Automation only works when the process underneath it works first
Eliminate the work that's slowing your team down.
We fix the process first, then automate what works — so you get lasting results, not faster failure.
This is the principle that most automation projects ignore — and why so many of them fail to deliver what was promised.
Connecting a broken process to an automation tool doesn’t fix the process. It just runs the same mistakes faster, at greater scale, with fewer people around to catch them.
At Changeable, we do this differently. Before any automation is designed or built, we do the process work. We analyse how work actually flows in your organisation, identify where time and effort are being lost, and redesign the workflow so it’s lean and logical. Then — and only then — we apply AI-powered automation to amplify what’s been fixed.
The result is automation that sticks, scales, and keeps delivering value as your organisation grows. We fix the process before we touch the technology.
Why most automation projects underdeliver
New Zealand organisations have been experimenting with automation for years — workflow tools, integrations, RPA, and now AI. Yet in most cases, the results fall well short of what was expected. The technology usually isn’t the problem.
The patterns we consistently see:
- Automation built around how the process was documented, not how it actually runs — the two are rarely the same
- High-volume, low-complexity tasks automated first because they’re easy — not because they deliver the most value
- Shadow processes and workarounds baked into the automation, making them permanent rather than fixing them
- No baseline metrics established before implementation, so it’s impossible to measure what actually improved
- Staff who weren’t consulted during design, so adoption is low and workarounds reappear quickly
- Point solutions that automate one task in isolation, creating new handoff problems between the automated and non-automated parts
Good automation starts with a clear-eyed view of current state — including the work that isn’t in the process documentation. That’s the foundation we build from.
What we can automate for your organisation
We work across a wide range of workflows. The most common areas where AI automation delivers measurable value for NZ organisations:
Administrative and operational workflows
Data entry, document processing, form handling, record updates, scheduling, and inbox management. These are typically high-volume, rules-based tasks that consume significant staff time without requiring human judgment. Well-designed automation here frees capacity immediately.
Reporting and compliance
Automated data extraction, report generation, exception flagging, and audit trail creation. For councils, public sector teams, and regulated businesses, this reduces the manual burden of compliance reporting while improving consistency and traceability.
Approvals and case handling
Routing requests through approval chains, triaging incoming cases, escalating exceptions, and maintaining a clear audit trail throughout. AI can handle the rules-based routing while keeping humans in the loop for decisions that require judgment.
Customer and stakeholder communications
Automated responses to common enquiries, status updates, acknowledgements, and follow-up sequences. When designed carefully — with appropriate human review checkpoints — these reduce response times and improve the experience for the people your organisation serves.
Data integration and synchronisation
Connecting systems that don’t talk to each other, eliminating manual data transfers between platforms, and keeping records consistent across tools. Much of the manual effort in NZ organisations sits in the gaps between systems — automation closes those gaps.
Internal knowledge and process workflows
AI-assisted document drafting, knowledge retrieval, meeting summaries, and internal workflow routing. Particularly valuable for teams with complex, high-volume knowledge work where AI augments rather than replaces staff capability.
How we design and deliver automation
Our approach is grounded in 20+ years of business analysis and process improvement discipline. It’s structured, methodical, and built around your actual current state — not a theoretical version of it. We follow four phases.
Phase 01
Process discovery and current state analysis
We start by understanding how work actually flows — not how the process map says it flows. This means talking to the people doing the work, observing workflows in practice, and analysing system and communication data to surface the full picture including shadow processes, informal workarounds, and the work that never gets logged.
The output is a documented current state that your team recognises as accurate — because it reflects what actually happens, not what’s supposed to happen. This is the baseline everything else builds from.
Phase 02
Future state design
With the current state clearly mapped, we identify what should change before automation is applied. This means removing unnecessary steps, clarifying decision points, eliminating handoff delays, and designing the process so it’s lean, logical, and ready to be automated reliably.
We document the future state in a format your team can review and agree on before a single line of automation is built. This step prevents the most common automation failure: building something technically correct that nobody uses because the underlying process was wrong.
Phase 03
Automation design and implementation
With a clean future state agreed, we design the automation. Tool and platform selection is driven by your existing environment, your team’s capability to maintain the automation, and the specific requirements of the workflow — not by vendor preference or what’s currently fashionable.
We work with a range of automation approaches depending on what’s right for the use case.
Automation is built iteratively, tested on real work before full deployment, and documented clearly so your team understands what it does and how to maintain it.
Phase 04
Embedding and continuous improvement
Automation isn’t a one-off project. We provide a structured handover that includes staff training and guidance, documentation, and a clear process for what happens when the automation encounters an exception or needs updating.
We also establish baseline metrics during implementation so you can measure the actual impact — time saved, error reduction, throughput improvement — and use that evidence to justify extending automation to the next priority area.
