Blog Summary
Digitising changes the format of work. Automating removes the human from it. Vendors often sell the first while charging for the second. Five plain tests reveal which one a proposal delivers. The real dividing line is judgment and handoffs, not paper versus pixels. A short scoring audit at the end grades any project before you sign.
Automation might be the most oversold word in enterprise software. Vendors attach it to dashboards, portals, and digital forms, and buyers sign the contract expecting the work to disappear. Then a person still keys the data, still makes the call, and still chases the stragglers. That gap between what was promised and what shipped has a name worth knowing before you spend: it is the quiet difference between digitising and automating.
Most guides stop at the definitions. This one does something more useful. We hand you a set of tests to run before you sign anything, because each one reveals whether a project actually removes work or simply moves it to a nicer screen.
Format changed, or judgment changed?
Digitising changes the medium. Paper becomes a PDF, a clipboard becomes a web form, and a filing cabinet becomes a database. Automating changes who does the work itself, so a step that used to need a person quietly becomes a step a machine handles. Both are legitimate, and neither is the villain here. The dishonesty lives in the labels, not in the methods.
A scanned invoice is digitised, but someone still has to read it and approve it. An invoice that validates, matches, and posts itself is automated, because no one touches it. McKinsey’s recent analysis estimates that about 57% of work hours are technically automatable today, yet it stresses a point most buyers skip past: the value comes from redesigning whole workflows, not from automating isolated tasks. So the real question is never whether a process is digital. It is whether the human has left the loop.
Five tests that separate digitising from automating
Run these five tests against any proposal, demo, or vendor claim. Each one resolves to a simple yes or no, and the pattern across all five tells you the truth that the sales deck will not.
1. The decision test
Ask who makes the decision once the project goes live. If a person still decides, you have digitised the inputs rather than automated the work. The data arrives faster and cleaner, which is genuinely useful, but the judgment never moved, and that is digitisation by any honest measure. Real automation encodes the rule and then acts on it, so remember that approvals, routing, and pricing are decisions, not just screens.
2. The unattended test
Picture the office completely empty at two in the morning. If the work still finishes, it is automated; if it sits in a queue waiting for the morning shift, it is digitised. Digitisation depends on a human turning up to push it along, while automation keeps running while everyone sleeps. This test quietly exposes the kind of automation that turns out to be a tidier inbox.
3. The exception test
Normal cases are easy to dress up as automated, so the truth shows up on the awkward ones instead: a missing field, a duplicate record, or an amount that looks wrong. Weak systems push those exceptions back onto a person without much ceremony, while strong systems route them, flag them, and keep the rest of the flow moving. This is where systems that reason over exceptions and act without a person waiting earn their cost. Ask the vendor to demo the ugly case, not the clean one they rehearsed.
4. The handoff test
Map every handoff in the process as it runs today, then map the handoffs that remain after the project ships, and count both honestly. If handoffs disappeared, you automated; if they merely moved to a prettier screen, you digitised. A surprising number of transformations do nothing more than relocate the same clicks into a new interface.
5. The throughput test
Digitisation often leaves both headcount and throughput exactly where they were, with the same team handling the same volume using better tools. Automation, by contrast, shows up in the numbers: the same team handles more, or fewer people handle the same load. If capacity did not move at all, treat the automation claim with suspicion, because throughput is far harder to fake than a polished demo.
By The Numbers
Where adoption stops and real value begins.
of organisations now use AI in at least one function. Yet roughly two-thirds have not begun scaling it.
of companies capture substantial value from AI. Those leaders are the ones not merely automating.
of US work hours are technically automatable today. Mostly by software agents, not physical robots.
Why the difference decides your budget
Digitising and automating do not cost the same, and pretending otherwise is where budgets quietly bleed. Digitising is cheaper, faster, and lower risk, while automating costs more upfront and then compounds in your favour over time. The trouble starts when scope and price disagree, and a team ends up paying automation rates for digitisation outcomes. Six months later the workload has not shrunk, and no one can quite explain why.
BCG found the gap between AI winners and the rest widening fast, with the leaders reshaping how the business works rather than only automating it. Before you approve the spend, separate the two scopes cleanly and know what a scoped automation engagement should actually include.
The honest sequence a good consultant follows
You usually cannot automate a process you have never digitised, because machines need clean, structured, digital inputs before they can act on anything. So digitisation is often step one rather than the enemy, and the honesty lies in naming each phase plainly. A good partner will tell you which one you actually need right now, and sometimes the answer is to digitise this year and automate the next. That advice costs them revenue today, which is exactly why it builds trust. It is worth looking for a partner willing to tell you the unprofitable truth, because the wrong one will keep selling automation language for digitisation work.
Watch Out: The Digital-Paper Trap
The most common failure mode is false automation. A portal looks automated while a human still keys data behind it.
Three signs you bought digital paper, not automation:
- A queue that only ever clears during working hours.
- An exception path that quietly dead-ends in someone’s inbox.
- A headcount plan that did not change at all after launch.
Where modern AI moves the line
For years, judgment marked the hard edge of automation: decisions stayed with people, while software handled the rote steps around them. Capable AI agents now shift some of that judgment, reading context, weighing rules, and acting within limits you set. McKinsey frames the change as moving people from execution toward orchestration, and it is already real in service workflows where the system resolves a case end to end.
The catch is that the line only moves when the workflow is redesigned around the new capability. Bolting an agent onto an old process changes very little, because the work itself has to be rebuilt to take advantage of what the agent can do.
Run the audit before you sign
We built a short audit for exactly this moment. It scores any proposed project across the five tests, with each one worth zero, one, or two points. Higher totals point to genuine automation, lower totals to digitisation wearing an automation label, and you can use the result to challenge scope, price, and promises in the same conversation. Copy the scorecard below and reuse it on every proposal that crosses your desk.
The Digitise-or-Automate Audit
| Test | 0 points (digitised) | 1 point (partial) | 2 points (automated) |
|---|---|---|---|
| Decision | A person still decides | Machine suggests, human approves | Machine decides and acts |
| Unattended | Waits for working hours | Partly runs after hours | Completes with nobody present |
| Exception | Lands silently in an inbox | Flagged but stops | Routed, flagged, and continues |
| Handoffs | Same handoffs, new screen | A few removed | Handoffs largely gone |
| Throughput | Capacity unchanged | Modest lift | Clear capacity gain |
The honest takeaway
None of this is an argument for automation over digitisation. Both earn a place in any sensible roadmap: digitisation cleans the foundation, and automation removes the work that sits on top of it. The harm is only in calling one the other, then paying accordingly. Buy the outcome you actually need, pay the price it genuinely deserves, and afterwards measure whether the work really went away.