Less manual work.
Sometimes the website is not the real problem. Information gets retyped, leads get lost between systems, reports are built by hand every month. Then we connect your systems, or build something that fits between them.
scroll ↓The same every week. Nobody enjoys it, everybody does it.
Your systems talk to each other.
What should never run on its own, doesn't.
Sometimes a rule, webhook or API is more reliable and cheaper than AI. Where language or analysis adds real value, AI becomes part of the system. People review the decisions.
A lead chain that becomes a deal within seconds.
For A'DAM VR we rebuilt the lead and CRM chain: from Salesforce to HubSpot, forms become deals within seconds, and a nightly job drafts invoices and checks subscriptions. Leads answered within 24 hours: 43% → 92%.
This used to be done by hand.
Eight connections running today. For each: what happened before and what runs by itself now. At the bottom, what it delivered.
Before. Group enquiries landed in Salesforce. The alert for a new form had been silent for over a year.
Now. The form creates a deal in HubSpot within seconds. No reply in five days? The salesperson gets a task.
median response time 43.5 h → 5 hBefore. Groups book days to weeks after the click. Google Ads never saw that booking.
Now. The form keeps the click ID. Every day, won deals go back to Google Ads as offline conversions.
first scheduled run: 100% accepted, 0 errorsBefore. A HubSpot invoice has no BCC field. The bookkeeper received nothing.
Now. At 04:20 a nightly run collects every new invoice. The PDF and journal lines in Exact's format go to accounts.
customer numbers assigned automatically · running twice books nothing twiceBefore. A reply to a cold email lands in the mail tool. The CRM knows nothing.
Now. A webhook puts the reply on the contact in HubSpot. The salesperson gets a call task for the same day.
reply → contact and task in 35 secondsBefore. Quote and group enquiries never reached Klaviyo. Email did not know who had already been in touch.
Now. Every enquiry and every click to call or email goes to Klaviyo as an event. A subscription only with the tick box.
eight flows, per language · Dutch opens 33%, 2× EnglishBefore. The pipeline had stood still since June. 43% of open deals were tests.
Now. Forms create a deal with click ID themselves. Google Ads fetches booked deals as a CSV. Order ID is deal ID, so nothing counts twice.
no extra tooling · test domains are now blockedBefore. Car washes and wash bays are scattered across maps and chain websites. Nobody had them on one map.
Now. Scripts merge 15 sources into one data model. The site builds every location page from it.
1,605 locations · prices 9 → 62, opening hours 55 → 209Before. Publishing every article by hand and submitting it to search engines.
Now. An article gets a date and goes live on that day by itself. IndexNow submits it straight away.
54 articles scheduled · position 18.5 → 7.3 in four weeksYou see for yourself how your site is doing.
Custom work and integrations come with a portal too. It shows what we measured, what we did this month, which agreements apply and how many improvement hours are left. No email needed to know everything still works.
03 Automation sprint. one processAPI before AIerror handlingloggingdocumentation from €2,500 excl. VAT +
A form comes in. Someone copies it into the CRM. Then an email goes out. At the end of the month the same information goes into a spreadsheet again. Processes like that grow on their own.
In an automation sprint we automate one business process. At the end there is a working integration, workflow or tool.
Who it's for. Teams that do the same manual work every week, retype information between systems, or run processes held together by loose spreadsheets, inboxes and tools.
First we map the process. What information comes in where? Which systems are involved? Which steps take time? And above all: which steps should not be automated at all? Then we build and test: an API integration, a workflow, a database, an internal tool or a mix. We only use AI where interpretation or language adds something. A plain software integration is often the better choice.
No slide deck about what might be automated. At the end there is something that works. Where needed with error handling, logging, human approval and documentation. It records what the system does when everything goes right, and when something fails.
A sprint solves one clearly bounded process. A new business platform, a large CRM migration or a system with dozens of integrations is scoped as a custom project first. Licence costs for third-party software are outside our development price.
Some automations are one clear integration. Others touch five systems, several user roles, sensitive data and processes that must not go down. The amount of logic, integrations, exceptions and checks sets the scope.
05 Custom work and systems. client portalsinternal toolsdashboardsdata pipelinesAPI integrationsworkflow automation priced on scope +
Sometimes a website is a small part of the question. Information from several systems has to come together, a team works with manual processes that have grown too big, or the software you need doesn't exist yet.
Then we don't force the problem into an existing package. We first map what needs to happen.
Who it's for. Organisations where the challenge sits in the systems, data and processes behind the website, not only in the front end.
Client portals, internal tools, dashboards, data pipelines, CRM processes, API integrations, reporting systems, booking logic, automation, or software that runs an existing business process better.
Not with a tech stack. First: what does someone need to be able to do? What information does that take? Where does it come from? Which systems already exist? What must stay reliable at all costs? And what has become needlessly complicated? Only then do we decide what to build.
Because 'custom software' says little about size. An internal tool for three people can be smaller than one complex integration with an external platform. We make the scope clear first, then estimate development, integrations, data, testing and maintenance.
What do you
do by hand?
Show us what runs by hand now. Sometimes one good integration is enough.