New Zealand businesses are already highly digital.
Recent MBIE research found that almost all NZ businesses use digital or online tools, and around half are already using AI. Yet 51% of businesses say they could get more value from digital tools.
That is an interesting gap.
Businesses are buying software, moving information online and introducing new technology, but that doesn't necessarily mean the work behind the scenes has become more efficient.
As a Business Analyst, this is something I look at differently.
I'm less interested in how many systems a business uses. I want to know what people still have to do manually to get a process from start to finish.
That usually tells us much more about where the real opportunities are.
A digital process can still be very manual
Take invoice processing as a simple example.
A business may use a modern accounting or finance system. All invoices eventually end up there, payments are recorded digitally and financial reports come out of the system.
On paper, that sounds fairly automated.
Now look at what actually happens.
An invoice arrives by email. Someone opens it, checks the supplier, enters the invoice details into the finance system, selects the coding, adds a cost centre or project, attaches the document and sends it for approval.
The accounting system is doing its job, but a large part of the process is still being completed manually.
That distinction matters.
Simply moving work from paper into software is digitisation. Automation starts when technology removes unnecessary manual steps from the process.
Excel is another good example
There is nothing inherently wrong with Excel. In fact, it can be exactly the right tool for many businesses.
The question is how the information gets there and what happens to it afterwards.
Imagine a business keeps all customer enquiries in an Excel spreadsheet.
If a customer submits an online form and their details automatically appear in the spreadsheet, that part of the process requires very little manual work.
If the customer calls or sends an email and an employee then copies their name, phone number, service required and other details into exactly the same spreadsheet, the outcome looks similar — but the process is completely different.
One captures information automatically.
The other requires somebody to do the work every single time.
Multiply that by 20, 100 or 1,000 enquiries and a small manual step becomes a significant operational cost.
The same applies to CRM and other business systems
A CRM can be extremely useful without making the customer journey automated.
For example, staff might still need to:
- copy customer information from one system to another;
- manually update statuses;
- create follow-up tasks;
- prepare standard emails;
- check another system for availability or pricing;
- create quotes;
- update spreadsheets for reporting.
The first question should be:
Why is all of this manual work still required?
The answer might be:
- an existing feature is not being used properly;
- two systems are not connected;
- the workflow itself needs to be redesigned;
- or the current technology genuinely cannot support what the business needs.
You cannot tell which one it is just by looking at a list of software.
Manual Customer Enquiry Process
Automated Customer Enquiry Process
Where should automation actually start?
Automation starts with understanding the current process.
This is standard Business Analysis practice.
I start from the beginning of the process and identify the trigger — what actually starts it. From there, I map the information sources, the steps people take, the systems involved, decisions, hand-offs, manual data entry and exceptions.
The goal is to understand what really happens from start to finish, not just what the process is supposed to look like.
Once the current process is clear, I look at where manual work creates the biggest impact.
Some manual steps are inconvenient but relatively low risk.
Others can directly affect the customer or the business. A typo in a customer's name may create a poor experience. An incorrect phone number could mean the business loses the lead completely. In finance or operational processes, a manual error can have much more serious consequences.
So the next step is to look at the manual parts of the process by risk, effort and business impact.
From there, the right answer may be to automate the whole process.
But it does not have to be.
Automating just one or two high-impact stages can make a significant difference.
For example, a process might move from an Automation Level of 2/10 to 5/10, or even 7/10, without replacing every system or removing every human decision.
That can already mean fewer errors, less repetitive work, faster processing and a much better customer experience.
The goal is not to reach 10/10.
The goal is to reach the right level of automation for the process.
So where does AI fit?
AI is useful when the problem actually requires capabilities that AI is good at.
For example, a process may involve interpreting unstructured documents, understanding free-text enquiries, extracting information that isn't presented consistently, classifying complex information or assisting with decisions where traditional fixed rules aren't enough.
That's very different from using AI to move a customer name from one system to another.
If a simple rule, integration or workflow can do the job reliably and at a lower ongoing cost, I would use that instead.
AI should have a business reason for being in the process.
The same rule applies to any technology decision: use the simplest solution that solves the problem properly and can support the business as it grows.
Before buying another system, look at the work
The MBIE finding that 51% of NZ businesses believe they could get more value from digital tools doesn't necessarily mean those businesses need more tools.
Some will. Others may already have most of what they need.
The opportunity could be sitting between their existing systems, inside an unnecessarily manual workflow, or in a process that grew organically as the business got busier and was never redesigned.
That is why I think measuring digital adoption alone only tells part of the story.
A business can have cloud software, online forms, Excel, a CRM, an accounting platform and AI — and still have employees spending hours copying information, checking routine transactions and chasing the next step.
For me, the more useful question is:
How much manual work is still required for this process to run?
Once we understand that, we can decide what actually needs to change.
It may be a quick process improvement.
It may be better use of an existing system.
It may be an integration or a straightforward automation.
It may require different technology.
Or there may be a genuine opportunity for AI.
The solution comes after understanding the process — not before it.
SOURCE