The original manifesto
North star
Protect the magic. Automate the machinery. Compound the learning.
The belief
The future of premium tour businesses is not replacing the experience with AI. It is building an intelligent operating system around the experience.
The customer does not come to Tally Ho for an automated tour. They come for judgment, charisma, timing, care, local knowledge, social sensitivity, and a feeling that the day unfolded in a way a generic product could not have produced.
That is the magic.
Everything around that magic is machinery:
- how the enquiry is understood;
- how the right product is recommended;
- how pricing and availability are checked;
- how a quote is drafted;
- how a guide is briefed;
- how post-tour follow-up happens;
- how reviews are interpreted;
- how partner intelligence is captured;
- how operational issues become process improvements;
- how content ideas become publishable assets;
- how decisions are surfaced before they become bottlenecks.
That machinery should become intelligent.
What AI-first means here
Tally Ho should become an AI-first tour company, not by making the tour artificial, but by making the company around the tour increasingly legible, automated, and self-improving.
AI-first means that when useful information enters the company, it should not disappear into an inbox, Slack thread, WhatsApp group, memory, or one person's head. It should become part of a loop:
- captured as evidence;
- cleaned into a usable record;
- interpreted into candidate insight;
- reviewed when judgment matters;
- promoted into knowledge, SOP, training, content, product improvement, or backlog;
- reused the next time a similar case appears.
The sacred layer
The tour itself remains human-led.
Human guides own:
- live guest care;
- interpretation of mood and energy;
- humour, warmth, and personal judgment;
- safety judgment in the street;
- taste and pacing;
- the stories, pauses, detours, and improvisation that make the day memorable.
Agents can help brief, prepare, learn, and follow up. They should not flatten the experience into a script.
The machinery layer
The machinery should become increasingly agent-readable and agent-operable:
- sales and enquiry triage;
- quoting and pricing support;
- partner relationship intelligence;
- product and content knowledge;
- review intelligence;
- guide notes and coaching signals;
- operational exceptions;
- weekly management reporting;
- content and authority building;
- data quality and source-of-truth checks.
The aim is not "more automation" for its own sake. The aim is less lost context, fewer repeated decisions, faster handling of routine work, clearer escalation, and more compounding learning.
The company habit
Every new workflow should ask:
Did this make the next similar case easier, smarter, faster, or more consistent?
If not, it may be a useful one-off, but it is not yet part of the AI-first company.