Service — Process automationMeasure first, then automate
Automate what eats your hours
Integrations between the tools you already use —ERP, CRM, email, spreadsheets— and AI agents for the work that used to need reading and deciding. First we measure how many hours go; then we automate, and measure again.
Terms
- Timing
- Process by process: automate one, measure it, move to the next.
- Price
- Per project, fixed in writing before we start.
- Tools
- The ones you have. Switching ERP or CRM is almost never step one.
Every company has a process someone does by hand because it's always been done that way: copying orders from email into the ERP, moving data from a PDF to a spreadsheet, building the same report every Monday, chasing whoever has to approve something. Each one eats a few hours. Added up, a whole job.
Automation doesn't start with the tool. It starts with measuring: how often it happens, how long it takes and what it costs when it goes wrong. That decides what gets automated first —what pays back most, not what looks best— and tells you afterwards whether it worked.
Who it's for
- Finance and admin teams keying in invoices, delivery notes or orders by hand.
- Sales teams building every quote from scratch.
- Operations and logistics coordinating by email and phone what could trigger itself.
- Companies with a CRM nobody updates because it means retyping what's already somewhere else.
What we automate
Tools that talk to each other. ERP, CRM, shop, email and spreadsheets connected through their APIs: Odoo, HubSpot, Salesforce, Shopify, Google Workspace, Microsoft 365 or whatever you use.
Reading what arrives as PDF. Invoices, delivery notes, orders and forms: the data is extracted, checked and entered where it belongs. Anything that doesn't add up is set aside for review.
Approvals and alerts without chasing anyone. Who approves what, when they're notified and what happens if they don't respond. With a log of everything.
AI for what needs judgement. Sorting emails, drafting replies, writing a quote from a request. Always with a person before anything with consequences.
Set up, and feeding itself. A useful CRM is one that fills itself: connected to email, the website and invoicing so the data comes in on its own.
Hours before and after. Every automation is measured against the starting point. If it doesn't save what was expected, we say so and fix it.
What you get
- The process measured before we touch it: occurrences per month, time and errors.
- The automation in production, logging what it does and alerting when something fails.
- The same measurement afterwards, so you know what it paid back.
- Documentation so your team can change it without depending on us.
What people ask us
- Which processes should we automate first?
- The ones that repeat often, follow clear rules and are expensive when they go wrong: entering orders or invoices, reconciling payments, recurring reports. Processes that need judgement can be helped by AI too, but they come later, once the basics work.
- Do we have to switch ERP or CRM?
- Almost never as a first step. Most current tools have an API, and connecting what you already have is cheaper and faster than migrating. If the diagnosis finds a tool is the problem, we say so with numbers.
- Do you use Zapier, Make or n8n?
- When they're enough, yes: they're quick to set up and easy to maintain for simple flows. When volume, cost or logic outgrow them, we write code. The choice is explained, and the system isn't locked into a platform.
- What happens when the process changes?
- The automation changes. That's why it's delivered documented and, if you want it, with monthly maintenance: an automation nobody knows how to touch gets switched off at the first change of vendor.
How it's bought
Every service follows the same path: a short diagnosis, the build with weekly deliveries and, if you want it, ongoing maintenance.
Related services
Let's talk
Tell us what you're building.