AI Impact | Part 1: Working with AI
home
Glossary
Fundamentals
AI (Artificial Intelligence)

Software that can process information, recognise patterns and generate content in ways that resemble human thinking.

LLM

Large Language Model. The technology behind ChatGPT, Claude and Gemini. Trained on text to understand and generate language.

Generative AI

AI that creates new content: text, images, code, audio or video. This is the category that includes all the tools discussed today.

Hallucination

When AI confidently says something incorrect. It does not know it is wrong. Always verify facts, especially names, dates and statistics.

Token

The unit AI uses to measure text. Roughly one token per word. Usage limits and pricing are based on tokens processed.

Context window

How much text AI can process at once. A larger context window means you can give it more background and longer documents.

Prompting and workflow
Prompt

What you type in. The instruction, question or brief you give AI.

Prompt engineering

The skill of writing prompts that get useful, specific results. Includes giving context, specifying format and iterating on output.

Project / Custom GPT

A saved workspace where context, files and instructions persist across conversations. Claude calls these Projects. ChatGPT calls them Custom GPTs.

System prompt

Hidden instructions that shape how AI behaves in a project or custom tool. Sets the role, tone and constraints before any conversation starts.

RAG

Retrieval-Augmented Generation. When AI searches through your uploaded documents to find relevant information before generating a response.

Visibility and discovery
GEO

Generative Engine Optimisation. Writing and structuring content so that AI recommends your business when travellers ask questions.

AIO

AI Optimisation. A broader term for the same concept as GEO: making your online presence visible to AI-powered discovery tools.

SEO

Search Engine Optimisation. The traditional approach to appearing in Google results. Still important, but GEO is the emerging complement.

Alt text

A text description attached to an image in a website's code. AI and screen readers use this to understand what the image shows.

AI capabilities
Agentic AI

AI that can browse websites, use tools and complete multi-step tasks independently, not just respond to a single message.

Vibe coding

Describing what you want in plain language and letting AI write the code. No programming knowledge needed.

Multimodal

AI that works with more than just text: images, audio, video and documents. Most modern tools are multimodal.

Fine-tuning

Training an AI model on your specific data so it learns your patterns, terminology and preferences. Mostly for larger organisations.

API

Application Programming Interface. A way for software to communicate with AI directly, used for building tools and automations.

Automation

Connecting AI to other tools (email, booking system, social media) so tasks happen automatically. Tools like Zapier and Make enable this.

09:00

Where Are
You Now?

Five dimensions of AI maturity. Five minutes. Your result will guide which exercises to focus on later in the workshop.

DTTT AI Maturity Model
AwarenessExplorationAdoptionIntegrationTransformation

Why this matters

The key data points from the presentation. Save these for later.

26%

of 25-34 year-olds use AI planning tools as their primary source of travel inspiration.

Source: GSIQ TACTIQ, 2025

-30%

drop in organic traffic reported by tourism organisations since AI search emerged.

Source: DTTT Member Survey

+30%

increase in AI referral traffic. Visitors still arrive, but the path has changed.

Source: SEO Clarity Analysis

searchSEO becomes GEO

Search Engine Optimisation is no longer enough. Generative Engine Optimisation means structuring your content so AI can find, understand and recommend your business. Your website is no longer the destination. It is training data for AI.

forumSearch becomes conversation

Travellers are shifting from keyword searches ("things to do in Småland") to conversational queries ("what can I do with a young family during Easter in Småland?"). AI synthesises answers from your website, reviews, social media and other sources. You cannot control the answer, but you can influence the inputs.

These cards are here as a reference. The facilitator covered the full context in the presentation. The rest of this tool helps you respond to these shifts practically.

09:50

AI Visibility

Travellers increasingly ask AI for recommendations instead of scrolling search results. AI synthesises everything it can find about your business and decides whether to recommend you. This section covers what determines whether you appear in that answer.

From search to conversation
person
Traveller
forum
Asks AI
auto_awesome
Synthesises
hotel
Recommends

This is called Generative Engine Optimisation (GEO). It applies to your descriptions, your website content, your articles and the way your information is structured online.

Take-home framework

Five things that make your content visible to AI

These apply everywhere: your listing description, your website pages, your blog posts, your social profiles. Use them as a checklist for everything you publish.

groupAudience

Who is this for? Families, couples, solo travellers, specific nationalities?

spaExperience

What does it feel like? Sensory details, pace, atmosphere, not just features.

pin_dropPractical detail

Location, duration, distance, languages spoken, what is included?

fingerprintSpecificity

Concrete details that only apply to you. Not generic claims any business could make.

favoriteEmotional pull

Does it create a feeling? Language that makes someone want to experience this.

Where this applies

What AI sees when it reads your listing

Your description on Booking.com, Google, your own website. AI recommends businesses whose descriptions answer the questions travellers ask. Click any example to compare.

hotelLakeside cabin
expand_more
restaurantRestaurant
expand_more

Score your description

Write your business description below, then rate it honestly on each criterion. This is the exercise: reading your own words through the lens of what AI needs to recommend you.

group
Audience: does it say who this is for?
spa
Experience: does it describe what it feels like?
pin_drop
Practical detail: location, duration, what is included?
fingerprint
Specificity: are the details concrete and unique?
favorite
Emotional pull: does it make someone want to visit?
10:20

Prompt Engineering Essentials

The quality of what AI produces depends entirely on how you brief it. This is the most important skill to develop. The three examples below use exactly the same tool. The difference is entirely in what the person typed.

Interactive framework

Anatomy of a strong prompt

A strong prompt is not just a question. It is a structured brief with multiple layers. Click each component to see what to include and why it matters. The more of these you provide, the better the output.

The pattern to remember

Role + Context + Data + Task + Format = a prompt that produces something genuinely useful. Most people only provide the task. The businesses that get the most from AI are the ones that invest in the other four components.

You now have the foundations. Time to see what becomes possible.Continue to Part 2 arrow_forward