Most small and mid-sized firms don’t lose ground because a rival has more people or a bigger budget. They lose to friction: slow handoffs, duplicate entry, stale data, and decisions made after the fact. Digitalization is the discipline of removing that friction. Done well, it shortens cycle times, raises decision quality, and makes your company easier to work with for customers, suppliers, and the people you want to hire.
Start with reality, then integrate
The place to start isn’t a shopping list of features; it’s a clear picture of how work actually flows today. Bring sales, operations, logistics, and finance into the same room and map reality, not the ideal. In my experience, failed “IT projects” are usually unresolved process disagreements wearing a software label. Once you have an agreed current state, prioritize integration before automation. If systems can’t talk, automation will only speed up the chaos. Connect orders, inventory, fulfilment, and invoicing so data flows correctly.
Design a layered architecture
Keep the architecture simple. Your ERP is the system of record: the dependable source for customers, products, inventory, orders, and financials. Around it is the system of work: the screens and mobile apps where people actually get things done, with role-based views that match how jobs are performed. On top sits the system of insight: live dashboards, alerts, and assistants that surface exceptions and recommend next steps.
A quick test: if a planner changes a delivery date in the warehouse app, can sales make a reliable promise and can finance invoice without manual fixes? If not, the problem isn’t the tool-it’s the architecture.
Treat the cloud as risk management
For SMEs, the cloud is risk management in plain clothes. It transfers the operational burden of uptime, backups, patching, and disaster recovery to people who live and breathe those tasks. It also removes the “indispensable admin” risk, when one person knows the on-prem server by heart. The hidden cost of staying local isn’t hardware; it’s stagnation, followed by an expensive catch-up later. Cloud turns unpredictable upgrades into a service rhythm you can plan for.
Keep AI practical and governed
AI deserves a practical brief, not a mystical one. Start where the ground is firm: searchable know-how so new hires can find procedures and product rules without interrupting veterans; drafting support so quotes, pick notes, and routine emails begin life mostly done; exception spotting so managers see orders likely to miss promise dates or invoices likely to be disputed; and natural-language queries that let a planner ask, “What’s blocking orders over €5,000 this week?” with a traceable answer. Keep guardrails in place, clear data lineage, get human approval for irreversible actions, and complete logs, and treat AI as a copilot, not an autopilot.
Your digital workplace is a talent strategy
Your systems say as much about your culture as your careers page. The people you want expect clean interfaces, mobile workflows, and self-service answers. They don’t expect to fill out PDFs or wait three days for a report. Modern tools aren’t a perk; they’re table stakes for attracting and keeping good people.
A 90-day kickstart
If you want a simple rhythm, plan three 30-day blocks.
- In the first month, map two critical flows: order-to-cash and procure-to-pay. Then, measure where time is lost and rework happens, and write a one-page definition of what “good” looks like.
- In the second month, connect the basics (orders, inventory, fulfilment, invoicing), clean your master data, and put a single operational dashboard in front of everyone.
- In the third month, automate the few repeatable steps that burn the most time, pilot a small AI assistant where the documentation is solid and the risk is low, and train by role judging adoption by outcomes, not attendance.
From there, work to a 6-6-18 horizon: by six weeks you should feel smoother flow in one connected process; by six months, two more processes should be integrated and your close should be faster; by eighteen months, structured documents should be flowing with key partners, and predictive signals should be part of everyday decisions.
What “good” looks like—and how to keep score
You’re on track when orders move from quote to invoice without retyping, a service rep can promise dates from one screen and be right, inventory accuracy supports reliable picking without heroic cycle counts, managers start the day with live exceptions they can act on, upgrades don’t require weekend war rooms, and new hires become productive in weeks instead of quarters.
At the board level, keep asking simple questions: which cycle times shortened this quarter and by how much; which decisions moved from weekly to daily or real time; where do we still retype data and why; how resilient are we to a site outage or a lost laptop; and which manual approvals disappeared because the policy now lives in the system. If that last list is empty, be suspicious, you may be automating around a policy gap.
The point isn’t to own the most tools; it’s to have the fewest handoffs and the fastest trustworthy insight. Start there – the rest will fall into place.