview to check and finalise. Advisor fees are calculated in the same run.
- What changed
- A weekly process that cost a person a full day of spreadsheet work is now an upload and a review.
Our platform
Loom
Loom. Where your agents run.
Loom is the platform we build customer agents on. Connect the APIs you already have, draw the workflow or describe it, give an agent its instructions and tools, put a person at the decision points, and follow every run step by step. That is why one process becomes a running agent in a single day. You get your own login. Change workflows yourself, or let our team run everything for you.
Connectors
Import an OpenAPI or Power Platform definition, or describe an endpoint by hand. API keys, bearer, basic and OAuth2. Every endpoint becomes a step in a workflow and a tool for an agent.
Workflows
Draw the process, or describe it in a sentence and let Loom draft it for you to review. Start it from a webhook, a schedule, a poll or by hand. Branch on conditions and test every step with real data.
Agents
Give an agent its instructions, a model of your choice, the tools it may use and a fixed output the next step can rely on. Optional access to your knowledge base.
Approvals
A pause in the workflow where a person reviews, edits and approves or rejects, with an optional time limit. The run waits as long as it needs to and continues where it stopped.
Runs
An inbox that puts “needs your approval” first. For every run you see the graph, the timeline, the agent’s conversation and every tool call.
Knowledge and LoomPilot
A knowledge base your agents can search, a playground to test what they will find, and LoomPilot, the assistant that answers questions about your workflows and runs and pins the answer as a live chart on a page.
An agent in one dayYour own loginDo it yourself or fully managedRuns as a service in our private cloud
Three ways to start.
Pick the one that matches what you bring. A process, a product idea, or a system that has to last for years.
1 day
An AI agent in one day
On Loom. From a simple check to a process with several steps, API calls, AI decisions and human approval. Running by the end of the day, with a person approving each action until you decide otherwise. Manage it yourself or let us manage it.
How the day goes3 days
Software in three days
In our AI software factory. We gather the requirements with you, set up the architecture, and let AI coding agents build the software under our engineers’ review. Screens, integrations and data connected. A working first version of your product (an MVP) in your hands on day three.
How the three days goweeks
A platform that lasts
When your organisation needs a system of its own. Multi-agent platforms, screening integrations, revenue and invoicing engines, and the APIs and data mappings between them. Weeks rather than days. See FAIR, Otentica and Finoptis.
See the casesAutonomous. And everything on record.
Useful automation needs clear responsibility. Before an agent changes anything, we agree what it may read, what it may change and when a person decides. Then we build those choices into the workflow.
Signal mode first
An agent decides and records before it is allowed to change anything.
A named actor
Every write carries the agent’s identity in the audit log of your system.
Approval where it matters
In Loom, a human decision is a step in the workflow. The run waits for it.
A record you can open
Every Loom run keeps its graph, its timeline, the agent’s conversation and every tool call.
Your system stays the source of truth
Agents read through the API and write back only what was agreed.
Your own keys
Model access runs through your own API key, so usage and data stay under your account.
The agent day
One process. One day. A running agent.
Bring one process and access to the system it lives in. By the end of the day an agent runs it on Loom, from a simple check to a process with several steps, API calls, AI decisions and a person at the decision points. You watch its first real runs before you go home.
| When | What we do together | What you have |
|---|---|---|
| Beforehand, 30 minutes | Pick the process, agree who decides, arrange access to the API | A one-page brief |
| Morning | Map the steps and the exceptions. Import the API into Loom and connect it | Your system connected to Loom |
| Afternoon | Build the workflow. Give the agent its instructions and tools. Add the approval steps. Test with real data | A running agent with a person approving each action |
| End of the day | Watch the first real runs together. Decide what to fine-tune | The agent in your process, every run on record |
| Afterwards | Fine-tune. Decide what may run without approval. Manage it yourself, or we do | An agent in production |
One process. A system with an API or an OpenAPI description. Access arranged before the day.
The AI software factory
A product idea on Monday. A working first version on Thursday.
In our software factory, AI coding agents write the code. Our engineers sit down with you, gather the requirements, write the product requirements document, cut it to what fits in three days, set up the architecture and review every piece. The factory is tuned so that a working first version of your product, an MVP with screens, integrations and data connected, is delivered in three days.
| Step | What we do | What you have |
|---|---|---|
| Requirements | We sit down with you, gather everything the first version needs and write the product requirements document | A requirements document you have approved |
| Scope | We cut the document down to what fits in three days and agree what comes after | A scope you can hold us to |
| Architecture | We set up the application’s architecture, the foundation the coding agents build on | A deployed skeleton with the right shape |
| Build | AI coding agents build the screens, the integrations and the data layer. Our engineers review every piece | A working product taking shape on a URL |
| Delivery | Everything connected and handed over, with a plan for what comes next | A working first version in your hands |
One product. The core flow from start to finish, connected. Certifications and multi-tenant hardening come in the weeks after, not inside the three days.
And when your organisation needs a system of its own.
Some work is a platform, not a workflow. Then we run a longer track: kickoff and a technical deep dive with your team and your software vendor, building in signal mode, an acceptance test with your staff, and going live one agent at a time. FAIR’s first release ran that way in six weeks.
Built by people who understand the whole system.
Our background is software architecture, application development and the recovery of projects that had gone off the rails. That is why we care about the unglamorous parts, the audit log, the write that can never be applied twice, the one source of truth. And why an agent from us acts inside your real systems rather than in a demo.
The way we work is simple. Engineers sit down with you and set the requirements and the architecture. AI agents write the code. Loom runs the agents. One team, led from Utrecht, with engineers in Tirana and Skopje.
Let’s talkA few things you may be wondering.
Where should we start with AI?
Start with a recurring task that costs time and has a clear result. Together we look at the available data, the systems involved and the decisions that need a person. That gives a practical first use case, and usually a good agent day.
Can an agent work with our existing software?
Yes, as long as your software offers an API or another way to connect. We check access, permissions and data quality before we choose an approach.
Do we have to use Loom?
No. Loom is how we build fast and keep people in control. If your situation calls for custom software, our AI software factory builds it.
Where does Loom run?
As a service in our private cloud. You get your own login.
Can we change things ourselves?
Yes. Workflows, connectors, approvals and knowledge are all in the interface. Or leave it to us and we manage everything for you.
How quickly is something running?
One day on Loom for an AI agent. Three days in our AI software factory for a working first version of your product. A platform of your own takes weeks.
What can be built in one day?
From a simple check to a process with several steps, API calls, AI decisions and human approval. Loom’s building blocks are what make a complex process possible in a day.
What do we get after three days?
A working first version of your product, an MVP. The core flow from start to finish, with screens, integrations and data connected, running on a URL. Not a certified, multi-tenant product yet. That is what the weeks after are for.
Who writes the code?
AI coding agents in our software factory. Our engineers gather the requirements, set the architecture and review every piece. You get code you can read and own.
How do we keep people in control?
We agree which actions an agent may take, where approval is needed and when it has to escalate. Those boundaries become steps in the workflow, together with monitoring and a record of every action.
What happens after go-live?
Monitoring, support and a next iteration. Yourself in Loom, or fully managed by our team.
Bring us one process. Or one idea.
Tell us which task your team keeps repeating, or which product you want to put in someone’s hands, and who decides. Within one working day you get a date. An agent day, or a three-day build.
Or email us: info@rutron.nlPrefer to call? +31 (0)30 899 3767
One day for an agent. Three days for a first version. Bring it along and see for yourself.