Can AI Fix It

Jack Harris, founder of Tally Ho Experiences, runs a seasonal London tour business from the Somerset countryside. Rather than training up temporary staff every peak season, he uses AI as an “elastic workforce” for the back office. That gives him and his assistant the headroom to take on complex, multi-supplier itineraries for the travel trade, without losing the personal touch.

Jack explains how getting the knowledge out of his head and into documentation laid the foundations, and how his “Love Note” workflow works. Guides leave a quick voice note after each tour, and an agent turns it into a personal follow-up email to the customer or travel agent within ten minutes.

Then it’s over to Catherine Warrilow and Paul Marden. Jack challenged them to set up an open-source agent harness, connect it to a messaging app and their own tools, and get it to draft an email, all in two hours. They take very different routes, with very different levels of stress. Jack picks a winner.

In this episode:

• Why founder-led businesses need to get what’s in their heads written down
• AI as an elastic back-office workforce for seasonal businesses
• Keeping customer relationships human when agents draft the emails
• What an AI “harness” is, explained with a brain in a jar and a horse
• Open source vs big-tech AI, and who benefits
• The rise of personal agents, and what they could mean for tourism in 2027

Chapters
00:00 Cold open
01:09 Welcome
01:58 Meet Jack Harris
02:51 Stepping out of the business, and documenting what’s in your head
05:23 Rebuilding the back-office stack
07:00 The elastic workforce
09:56 Training an assistant vs training an agent
12:12 Keeping relationships human
14:11 Escaping the tyranny of the desk
16:20 Love Notes: from guide voice note to follow-up email
19:24 What’s a harness?
21:01 The two-hour challenge
22:05 Paul’s build: Hermes, Telegram and HubSpot
26:27 Catherine’s build: Claude as co-pilot, Telegram and Gmail
32:21 Speed, models and billing
35:42 Software glued together vs genuine AI
38:16 Open source and who benefits
39:59 Personal agents and the future of travel
44:14 And the winner is…

Guest: Jack Harris, Tally Ho Experiences
Hosts: Catherine Warrilow and Paul Marden

What if your ticketing system told you to staff 12 people on Friday instead of 20, and then drafted the member email too? Ivan Diaz from Convious shows us how.

AI in business is almost a cliche, it’s discussed on LinkedIn and at conferences. But it’s tough to get started, not only is there a technical leap to make but it’s difficult to learn the tools if you don’t understand what the art of the possible is.

In this episode of Can AI fix it we meet Jonathon Taylor, Head of Marketing for the League of Adventure, the owners of Phantom Peak, the award winning immersive experience. Jonathan is part of a dynamic AI first team, who are discovering the many ways that AI can improve guest experience and save time and money - and he’s going to share 3 of those ideas with us.

If you already have an established, authentic brand with great reach on social media, using AI for content production poses a risk. So how can attractions like these make use of tools like Claude and Chat GPT?

In today’s episode we meet Nik Wyness, the Head of Marketing and Engagement at the Tank Museum, and we’ll be looking at ways to use AI tools to help monitor and respond to online customer reviews.

Catherine Warrilow and Paul Marden talk about how AI at visitor attractions can be transformational for business process, and more than just a bit of image editing or copy writing.