Customer Spotlight
Twenty minutes, one AI agent and considerably fewer "I definitely told you I was traveling" conversations.
Graeme Barron travels a lot for work.
As General Counsel at GWI, he regularly travels from his home in the northeast of England to London, alongside international trips to offices including New York and Athens.
With that much travel in the calendar, plans can quickly become a lot to keep track of, especially when dates, timings or itineraries change.
Graeme and his wife would compare notes on upcoming trips and keep each other up to date as plans changed.
Rather than relying on memory or repeatedly checking different calendars and bookings, Graeme saw an opportunity to make things easier.
So he decided to see whether AI could help keep everyone up to date automatically.
From travel bookings to a weekly family update
Graeme had already built a personal "chief of staff" agent in Claude to help with things like meeting preparation and weekly planning.
He added one more task: every week, create an update covering his work travel for the next four weeks.
Using Perk's Model Context Protocol (MCP) connector, the agent pulls Graeme's upcoming itinerary information from Perk. It then combines that with information from his Google Calendar and Gmail to add context around each trip, not just where he's going and when, but why.
Every Friday morning, the agent prepares an email draft addressed to his wife.
The email includes upcoming journeys, hotels and arrival times, alongside the context behind each trip. It also produces a calendar view, making it easier to see at a glance when Graeme will be away.
Because the agent checks his travel again every week, it can also highlight anything that has changed since the previous update, such as a new arrival time.
Graeme reviews the draft, then hits send.
Built in around 20 minutes
There was no lengthy implementation project behind it.
Graeme connected Claude to Perk, Gmail and his calendar, then iterated on the instructions and format until the update contained the information he wanted.
"It honestly took no more than 20 minutes. It's really simple to do."
The agent itself can't send the email automatically, keeping a human in the loop. That means Graeme can review everything before anything leaves his inbox.
An hour of shadow work turned into a quick review
Before building the agent, Graeme and his wife would sometimes spend around an hour on a Friday evening comparing diaries and going through upcoming travel.
Now, that information is gathered automatically.
"It's just easy for me. I don't have to do anything."
The weekly overview has turned out to be useful for Graeme too. Instead of moving through his calendar week by week, he gets a single view of what's coming up.
Connecting travel to the tools you already use
Graeme's experiment shows the potential for personalizing the experience when you connect Perk to your AI tools. It wasn't a company mandate or a workflow built by the IT team. It was Graeme making the data work the way he needed to reduce the shadow work.
For Graeme, that happens to be a weekly email that keeps family life running more smoothly.
For someone else, it could be a calendar update, a trip briefing or another workflow entirely.
As Graeme puts it:
"I think the MCP is such a great thing. Just getting the information to where you need it. Super useful."
And, arguably, marriage-saving too.

:format(webp))