Contour was already using AI. Around half the team hold Claude licences, and people had built working skills for planning reports, cover letters and RFI responses. What they lacked was a map. Flowstate's AI Discovery gave them one.
The Problem
Contour's team was already moving on its own. AI was part of the working day across the firm, with around half the team holding Claude licences and people building their own skills for planning reports, cover letters and RFI responses. An internal AI Integration Group, led by a director, was drafting an AI Use Policy and user guide. On Flowstate's AI Maturity Index the firm scored 56% (Making Progress): strong on personal use (72%) and leadership (69%), weak on Systems and Information (32%).
Underneath, much of the project and client administration still ran through people and spreadsheets. Half of those surveyed said their systems are mostly connected manually, which is common in a growing firm and left less time for client work.
Contour wanted a fact-based view of where AI adds value, a clear first move ranked by impact and effort, a way to absorb rising workloads without leaning on the same senior people, and clear rules on what AI can be trusted with. The expertise was there. The map was not.
“It was useful to see some immediate things identified that could contribute to an increase in efficiency and a smoothing out of our workflow.”
Oliver Ryan, Town Planner, Contour Consultants
Where Time Could Be Won Back
- Project administration and time recording that sits around every job.
- Billing cycles that take coordination across the team each month.
- Report preparation, where much of the structure repeats from job to job.
- Review and sign-off, where senior time is in high demand.
- Keeping project status, budgets and client enquiries visible in one place.
What We Delivered
Discovery. Participants completed a pre-workshop questionnaire, then the AI Integration Group joined an August 2026 workshop, run as two three-hour sessions. We mapped the invoicing cycle, the tech stack (Abtrac, Microsoft 365, SharePoint, Claude, Copilot) and the points where work stalls.
Roadmap. Presented in September 2026, a 12-month plan that separates the quick wins from the more transformational work:
- Quick wins: tidy the data foundations and put AI to work on the repeat tasks the team does every week, using the tools already in place.
- Next, automate the routine, starting with the invoicing cycle and the admin that sits around each project.
- Transformational work over the year: AI agents handling report writing at scale, with a CRM as a key component, bringing enquiries, projects and client relationships into one connected view.
Rollout approach. A beta group tests first, with sign-off before anything goes practice-wide. Confidentiality is protected, and professional judgement, VCAT work and expert evidence stay with the planner.
What Changed
- The leadership group now shares one view of where AI helps, with invoicing named as the first build.
- A 56% baseline on the AI Maturity Index to measure progress against.
- A staged 12-month roadmap with beta testing built in, and clear boundaries on what AI is trusted with.
- Strong marks from the workshop (two of five responded, scored out of 10): value 8.0, clarity on next steps 9.0, facilitation 9.5, likelihood to recommend 9.0.
“Flowstate got into the detail of our business quickly. They understood the pressures that come with growing a firm like ours, and turned all of our conversations into a clear roadmap, ranked by what will make the biggest commercial difference. They were easy to work with, communicated well the whole way through, and I'd happily recommend them.”
Mel Crawley, Head of People and Operations, Contour Consultants
The Compounding Effect
Discovery is where AI work either gets sequenced or gets scattered. Contour's team arrived with real momentum and individual wins. They left with a sequence, a baseline and a rollout approach the whole firm can follow.
The expert judgement stays with the planners. The repeat work is where AI goes first.