What you receive at the end
Every automation engagement produces a complete, documented set of outputs:
- A current state process analysis documenting how work actually flows, including shadow processes and workarounds
- A future state workflow design agreed with your team before implementation begins
- One or more live automations, tested on real work and documented for your team to maintain, that can include agents that handle the repetitive work your team shouldn’t be doing
- Integration documentation covering the systems and tools involved
- Staff guidance and training materials appropriate to the automation deployed
- Baseline metrics and a measurement approach so you can quantify the impact
- A roadmap for extending automation to the next priority areas in your organisation
The scope and timeline of each engagement varies based on the complexity of the workflows involved and the number of automations being built. A focused single-workflow automation can be completed in two to three weeks. A multi-workflow engagement across a function or department typically runs four to eight weeks.
Who this is for
SMBs feeling the strain of manual processes
Small and medium businesses often reach a point where the manual effort required to keep things running starts to limit growth. Staff are spending hours on tasks that should take minutes. Errors creep in. Capacity is consumed by administration instead of value-creating work. Automation targeted at your highest-volume, most repetitive processes can release significant capacity quickly — without hiring.
Councils and public sector teams
Local government and public sector organisations carry heavy administrative and reporting workloads, often with constrained budgets and limited capacity to grow headcount. AI automation in areas like service request handling, compliance reporting, records management, and internal approval workflows can deliver meaningful efficiency gains while maintaining the audit trail and accountability standards that public sector work requires.
Enterprises standardising across teams
Larger organisations often have inconsistent processes across departments — each team has evolved its own way of doing things, creating fragmentation, handoff problems, and inconsistent outcomes. Automation at enterprise scale requires both process standardisation and careful change management. We bring both. We design the change that makes automation actually stick
Organisations that have tried automation before
If you’ve been through an automation project that didn’t deliver what was promised, you’re not alone. In most cases the technology wasn’t the issue — the process underneath was. We’re well suited to these engagements because we start from the process, not the platform. We’ll give you an honest assessment of what went wrong and what a well-designed second attempt looks like. For SMBs who are learning about automation, see how a practitioner builds automation from scratch
Have a question about automation?
What is AI workflow automation?
AI workflow automation uses artificial intelligence and integration technology to handle repetitive, rules-based tasks within a business process — reducing the manual effort required, improving consistency, and freeing staff for higher-value work. At its simplest, it might be an automated email response or data transfer. At its most sophisticated, it involves AI agents that can interpret unstructured data, make routing decisions, and interact with multiple systems.
What's the difference between automation and AI automation?
Traditional automation follows fixed rules: if X happens, do Y. It works well for structured, predictable tasks. AI automation goes further — it can handle unstructured inputs (like the content of an email or a scanned document), make decisions based on pattern recognition, adapt to variation in the data, and improve over time. Many workflows benefit from a combination of both: rules-based automation for the predictable parts, AI for the parts that require interpretation.
Do you fix the process before automating it?
Always. Automating a broken or inefficient process just produces broken or inefficient results faster. Our engagements always begin with a current state analysis and future state design before any automation is built. This is the step most automation providers skip — and the main reason automation projects fail to deliver what was promised.
Will automation replace our staff?
The honest answer is: it depends on how it’s designed. Our approach is explicitly oriented toward augmentation — freeing staff from low-value manual work so they can focus on judgment, relationships, and the work that actually requires human involvement. We also build change management guidance into every engagement so staff understand what the automation does, how it affects their role, and where they remain in control. Automation that staff don’t trust doesn’t get used.
What tools and platforms do you use?
We’re platform-agnostic. Tool selection is driven by your existing environment, your team’s capability to maintain the automation, and the specific requirements of the workflow. Common platforms we work with include, Zapier, Microsoft Power Automate, Make, n8n, and custom API integrations. For AI-specific automation, we work with a range of document intelligence, language model, and agent frameworks depending on the use case.
How do we know if a process is ready to automate?
A process is ready to automate when it’s clearly defined, consistently executed, and produces reliable outputs. If your team runs it differently depending on who’s doing it, or if there are frequent exceptions and workarounds, the process needs to be redesigned first. Our discovery phase identifies exactly where each workflow sits on that spectrum — and what work needs to happen before automation can be applied reliably.
How long does an automation engagement take?
A focused single-workflow engagement — from discovery through to a live, tested automation — typically takes two to four weeks. Multi-workflow engagements across a department or function generally run four to eight weeks. We’ll give you a clear scope and timeline estimate at the discovery session before any commitment is made.
How do we measure whether automation is working?
We establish baseline metrics during the process discovery phase — before any automation is built — so you have a genuine before/after comparison. Common measures include processing time per transaction, error or rework rate, staff hours per week on the automated task, and throughput volume. These metrics also become the basis for justifying the next automation investment within your organisation.