AI Enablement with Claude Cowork
How Hamilton & Waikato Tourism made a chat window into a working infrastructure
The Client
Hamilton & Waikato Tourism is the Waikato's official Regional Tourism Organisation, funded by four councils to market the region to the world. Caves, surf breaks, a hobbit village, and 400 kilometres of cycle trail. A small marketing team in Hamilton sells all of it to domestic and international visitors.
Like most marketing teams, they were already using AI. Also like most marketing teams, they were using it one conversation at a time. Open a chat, ask for a draft, close the chat, start again tomorrow.
The Challenge
Natasha Chevtchenko is the only person on digital marketing in a team of six. She produces six to ten posts a week across social channels that each have a different audience and a different voice. Reporting takes her about a day, pulling and interpreting numbers from Google Analytics and Meta while the platforms keep changing underneath her.
She had taught herself AI by using it. With no training and no developer, she had built custom AI workflows, one for LinkedIn posts, and an agent wired into Asana to sort out her tasks and set her day. Every attempt ran into the same wall:
Every tool started from nothing, and none of them knew anything about each other.
The brief, the personas and the brand guidelines had to be re-explained in every session.
The copy still came back off brand, so the time saved drafting went straight back into rewriting.
Everything Natasha and her team had built lived in a browser tab, in tools that could not see their files or each other's work.
More self-teaching was not going to close that gap. She wanted a straight answer on three things: what agentic automation actually looks like in a marketing workflow, what it takes to get there, and whether any of it is possible with no developer and no plan to hire one. Natasha was after a single setup that already knew the brand, could reach the team's files and could be built and run by a marketer.
"This is something I tried to understand, and I was like, oh, I probably need to learn a little bit more."
The Journey
Enlighten Designs' Principal Technologist, Sam McKoy, worked directly with Natasha on why her existing setup kept hitting the same wall.
The first move was a diagnosis, not a lesson. The existing tools were not stalling because of bad prompting. The writing was never the weak point. The forgetting was. Everything else followed from that.
Sam also showed Natasha his own everyday use of Claude, and how he takes the repetitive parts out of his week using Claude, Claude Cowork, connectors and skills. Everything he demonstrated was something he had built for his own work.
The Solution
From there, the two of them worked through the four building blocks that turn Claude from a chat window into working infrastructure, each one mapped against Hamilton & Waikato Tourism's own workflows rather than a demo:
A written brief Claude reads every time. So it already knows the organisation, the region and the brand guidelines before anyone types a word. The context gets written down once, not re-explained daily.
Skills. Repeatable instructions for the tasks that come round every month, written once and reused, so the tenth LinkedIn caption is as good as the first.
Projects. Keeping the work for a campaign together, with its own context, so nothing gets lost when someone else picks it up next month.
Scheduled tasks. Work that runs without anyone opening a laptop. That is where a marketing team of this size gets real leverage.
Because every block was worked against the team's own jobs, the question was never "what could this do" but "which of these is worth doing first".
Enlighten did not take the work away. Sam gave Natasha the four building blocks, and Enlighten Designs Marketing Associate Drishti Ratna showed her what they look like inside a marketing role, so she could apply them to the rest of her work herself.
"I came in thinking we were already doing this fairly well. I left with a list of things we should have set up a year ago."
"It gave me a great guide on where to start and what to do next."
The Results
Within a week, and without any prompting from Enlighten, Natasha had:
Moved to the paid plan
Written her first skill
Started building her first scheduled task
She could also name which of her recurring jobs justify a skill, which belong in a project, and which tasks of her monthly cycle should run on a schedule. Claude Cowork was the piece she saw the most immediate use for, given the number of channels she runs and the different message each one needs.
The session gave Natasha enough to start, and she was in Claude the same day. Within a week she had a paid account, had built her first skill and had scheduled her first task. The context that used to be re-explained every session now lived in the skill, and the task ran without her opening a tab.
Natasha came back with sharper questions that only surface once you're building. Some were about wiring it in: could Claude connect safely to the organisation's SharePoint, and how should she bring the wider team in. The rest were about where context should live. Whether brand guidelines belonged in an organisation-wide instruction or a shared skill, and whether a Project retains memory or can see chats and tasks outside it.
Those are the questions of someone building, not someone still deciding whether to start.
What happens next
Natasha is building against the shortlist. The written brief and the skills for the weekly jobs come first, because those pay back fastest.
Enlighten is working through her questions on settings, desktop security, SharePoint and organisation-wide guidelines. That is the groundwork for bringing the wider team on.
We are measuring the next full monthly cycle against the starting point above: how long reporting takes once the numbers run through Claude, how much of the six-to-ten-posts-a-week content load moves through skills, and which of the four blocks earned their place. We will check back in after the first full cycle and share what we find.
"We were already using AI every day. Enlighten showed us how to make it actually work for our team."
Project Overview
Client: Hamilton & Waikato Tourism
Sector: Regional tourism marketing
Region: Waikato, New Zealand
Team: Six people, one digital marketing lead
Purpose: Claude enablement
Claude products: Claude, Claude Cowork, skills, projects, scheduled tasks, connectors
Features used: Skills, projects, scheduled tasks, connectors
Account Manager: Doug Wilman
Delivered by: Enlighten Designs. Sam McKoy, Principal Technologist, and Drishti Ratna, Marketing Associate
Enlighten Designs is a member of the Anthropic Claude Partner Network. We help New Zealand organisations move from experimenting with Claude to running on it: enablement, deployment and ongoing engineering support.