All Case Studies

    Case Study

    From "I need more people to be the master" to 18 leaders building their own AI assistants and a company-wide AI fleet

    18 managers and leaders across Papua New Guinea and Australia, trained hands-on, then built one-on-one.

    Problem: Operational VisibilityProblem: AI Capability
    R&A Marine Services logo

    Company

    R&A Marine Services

    Industry

    Shipping, telecommunications and energy, Papua New Guinea and Australia

    Problem

    Too much work came back to the CEO, and his team across Papua New Guinea and Australia had little experience with AI.

    What we built

    • • Claude Teams set up for the company, with the business's own rules, context and voice
    • • Hands-on training for 18 managers and leaders, in four groups
    • • One-on-one builds, each one solving a real bottleneck for that person
    • • A company-wide AI fleet map and builders kit, so every assistant is built the same way and stays visible

    Outcome

    • ✔ 18 managers and leaders trained
    • ✔ 12 assistants in daily use, built with 10 people
    • ✔ A second group of 13 lined up

    The Situation

    Where things stood

    R&A Marine Services is a family-led business working across Papua New Guinea and Australia, employing about 300 staff across shipping, telecommunications and energy projects.

    Dem, the CEO, had already seen what AI could do for his own work. But too much still came back to him: reports, documents and processes that others were slow to finish or that he ended up doing himself. So he came to Nina with a simple ask: train my people.

    "I'm doing triple the work. I want my people to be more capable than this, so I have to do less."
    Dem, CEO, R&A Marine Services

    The Bottleneck

    What was actually in the way

    • The managers were spread across offices and remote sites in Papua New Guinea and Australia
    • Experience ranged from daily AI users to people who had only ever used it like Google
    • Everyone was flat out with client work, so time for learning was tight
    "Mostly all of us have it on our phones, and we're using it like Google."
    Sanjoy, CTO, R&A Marine Services

    The Fix

    What we built

    1. Setting up Claude Teams

    Nina spent a day setting up Claude Teams for the company, with the business's context, voice and brand built in. She also set the house rules: company data stays in company systems, and a person signs off anything client-facing, financial, legal or safety-related.

    2. Training in four groups

    18 managers and leaders across Papua New Guinea and Australia were split into four groups of about five, with the leadership team going a week ahead. Each 90-minute session was hands-on, on their own keyboards. The thread running through all four sessions was one real thing each person needed: their own daily or monthly report.

    • Session 1: Platform introduction and AI. How Claude works, how to connect it safely to company data, and teaching it the business.
    • Session 2: Unpacking Claude and Projects. Projects, and teaching Claude each person's voice and report format.
    • Session 3: Skills. Turning how they do their report into a process skill, so Claude knows their standards and their way of working.
    • Session 4: The build framework. The six components of any AI assistant, tying everything together into their first assistant, and choosing their next bottleneck to solve.

    Every session was recorded into a training playlist, with student guides, handbooks and builder packs. So whether people joined live or caught up afterwards, they had everything they needed.

    3. Switching to one-on-one

    Session 5 was meant to be a half-day builder session for each group. But with internet dropping out and cameras off, it was hard to see who was keeping up, and speaking up in a big group didn't come easily for everyone. So Nina changed the plan. She sat down with the leaders to name each person's biggest bottleneck, then built with each person one-on-one. The connection was better, the questions came, and the biggest surprise was how much everyone had already learned.

    4. Building a company-wide AI fleet map

    Every assistant now goes into one fleet map that the whole organisation can see: who's building what, how well it's working, which bottleneck it solves, and how much time it's giving back. A builders kit guides everyone to build the same way, and a copy of every file and process map is saved in Dropbox. That makes it easy to review every assistant, spot where one person's assistant could help someone else, and keep everything safe, responsible and in line with company policy.

    Tool layer

    • Claude Teams: the company's shared Claude account
    • Dropbox: for copies of every assistant, skill and process map
    • Asana: for tasks and follow-ups
    • Otter: for voice notes and walkthroughs
    • Microsoft Teams: for remote training and one-on-ones

    The Outcome

    What changed

    • ✔18 managers and leaders trained across four groups
    • ✔12 AI assistants in daily use, built with 10 people, and more being tested
    • ✔Eva's month-end accounts payable report went from a full day to about 15 minutes
    • ✔Martin's nightly site report went from two hours to 15 minutes, giving him up to 10 hours back a week
    • ✔Every assistant is visible in one company-wide fleet map, with its time saved tracked
    • ✔A second group of 13 is lined up, and a company-wide plan has started with HR

    The Human Payoff

    What it actually meant

    Nobody is being replaced. The team is reshaping how work gets done: they're faster, more confident, and spotting and fixing their own bottlenecks. Dem gets a team that can carry the work. And with the fleet map, everyone can see where AI is moving the needle and where to focus next.

    "Actually, I'll get Claude to do it."
    Miranda, Chief Compliance and Legal Officer, R&A Marine Services
    "This is the whole point of this: is to help people set things up so that their job is easier."
    Dem, CEO, R&A Marine Services

    Client Feedback

    "I'm really enjoying it. Like for myself, I think it's great, and I think for the company, I can see endless opportunities."

    R&A Marine Services logo

    Miranda, Chief Compliance and Legal Officer, R&A Marine Services

    Papua New Guinea and Australia

    More Results

    Related case studies

    Is your team half-using AI, and you can't see where it's all going?

    People are trying different tools and building their own things. But you can't see who's building what, what's being doubled up, or whether any of it is actually any good. If that sounds familiar, let's have a chat. I'd love to help you train your team and bring it all together, so you can see exactly where AI is moving the needle.