When a business begins an AI discovery process, the expected outcome is usually a list of automations, agents and other AI opportunities.
Sometimes discovery reveals an opportunity with even greater impact: creating the operating foundation the business needs for its next stage of growth.
We are seeing this pattern through our AI Navigator work with New Zealand SMEs. Once we map an end-to-end workflow, review the information moving through it and speak with the people doing the work, the same issue often appears. The ERP, MRP, inventory, job management or point-of-sale system does not match the business's actual requirements.
The signs usually appear in day-to-day operations. Staff re-enter information between systems. Important steps happen in spreadsheets. Someone maintains a private list because it gives them the view they need. Reports take hours to assemble. Paper, email and whiteboards carry parts of the process that the platform does not.
These are signs that the business has developed beyond its current operating setup. They point to an opportunity to give people better tools, improve visibility and create capacity for growth.
This is still a useful AI discovery outcome. Structured, accessible operational data is the foundation for more capable AI applications. If discovery exposes the work needed to create that foundation, it has identified one of the highest-leverage steps the business can take.
It also does not automatically mean replacing the core platform.
Two common starting points
Most SMEs we encounter sit somewhere between two positions.
The first group has never invested in a dedicated core system. The business may have grown successfully with spreadsheets, paper forms, shared inboxes and the knowledge of experienced staff. Each tool solved an immediate problem, and together they became an informal operating system.
This can work surprisingly well. Growth eventually creates new demands: transaction volumes rise, more people join the team and customers expect faster service. At that point, a dedicated platform can give the business a shared view of capacity, stock, job status or margin and make successful ways of working easier to repeat.
The second group has invested in a core platform, sometimes at considerable cost. The business may now have evolved beyond the requirements that originally shaped the decision, or discovered through use that some important workflows need different support.
The platform may cover finance but not production. It may manage jobs but provide a poor experience for field staff. It may hold inventory data but lack the connections needed for ecommerce, quoting or purchasing. In other cases, the right functions exist but were never configured or adopted properly.
The spreadsheets and manual processes that grow around the platform show where the next opportunity lies. They can guide a focused improvement programme without writing off the investment already made.
Workarounds are evidence
It is easy to dismiss a spreadsheet as a bad habit. During discovery, it is often more useful to treat it as evidence.
A spreadsheet may show that staff need a view the core system cannot provide. A paper form may reflect a user experience problem on the factory floor. A manual reconciliation may point to systems that hold related information but cannot exchange it. A weekly reporting exercise may reveal that the required data exists, but cannot be accessed in a useful structure.
These workarounds help define the real system requirements. They show what information people need, what decisions they are trying to make and where the current workflow needs more support.
That is why platform decisions should begin with the work, not a software catalogue. Map the process from the initial customer request through delivery, invoicing and reporting. Identify where information is created, changed, approved and reused. Then assess whether the current system supports those requirements.
Why AI discovery can reveal the opportunity faster
AI-assisted discovery can help analyse interviews, process maps, system exports, documents and recurring exceptions across a business. It can make repeated patterns easier to see and help teams compare what the process is meant to do with what happens in practice.
The value comes from combining those tools with operational judgement. A model can help organise evidence, but people still need to validate the workflow, understand why an exception exists and weigh the commercial trade-offs.
The result may be an AI use case. It may also be a clearer system requirement, a data-quality opportunity, an integration gap or a path to stronger adoption. Each is a productive outcome because it moves the business closer to an operating environment where AI can be applied reliably.
Five ways to create the next operating foundation
Once the opportunity is clear, there are five practical response paths. Many businesses will use a combination of them, staged at a pace their team can absorb.
1. Turn an informal process into a repeatable workflow
AI-assisted development tools can make focused internal applications more accessible to SMEs. A business may be able to turn a spreadsheet or paper process into a simple workflow for collecting information, applying rules, creating approvals and maintaining a reliable history.
This can give a growing team greater consistency and visibility without rebuilding the whole business platform. It addresses a defined need while preserving the systems that already work.
2. Unlock more value from the platform already in place
Sometimes the software already has more to offer. Cleaning master data, redesigning a workflow, introducing clearer ownership, improving training or enabling an existing module can remove a surprising amount of manual work.
