Your AI project, start to finish.
This page walks you through exactly how an AI project runs with us: what happens in each phase, where the decisions sit, when the price is fixed, how you can pay, and what it takes to run once it is live. No jargon, no surprises. By the end you should know what to expect at every step.
Six phases, two decisions.
The mapEvery AI project follows the same shape, whether it is one automated process or a platform build across the whole business. Only the size of each phase changes. You make two decisions along the way: whether the idea is worth scoping, and whether the fixed price is worth building. Everything else is our job.
See the capability
90 minutesNo charge, no obligationBefore anything is scoped or priced, we show you what modern AI can genuinely do against your own workflows: drafting, summarising, answering from your documents, automating handoffs between systems. A live demonstration, not slides. The point is to replace hype with something you have seen working.
- Live demonstration against workflows like yours
- We listen to where time actually goes in your business
- Honest talk about what AI does badly, as well as well
- The right people in the room for 90 minutes
- Two or three processes that eat your team's time
Discovery & readiness
Typically 1 to 2 weeks elapsedAI amplifies whatever it is pointed at, so before we design anything we check the foundations: where your data lives, who can see what, how your Microsoft 365 and identity are set up, and what your security and privacy obligations are. If permissions are loose or data is scattered, we tell you now, because fixing that first is cheaper than fixing it later.
- Your data: where it is, what shape it is in, who owns it
- Identity and permissions, so AI only sees what it should
- Licensing you already pay for that can carry part of the load
- Privacy, security and any industry obligations
- Access to review the environment (read-only)
- A short session with the people who run the process today
Scoping: the plan and the price
Price is fixed hereWe design the solution properly: what is in, what is out, which systems it connects to, where it will be hosted, and how many hours each phase needs. That becomes a written scope with a fixed price and a payment schedule. Nothing in this document is vague, because this is the document you decide on.
- Each phase, with the hours assigned to it
- Every system we connect to, and how
- Hosting, licensing and running costs, itemised
- What is explicitly out of scope
- The payment schedule, milestone by milestone
- A decision-maker to walk the scope with us
- Questions. Ask us anything that is unclear
You now know exactly what it costs, what it does, and what it takes to run. If the numbers do not stack up, stopping here costs you nothing further, and you keep the readiness report and the scope. If you go ahead, the build starts against the plan you approved.
Build & integrate
Pilot first, alwaysWe build a working pilot for a small group first, then extend it. You see progress at a regular demo rather than waiting for a big reveal. This is where connectors are wired to your systems, and on larger projects where the data lake is stood up, so the AI works from clean, governed data rather than whatever it can find.
- Working pilot in real hands early
- Connectors built to your line-of-business systems
- Data lake stood up where the project needs one
- Regular demos, and course-corrections while they are cheap
- A pilot group of 5 to 10 people who will actually use it
- Feedback at each demo, even one line
Go-live & training
Everyone, not just the pilotSecurity review, then rollout to everyone who needs it, with hands-on training so people actually adopt it. Adoption is where AI projects live or die, so training is built into delivery rather than sold as an optional extra.
- Security review before anything goes wide
- Rollout in stages, so support stays personal
- Hands-on training sessions for your team
- Written guides your people keep
- Time in calendars for training sessions
- A visible sponsor. Adoption follows leadership
Run & improve
OngoingOnce live, the system is monitored, patched and supported like the rest of your environment. We measure what it is actually saving you, review it with you, and improve it in small steps. Running costs are known before you go live, because they were itemised in the scope at phase 3.
This is also where the value compounds. The first automation is rarely the last: once your team sees one process running itself, the next three candidates surface on their own. Most clients run this phase as an AI & Automation Uplift Plan: a monthly plan with build hours baked in, so improvements and new automations keep landing without a new project each time.
- Monitoring, updates and support under your managed agreement
- Usage and value reviewed with you, with real numbers
- A backlog of improvements, done in priced, agreed steps
- New automations added through your uplift plan's monthly build hours, or quoted separately when they are bigger than the plan
Payment options.
Agreed at scopingThe payment structure is part of the scope document, so you approve it alongside the price. Three shapes cover almost every project. Published guide: bespoke AI builds run from $15k to $250k+ excluding GST, scoped to the job.
Hosting and running costs, named up front.
No surprisesAn AI system has ongoing costs the way a vehicle has fuel. Every one of them is itemised in your scope before you decide. These are the categories to expect.
AI & Automation Uplift Plans.
Improvements, every monthA project gets you live. An uplift plan keeps you moving: your automations managed and monitored, build hours every month for improvements and new automations, and a regular meeting where the numbers are on the table. Every automation we manage carries documented ownership, testing, monitoring and an estimated value or time-saving measure, so you always know what each one is worth. All prices are NZD and exclude GST.
- AI readiness baseline and a 90-day roadmap
- AI acceptable-use policy for your business
- Management of up to 2 simple automations
- Microsoft 365 and Copilot guidance
- 3 uplift hours every month
- Quarterly staff training
- Monthly performance report
- Business-hours support
- Management of up to 6 automations
- 10 uplift and build hours every month
- One internal AI knowledge assistant
- Microsoft 365, CRM, accounting and PSA integrations
- Monthly ROI and optimisation meeting
- Quarterly training
- Priority support
- Management of up to 15 automations or AI agents
- 24 uplift and build hours every month
- Departmental copilots and custom API integrations
- Security, privacy and ISO 27001-aligned governance reviews
- Monthly executive steering meeting
- Quarterly roadmap workshops
- Priority incident support and AI cost management
- 12-month initial agreement
- One-off onboarding: $2,500, $5,000 or $10,000 respectively, plus GST
- Microsoft Copilot, ChatGPT, automation-platform and API consumption costs are billed separately, and itemised so you can see them
- Unused monthly build capacity does not roll over
- Complex projects exceeding the monthly capacity are quoted separately, through the scoping process above
- Ad-hoc improvement work outside a plan is billed at $269 + GST per hour
- Move between plans as your automation count grows; the plan should follow the value, not lead it
Your systems, joined up.
Connectors & the data lakeOn larger projects the real value comes from connection: the AI reads from and writes to the systems you already run, through governed connectors, with a data lake in the middle so answers come from clean, current data.
A small engagement, and a big one.
Sample plansThe same six phases, at two very different sizes, with hours tagged to each phase so you can see where the effort actually goes. Most first projects look like the left column. The right column is what a whole-of-business build with system integrations and a data lake looks like.
What you can hold us to.
Both ways- A fixed price before any build starts, and no change without written agreement
- A working pilot in real hands early, not a big reveal at the end
- Every hosting and running cost itemised before you decide
- Security and privacy reviewed before anything goes wide
- An honest "stop here" if the value is not there
- The people who actually run the process, available for short sessions
- A decision-maker at the two decision points
- A pilot group willing to use the thing and say what is wrong with it
- Time in calendars for training when go-live comes
Start at phase 1.
It costs nothing.
The 90-minute capability session is free and obligation-free. You will leave knowing whether AI is worth pursuing for your business, and if it is not, we will say so.
