TMNZ Leverages n8n and AI for Productivity and Innovation in FinTech

How TMNZ scaled automation across teams and deliver 400 hours of productivity gains each month through n8n and AI.

Back to Case Studies

Background

Tax Management New Zealand (TMNZ) are tax revolutionaries. Approved by Inland Revenue, TMNZ has helped over 25,000 businesses of all sizes, from large corporates to sole traders, delivering innovative tax payment solutions to help New Zealand’s economy grow. As a purpose-driven business, TMNZ invests 100% of profits into Whakatupu Aotearoa Foundation, for New Zealand’s communities and the environment.

Challenge

TMNZ were looking for a way to better automate processes which would ultimately help their customers optimise tax payments and obligations. As a Fintech, they had to ensure any automation tool they chose would enable iron-clad security including self-hosting, version control, role-based access, and secrets management.

Teams were already leveraging tools including Zapier and Make to automate isolated workflows, but were heavily limited in what they were able to achieve. While larger players like MuleSoft and the Microsoft suite are popular in Aotearoa New Zealand, TMNZ wanted a solution that could be self-hosted without being cost-prohibitive.

With around 60 employees, they needed a tool that anybody across the business could use with minimal training. At the same time, TMNZ engineers needed to have fine-grain control and the ability to turn to code when needed, as well as a solution that would unite disparate tools in their tech stack.

“n8n hits the sweet spot. Our engineers can easily adapt workflows to suit our business needs. We can self-host locally, which from a data privacy perspective is really important for Inland Revenue, tax agents and businesses that we work with. And it is actually easy for less technical staff to help build workflows across Marketing, Finance and HR.” said Eric Troebner, CTO, TMNZ.

Solution

Today more than a third of TMNZ staff are using n8n to automate tasks from market research and HR processes, to agentic workflows, running around 150,000 executions each month.

It’s not all just about time savings either, a big proportion of the 400 hours savings gained each month (and growing) are through productivity improvements and enabling a cultural shift in the way that the organisation works.

n8n also satisfied TMNZ’s security requirements, enabling role-based access and version control across different production environments for reviews and controls. Alongside, SSO and read-only permissions allow stakeholders to stay up-to-date with workflows without touching the code base.

“Getting quickly to a point where we actually understand how everything hangs together is really powerful. Even if we then sometimes go back to the engineering team and implement workflows in our full stack platform, the logic is far easier to implement in n8n first,” said Troebner.

AI Integration

The team has recently started developing and productionising AI POCs, something that n8n wasn’t originally considered for. TMNZ runs an agnostic LLM stack with a number of foundational models including OpenAI, Azure, and Anthopic, atop a self-hosted infrastructure including PostgreSQL databases and vector stores. They now leverage n8n as part of this stack, for agentic AI use cases including invoice processing, data enrichment and enablement of Marketing, HR, Finance and Operations teams. The team has also created a number of MCP servers and AI agents that are more interactive. Using n8n, these can be called and combined with LLM models, using Open WebUI as a user experience layer.

Troebner says that n8n fits well into this AI stack based on TMNZ’s main criteria to have AI technologies modular, interoperable, observable and open source based where possible.

Impact

Already n8n has had a significant impact at TMNZ, with every POC delivering value to each team of the business. One of the most profound impacts the team has noticed is how quickly AI agents, workflows and experiences can now be spun up and combined.

TMNZ’s AI POCs have proved particularly impactful, saving up to an hour per execution. Troebner and his team are using the results to continuously iterate on workflows, to improve resilience and add new functionality as needed. Even negative results guide the team towards improved automations, illustrating which workflows are worthwhile investing time into to achieve a positive ROI from AI investments.

“Ultimately, an important differentiator is how good your organisational efficiency is, how good your automation is, how good your customer service is, all this is driven through the enablement of technology. That’s where n8n is helping to shift our culture from ‘this is too hard to solve, let’s just use a spreadsheet for now’ to ‘what can we do to systemise, automate and use AI to scale this’. Just understanding what's going on in a workflow is really powerful for all stakeholders,” said Troebner.

What’s next

With n8n up and running across the organisation, TMNZ is now productionising its environments more for scalability to run up to one million executions per month, as well as rolling out enhanced AI evaluations to ensure workflows are meeting FinTech industry standards and regulation in production.

“We have event-driven systems and we use n8n as part of that. We’re currently running on a relatively simple setup but we’re starting to scale this out, using queues, load balancers and monitoring to support the growing demand on n8n,” said Troebner.

Ultimately, the impact of n8n at TMNZ is just beginning.

“n8n will be a massive enabler of driving business transformation, being able to rethink your sales process, being able to accelerate your internal processes and freeing up a significant amount of time for our people to provide the best possible customer experience.

In the past, we’ve always hit a wall when it came to processes that weren't 100% deterministic, where you couldn't map out every single path, and now you can break through that barrier, with n8n.”

"n8n is helping to shift our culture from ‘this is too hard to solve - let’s just use a spreadsheet for now’ to ‘what can we do to systemise, automate and use AI to scale this’. Just understanding what’s going on in a workflow is really powerful for all stakeholders."

Eric Troebner
Eric Troebner

CTO, TMNZ