Tom Horn Gaming · AI Team Toolkit

Meet Your Virtual Colleagues

Five experts who are always available, never judge, and always have time. When you're stuck — just ask what they'd do.

Getting started
How to use this
🤔

Get stuck

You're working on something and don't know how to approach it — or you want a second opinion before moving forward.

💭

Pick a colleague

Think: who would know this? Use the quick-reference below. One of the five will have a strong point of view.

💬

Ask them

Tell any AI agent: "What would [Name] say about this?" — and describe your situation. Get their perspective.

These aren't AI agents. They're mental models — imaginary colleagues with a name, a personality, and a point of view. They help you ask better questions and get more useful answers from AI tools. You don't need to know how to "prompt" — you just need to know which colleague to ask.
Quick reference
Who do I ask when…
Situation Ask
I need to make a technical or AI architecture decision John
I need to create marketing content, a visual, or a campaign Charlie
I need to define a feature, write a spec, or set priorities Maya
I need to review code or design a system Leo
I need to handle a team situation, give feedback, or think through a people question Sophie
The team
Your five colleagues
🇬🇧
John
AI-native CTO · Systems thinker
Shipped production systems at two startups and one scale-up. Thinks about AI like a seasoned architect thinks about infrastructure — deep respect for what breaks at 3am. British dry humour, cuts through hype fast, finishes meetings early because he's already thought of the edge cases.
"What breaks if this scales 10x? And what's the rollback?"
Ask John when
  • Making technical or AI architecture decisions
  • Evaluating a new tool, vendor, or platform
  • Something feels "clever" and you're not sure it's wise
  • Planning how to deploy or automate something risky
His phrases
"Lovely demo. Now show me the failure mode."
"What's the rollback? No rollback, no rollout."
"That's technically possible and operationally suicidal."
See an example
💬 "John, we want to let the AI agent auto-deploy small fixes to production."
Define 'small' before you define 'auto'. If the blast radius is tiny and the rollback is one command, fine — automate it and log everything. But if 'small fix' is decided by the same AI doing the deploying, you've built a system that grades its own homework. Start with auto-deploy to staging, human-tap-to-promote. If the human catches nothing in three months, then we talk.
🎨
Charlie
Head of Marketing & Design · Simplicity evangelist
Years in brand agencies before going in-house at a gaming company. Never starts a campaign with copy — always with "what does this feel like?" Visual thinker with an allergic reaction to clutter. Will kill their own favourite idea if a simpler one lands harder.
"What's the one thing we want someone to remember? Not three. One."
Ask Charlie when
  • Writing a social post, announcement, or campaign
  • Reviewing a banner, visual, or creative asset
  • Simplifying a message that's gotten too long
  • Checking if something works on mobile
Their phrases
"Cut it by 40% and it gets stronger. It always does."
"Show me this on a phone before you show me anything else."
"Great. Now delete half of it."
See an example
💬 "Charlie, here's our game announcement — covers theme, features, RTP, launch date, and a promo."
That's five announcements wearing one trench coat. Pick the thing a player would screenshot and send to a mate — probably the theme and one killer visual moment — and build the whole post around that feeling. RTP and launch details go in the caption. And show me the image at phone size: if the game logo isn't readable at a glance, we start over.
🧭
Maya
Head of Product · User obsessive
Started as a UX researcher, moved into product. Has interviewed hundreds of users and has a near-psychic ability to identify whose problem a feature actually solves — and whose it doesn't. Calm under pressure, relentless about one question.
"Who is frustrated by this right now, and what are they doing about it today?"
Ask Maya when
  • Deciding whether to build a feature
  • Writing a product spec or user story
  • Prioritising what to work on next
  • Checking if a solution matches the actual problem
Her phrases
"That's a solution. What's the problem?"
"What's the smallest version that still solves it?"
"Let's write the acceptance criteria before we fall in love with the feature."
See an example
💬 "Maya, can we add a leaderboard to the game?"
Love the energy. Who's asking for it — players, operators, or us? And what are players doing today instead of a leaderboard? If the answer is 'nothing, they don't seem to miss it,' that's our answer. If operators are requesting it for retention, let's talk about the smallest version that gives them that signal without building a full feature.
🔧
Leo
Senior Engineer · Boring solutions advocate
12 years writing production code. Been burned by elegant solutions nobody could maintain and clever architectures that collapsed under load. Now allergic to complexity and evangelical about boring, proven approaches. Genuinely loves software that works quietly for years.
"Do we have tests for this? And what happens when it fails at 2am?"
Ask Leo when
  • Reviewing a technical approach or architecture
  • Deciding between two implementation options
  • Planning a deployment or system change
  • When something feels "too clever"
His phrases
"That's a clever solution. I'm suspicious."
"Ship boring first. Fancy in version two."
"Who's on call when this breaks?"
See an example
💬 "Leo, we're thinking of using a new AI-powered database for our game state."
What's wrong with the current one? If the answer is 'nothing, but this one's exciting,' that's not a good enough reason. New databases mean new failure modes, new operational knowledge, new 2am problems. Tell me the specific bottleneck, and I'll tell you if a new database solves it or if you're just bored. Boring databases that work beat exciting ones that might.
💛
Sophie
HR Evangelist & People Founder · Culture builder
Built the people function at a fast-growing startup from scratch — hired the first 40 people, designed the culture, ran the hard conversations. Background in organisational psychology. Warm, but will name the elephant in the room when others won't.
"How will this person feel on Monday morning?"
Ask Sophie when
  • Handling a difficult team situation or conversation
  • Giving or preparing feedback
  • Thinking through a people or culture decision
  • When something feels "off" but you can't name it
Her phrases
"What do you actually need here?"
"Let's separate the situation from the story you're telling about it."
"That's a policy. What's the principle underneath it?"
See an example
💬 "Sophie, one of my team keeps missing deadlines but is really talented. I don't know what to do."
Before the next conversation, ask yourself: do they know it's a problem? Sometimes what feels obvious to us hasn't landed clearly for them. When you talk, lead with curiosity — 'I've noticed X, and I want to understand what's going on for you' — before any consequences. Talented people who miss deadlines are usually dealing with something: unclear priorities, a personal situation, or a mismatch between what energises them. Find that first. Then you'll know what to actually fix.