For tour, activity, and experience operators

Map the workflow
before the AI.

Most operators don’t need another tool demo. They need to know which workflow is worth fixing, what AI can safely help with, and where a person still needs to check the work.

MAP A WORKFLOW →Take the scorecard

Scroll

Pressure release

Done with AI hype?

Tell me your pain points. We’ll map the high-value workflow and get it implemented.

Do AI in my company

Messy problem → mapped workflow → working system.

“I would have no idea how to connect seven APIs into one business intelligence unit to do what you did.”
Bruce Rozard, Co-founder of Arival, on the Experience This Podcast

Built inside

Tally Ho

A live experience business where AI is tested against bookings, marketing, sales distribution, reviews, finance, guide notes, and remote-owner drag.

Human ↔ machine

Product judgement, technical fluency

Enough technical understanding to ask the right questions, steer AI-assisted building, and keep the finished system pointed at the human problem.

First move

Workflow mapping

Find the awkward workflow, decide where judgement belongs, then set the guardrails before any agent touches the work.

The Operator Reality

FIG. 01

The first job is finding the drag.

Seasonal pressure, inbox volume, unused reviews, fragmented context, and handover gaps are where the cost usually shows up. Map those first. Then decide what AI can safely help with.

Seasonal pressure

Revenue is seasonal. Staffing, cash, and customer questions still need answers every week.

Founder bottlenecks

The pricing logic, supplier quirks, guide judgement, and tiny exceptions live in one head, usually yours.

Inbox drag

Enquiries, amendments, advisor requests, partner replies, follow-ups. The volume takes hours away from sales, guests, and operations.

Reviews left unused

Reviews contain product signals, objections, guide feedback, and sales language. Without a workflow, you don't use that information.

Fragmented tools

Guest, booking, payment, and follow-up context sits in separate systems. You lose time checking multiple places before making a decision.

Handover gaps

Guest context, guide notes, exceptions, and last-minute changes move between people and tools. Important details are easy to miss.

If one problem is already clear, map that workflow. If the whole operation feels blurry, take the scorecard first. Take the scorecard →

The Thesis

FIG. 02

The experience is the moat.
The machinery around it is fair game.

The guide’s timing. The feel for the group. The judgement that knows when to move, pause, skip, or linger. That part shouldn’t be flattened into a chatbot.

The machinery around it is fair game: reviews, inboxes, reporting, CRM follow-up, guide notes, handovers, and the decisions that currently rely on someone remembering the weird detail at the right moment.

The point is simple: give the team more time and attention for the part guests actually came for. Map the workflow →

In Practice

FIG. 03

The proof is the operating mess.

Inside Tally Ho I’ve been building around the unglamorous work that usually gets left until later: inboxes, reviews, finance, reporting, guide notes, CRM, and team knowledge. It’s less glamorous than the keynote examples, which is usually why it works.

What this looks like in the lab

The lab is a real operating business. When something breaks, it breaks into my week.

SystemNoteStatus

Enquiry and advisor drafts

Inbound messages parsed, context gathered, replies drafted, and a human still checks anything a guest might see.

LIVE

Finance and reporting context

Stripe, Xero, and bank feeds read in context, so money questions stop hiding across three tabs and a vague feeling.

LIVE

Review intelligence

Guest reviews turned into guide coaching, product signals, objections, and usable sales language instead of a Friday skim.

LIVE

Persistent business memory

The system keeps business context available, so I don't start every morning by explaining the same operating context again.

LIVE

CRM intelligence

Drafts that already know the guest history. Quotes that went quiet get chased before I remember they exist.

BUILDING

Guide and team knowledge

Guide notes, shift context, and the small operating details that normally vanish between WhatsApp, memory, and hurried handovers.

BUILDING

Start with the workflow: what moves, what breaks, what data matters, who checks it, and where the risk sits. Map your workflow →

Why Me

FIG. 04

I’m not an AI consultant who wandered into tourism. I’m an operator who dragged AI into the work.

tallyho.com is the London tour company I’ve spent fifteen years building: boutique private tours, high-net-worth families, corporate events, bicycles, pub tours, taxi tours, out-of-town days, and premium British motors.

Four years ago I left London for the Somerset hills, which was a brilliant way to battle-test the business. If the team needed me for every repeated question, the system had failed. That is where the automations started, then the AI.

The work started in the machinery around the guest experience: marketing, sales distribution, guest communication, finance, reviews, guide notes, and handovers. The useful question isn’t “what can the model do?” It’s “what does the team need to remember, decide, sell, check, or hand over when nobody has spare attention?”

I also picked up the travel-tech scar tissue at Whym, a Silicon Valley-backed last-minute ticketing startup. I was London manager, then acting CEO. We were early enough that some major attractions still wanted guests to print a PDF. The future had arrived, provided you could find a printer. Useful, though: it taught me to translate between human problems, product questions, and software.

15 years

tallyho.com

A London tour company with private clients, corporate work, bikes, pubs, taxis, day trips, and premium British motors.

4 years remote

Founder out of the room

If guest context, guide notes, repeat questions, or handovers still needed me in London, the system was not good enough.

Human ↔ machine

Judgement into systems

Guide instinct, guest context, supplier quirks, and founder decisions turned into workflows, checks, and usable software.

Talks, podcasts, workshops

I speak and run working sessions for tour, activity, and experience businesses on practical AI, review intelligence, OpenClaw, and Hermes. The work is turning operator judgement into systems the team can trust: scorecards, handovers, and clear approval points. Less AI theatre, more useful machinery. Talks and booking →

Start The Conversation

FIG. 05

Send the awkward bit.

Bring a rough problem, a specific workflow, or a half-formed enquiry. If there’s a fit, I’ll point it at the right next step: mapping session, scorecard, sprint, talk, or partnership conversation. The next step should make the work clearer, not add another platform to the pile.

I know the awkward bit

Name the process, decision, or handover that keeps costing time. We can map it from there.

Map it

I can feel the drag

Use the scorecard if the operation feels blurry and you want a structured starting point.

Take the scorecard

I want to talk it through

For sprint, speaking, partnership, or broader consultancy questions, use the form. Rough is fine.

Start the conversation

Speaking or partnership? Use the same form and pick the subject.

Or find me on LinkedIn. I check DMs more than I should.

Not ready to get in touch?

Take the AI readiness scorecard Prefer to read first? The Articulation Method, free, 7 pages