News
Insights from the field
What we learn from hundreds of thousands of cases per year. On AI in the back office, process automation and humans in the loop.
Automating order processing in practice
How do you automate order processing without your team losing control? An AI coworker handles orders end-to-end in your ERP. With real numbers from wholesale and construction.
Automating order processing in Exact Globe
Customer orders arrive by email and PDF, then have to be keyed into Exact Globe by hand. An AI coworker reads, checks and books the sales orders directly, with your team on the exceptions.
Automating order entry in AFAS
Retyping customer orders into AFAS by hand costs time and creates errors. An AI coworker reads orders from email and PDF and books them as a sales order in AFAS, under control.
Automating invoice processing
Purchase invoices cost more time than you think. See how an AI coworker reads invoices, matches them against the purchase order and the receipt, and posts them under control in your ERP.
Automating three-way matching
Automating three-way matching means lining up the invoice, purchase order and goods receipt automatically. You catch discrepancies early and keep a grip on your margins, even at high volume.
Automating service tickets in construction and installation
Re-typing service tickets by hand costs the back office hours every week. Here is how an AI coworker reads, checks and books field tickets into your project administration, with a person in the loop where it matters.
Automating order processing in wholesale
Wholesalers process thousands of orders a month across email, PDF and EDI. An AI coworker takes over that order flow and books straight into your ERP, with a human in the loop for exceptions.
What is an AI coworker?
An AI coworker runs back-office processes on its own and brings in your team when an exception comes up. Here is what an AI coworker is, and why it is something other than a chatbot.
Procure-to-pay end-to-end: from requisition to payment by one AI coworker
Procure-to-pay is split across purchasing, finance and approvers. An AI coworker turns it into one stream: from purchase requisition to SEPA batch, with the same controls your team uses.
RPA vs AI agents
RPA follows fixed rules and excels at high volume. An AI coworker interprets unstructured input and handles exceptions. When do you use which, and how do they work together?
AI agent vs chatbot
A chatbot talks, an AI coworker acts. What exactly separates an AI agent from a chatbot, and what does it mean for your back office? A clear, no-nonsense overview.
Quote and contract administration: an AI coworker as sales-ops' right hand
Between a customer's yes and a signed contract sits a lot of repeat work: lookups, recalculations, version control. An AI coworker handles that next to your sales team, while your commercial people keep selling.
Automating business processes with AI: where do you start?
Which process should you tackle first with AI? A clear selection framework based on volume, regularity and unstructured input points you to the process where the gains are largest.
Calculating the ROI of AI automation
How do you weigh the investment in an AI coworker against the return? We walk through the costs and benefits that matter and give a concrete worked example you can run for yourself.
HR administration: onboarding, sickness and contract changes by an AI coworker
HR work is people work, but around it sits a stack of administrative tasks that crowds out the important pieces. An AI coworker takes over onboarding mutations, sickness registration and contract changes.