This path can extend the value of the existing investment and improve adoption with less disruption than migration. It still requires clear process ownership and change management.
3. Connect best-fit systems into one workflow
Where specialist platforms each do their job well, integration can let the business keep that capability while creating a joined-up experience. Automations can move data between systems, trigger follow-up work and keep records aligned. AI agents can also complete defined tasks across this environment, provided they have clear permissions, reliable inputs and appropriate human review.
This gives employees more time for customers and higher-value work while making the end-to-end workflow visible.
4. Create an accessible operating, interface or data layer
A business may keep its core systems while adding a simpler layer through which staff can find information and complete common tasks. That layer can bring together data from several sources, present it in a usable way and provide controlled access to the tools people need.
We took this approach with a growing field-services business that already used specialist software for an important part of its work. The larger opportunity was to connect the workflow around it, including leads, job scheduling, field activity, timesheets, documents and customer handover. N16 built a role-based operations platform that connected with the specialist and accounting systems. The business retained the tools that worked well and gained a core operating experience designed around its team.
This is also where an AI operating system can create significant value. Claude, ChatGPT or another approved model can be connected to trusted business information so staff can ask questions, produce reports or initiate governed actions without navigating every underlying platform.
The technical design will depend on the systems involved. Some provide suitable APIs. Others require controlled extraction into a structured data store before the information can be used safely and reliably. Access controls, data ownership, traceability and review steps remain essential.
5. Invest in a new core platform when growth demands it
Migration becomes the stronger option when a new platform can materially improve the core workflow, make essential data accessible and support the scale or business model the organisation is moving towards.
It is a significant investment and an opportunity to design better ways of working. The decision should account for implementation cost, data migration, training, process redesign and the team's capacity for change. A well-defined requirements assessment helps the business select confidently and gives implementation partners a clearer brief.
Start with the diagnosis
The AI market is naturally organised around different types of intervention. Training providers help teams build capability. Automation specialists connect systems and remove manual work. Software developers build new applications. Each can be the right answer when the need is already clear.
N16 starts one step earlier.
Through AI Navigator, we map the end-to-end workflow, interview the people doing the work, assess the data and systems, and identify where investment can make the greatest difference. The recommendation may be staff training, better use of the current platform, an integration, a focused application, an AI operating layer or a full migration.
This gives the business a clear basis for choosing its delivery path. It also means the recommendation can match the organisation's current capability, budget and appetite for change.
How to choose the right path
Before deciding to improve, connect, extend or replace a system, an SME should be able to answer a practical set of questions:
- Does the platform support the critical end-to-end workflow?
- Is the underlying data accurate, consistently maintained and clearly owned?
- Can the business access that data through APIs or controlled exports?
- How much staff time is spent on re-entry, reconciliation and workaround reporting?
- What would make the platform easier for people to use, whether that is improved design, added functionality or training?
- What capacity could be released by reducing spreadsheets, paper and key-person dependency?
- What will the business need from its systems over the next three to five years?
- Which AI applications would become possible if the data were structured and accessible?
- What disruption and change capacity can the business realistically absorb now?
These questions turn a broad technology discussion into a business decision.
Build the foundation and the AI opportunities expand
The immediate outcome may be a smoother workflow, better integration or a clear decision about the core platform. The longer-term value is a business that can use its operational information more effectively.
Once data is structured, accessible and governed, AI can do more than draft content. It can help staff interrogate performance, prepare quotes, identify exceptions, compare documents, coordinate routine work and surface the information needed for decisions. Agents can take on defined administrative steps while people retain oversight of higher-risk actions.
For some SMEs, the right move will be a new ERP, MRP or inventory platform. For others, the best return will come from improving what they have, connecting it properly or adding an accessible layer over the top.
The important outcome of AI discovery is a clear view of the opportunity and an investment path that fits the business as it operates today. That gives the organisation a stronger data foundation, more capacity to grow and a practical route to the AI capabilities it wants next.
If your business has grown beyond the systems and workarounds that brought it this far, AI Navigator can help identify whether the next step is to improve, integrate, extend or replace them.