Hackathon: consultants build a three-way matching agent in an afternoon
On April 30, teams from BPM Company and Mendify built a working three-way matching agent for accounts payable in a single afternoon, hands-on in the AgentsLab platform with Claude Code as a sparring partner.
Customer service tickets: from Zendesk or Freshdesk message to resolved in the ERP
A ticket in Zendesk or Freshdesk isn't a resolution. The resolution lives in the ERP: a credit note, a return, an address change. An AI coworker bridges the last mile.
AI coworker in Bouwworks: project administration and service tickets at production scale
Bouwworks is the system that ties projects, hours and subcontractors together in construction, installation and infra. An AI coworker works directly in Bouwworks, links service tickets to projects, and keeps the administration current.
Accounts payable with an AI coworker: from invoice to payment batch
Accounts payable is more than posting invoices. It's matching, approval, planning payment batches and managing supplier relationships. An AI coworker runs that whole chain, with human review where it matters.
Master data management: keeping customer and item data clean with an AI coworker
Duplicates, stale prices, missing VAT codes: master data rots by itself. An AI coworker keeps customer and item data clean against your own rules, without an MDM tool on the side.
AI coworker in SAP S/4HANA: native integration via BAPI and OData
SAP sits at the heart of many Dutch enterprises. An AI coworker posts directly into S/4HANA through BAPIs, OData and RFC, with no middleware and the same audit trail as a person.
AI coworker in Dynamics 365 Business Central: native to Microsoft's cloud ERP
Business Central is Microsoft's cloud ERP for the mid-market. Our AI coworkers work inside it via the standard Web API, with the same permissions and audit trail as a user.
AI coworkers in Odoo: process automation via XML-RPC and JSON-RPC
Odoo is the open-source ERP gaining ground fast in the Dutch mid-market. AI coworkers work inside it natively via XML-RPC and JSON-RPC, with no extra modules installed on your Odoo server.
Month-end and quarter-end close: from nine working days to three with an AI coworker
Close is the moment finance teams put in extra hours to supply the rest of the business with numbers. An AI coworker handles the preparatory checks, so the controller can focus on the explanation.
Agentic AI versus RPA: why RPA projects stall and what takes their place
RPA works as long as the process is perfectly predictable. The moment an exception shows up, it stalls. Agentic AI is about the opposite: context, reasoning and escalation.
Automating invoice processing in 2026: from OCR to AI coworker
OCR reads, but doesn't understand. An AI coworker matches, posts and escalates, end-to-end in your financial admin.
AI coworkers in professional services: accountancy, consulting and legal
Professional services live on billable hours. File work, intake and reporting eat into them. An AI coworker keeps the non-billable work small, so partners and consultants have more room for real work.
Accounts receivable with an AI coworker: from reminder to relationship
Collections is repetitive work, but also delicate. An AI coworker sends reminders, segments by payment behavior and escalates to the account manager, without damaging the customer relationship.
Order processing: from 10 minutes to 27 seconds per order
A wholesale customer was processing sales orders in ten minutes apiece. Six weeks later the AI coworker was running in production, average handling time: 27 seconds.
AI coworker in logistics and transport: from order intake to ePoD
Logistics has the operation under control, but around it sits a layer of administrative work: order intake, CMR, customs, billing. An AI coworker handles that layer, at the speed of the warehouse.
AI recruitment agent: from 240 CVs to 3 candidates for the first interview
Recruiters spend up to seventy percent of their time on screening. An AI coworker takes that work on, with transparent criteria and the recruiter making the final call.
AI inside Exact: how an AI coworker natively works in Online and Globe
No CSV exports, no middleware. An AI coworker works inside Exact Online and Globe the way a human does, through the REST API and XML/SDK, with the same audit trail.
AI coworker in manufacturing: from order to work order in four minutes
The step from customer order to work order needs coordination with purchasing, planning and engineering. An AI coworker handles that step, with a planner who stays in control of capacity.
GDPR for AI coworkers: where your data sits and why it matters
AI adoption often stalls on one question: where does our data go? For AI coworkers working in Dutch and European processes, the answer is clear: EU hosting, role-based access and case-level audit trails.
Why AI adoption succeeds or fails on people, not technology
The biggest reason AI projects stall isn't the technology, it's adoption. That's why 'human in the loop' isn't a concession but our most important design principle.
Multichannel retail: one order stream from Shopify, bol.com, Marktplaats Pro and the store
Retailers today sell through four to eight channels at once. For operations behind the scenes that has to be one stream. An AI coworker makes orders, stock and returns a single reality.
Shared inbox triage: how an AI coworker gets info@, orders@ and service@ under control
Shared mailboxes are the nerve center of operations and the biggest time sink at the same time. An AI coworker classifies, routes and executes, without forcing you to reorganize the team.
AI coworker vs. chatbot and copilot: which tool for which work?
Chatbots answer. Copilots assist. An AI coworker does the work. What that difference means for your automation choices.
AI coworkers in the public sector: finding responsible scale at municipalities and agencies
The public sector can't afford automation that isn't explainable. An AI coworker that runs in NL, doesn't train on your data and has every decision approved fits inside those frames.
Process automation for Dutch wholesale: where AI coworkers make the difference
In wholesale, volume isn't news. The real challenge is variation: hundreds of customers, thousands of items, dozens of formats. That's where an AI coworker shines.
Automating service tickets in construction and installation: 10 minutes to 10 seconds
Service tickets from subcontractors arrive by email, PDF, portal, sometimes handwritten. An AI coworker pulls the data, links the right project and posts in Bouwworks or AFAS, in seconds.
Build vs. buy: build your own AI agent or pick a platform
Building your own AI agent sounds attractive: full control, no vendor lock-in. But the R&D piece you're signing up for is larger than most teams estimate. An honest comparison.
What does an AI coworker cost? Pricing model, TCO and payback
An honest business case looks beyond a software subscription. What does an AI coworker cost, implementation and maintenance included, and when does it pay back?
The business case for an AI coworker: how to calculate the ROI
AI automation only matters when the math works. Four variables drive the ROI, and one question decides whether you should even start.
AI governance and policy: what at minimum belongs in an AI policy?
An AI policy isn't a disclaimer, it's a working document that sets scope, roles and escalation. What needs to be in there at minimum, and the questions you can't forget.
Exact Online and AI coworkers: automation in the financial admin
Exact Online is built for daily use, but processing invoices, orders and projects often stays manual. An AI coworker takes that on, directly in Exact, no middleware.
Microsoft Copilot vs. an AI coworker: where the line sits
Copilot helps a person be faster. An AI coworker does the work itself. Two different things, two different contracts, two different ROIs. The practical boundary explained.
From pilot to production: why so many AI projects get stuck
Most AI pilots don't reach production, not because the technology fails, but because the bridge from 'proof of concept' to 'daily use' never gets built.
Agent evaluation and observability: how to measure an AI coworker in production
An agent in production doesn't have unit tests, it has evaluation sets. How to make an AI coworker measurable without going through every run by hand.
Document classification: from one mailbox with everything, to one process per document
Most back-office mailboxes receive everything mixed together: orders, invoices, complaints, service tickets. An AI coworker classifies and routes, document by document, in real time.

Partnership Announcement: Lighthouse E2I & AgentsLab
Lighthouse Empower to Improve and AgentsLab are joining forces to structurally strengthen service organizations in construction, installation and infra with the AI Service Agent.

AgentsLab at Improven's Friends 2 the Firm
On November 20, 2024, AgentsLab is presenting at Improven's Friends 2 the Firm how AI coworkers are already transforming real business processes today.

How Future-Proof is Your Business?
On November 28, AgentsLab, IT Synergy and FACET are hosting the Future Proof event in the Van Nelle Fabriek: on AI, NIS2 and data as a strategic advantage.
Empowering Your Team with AI: The AgentsLab Advantage
AgentsLab's AI coworkers adapt to your team's needs, from customer communications to invoice processing and administrative support.
Harnessing AI Coworkers for Enhanced Productivity with AgentsLab
AgentsLab offers specialized digital team members that integrate seamlessly into your organization and boost productivity by automating routine tasks across departments.
Exploring OpenAI's New o1 Model Series
The OpenAI o1 model series, with o1-preview and o1-mini, targets accuracy and reliability in complex tasks. What does that mean for AI coworkers in practice?
Curious what an AI coworker can do for your process?
Book a no-strings Quick Scan and explore the options.
Book a Quick Scan