The person decides: bots that help without solving the game for you

A simple question with an uncomfortable answer

If a team game can be played solo because bots cover the missing roles, and that same game is used to prove you learned something, an obvious question shows up: what stops the bot from solving everything while you pass without having done anything?

It's a question we took seriously while designing the Academy line of knowledge games, because the answer determines whether the certificate on the other end means anything at all. The fix isn't a hidden technical restriction: it's an explicit design rule, the same one across all three roles of every game.

The rule: the bot proposes, it never solves

In Firewall Humano — and in the knowledge games joining this line — a bot never executes a decision that a connected teammate could make instead. Instead, it proposes it over the team chat and waits: "I'd report this one: the domain is three days old and doesn't match our usual vendor. Are you going to report it?" If a person holds that station, the final call is always theirs.

That's not a small detail. In Firewall Humano, marking a technical signal, verifying a request through another channel, or deciding what to do with a message are exactly the actions that generate competency evidence. If a bot resolved those for you, there would be nothing left to demonstrate — you'd just be watching a machine play.

When a bot does decide

There's one exception, and it's just as deliberate: if no human is holding the station that has to make that call, the bot does make it. If you're playing solo as Inbox, with bots covering Analyst and Context, those bots inspect, call and mark signals without asking your permission — because there's no person waiting at those stations to decide. But the final call on every message, the one that moves the Security and Operations meters, always stays in your hands if you're playing Inbox.

The full rule is even more precise: a bot only takes over a decision that no human on its team could make. If you're playing Analyst, the bot Inbox does decide with whatever the team investigated, because nobody else can — otherwise the shift would sit stuck, waiting on a decision from a person who doesn't exist. It's the same standard across all three stations: there's never a decision nobody makes, and there's never one made for you that shouldn't have been.

So what do the bots actually do

Not deciding for you doesn't mean staying quiet. Experiencia RPG's cooperative bot layer (documented in detail in our technical bots guide) gives them a repertoire of active behaviors:

A bot can investigate and share what it finds: inspect a technical layer, check the org chart, call a registered number, and report back over chat with the same level of detail an attentive human teammate would give. A bot can explain its reasoning when it proposes something, because in a knowledge game the reasoning matters as much as the conclusion — it's not enough for the bot to say "this is phishing," it has to say why. And a bot can prioritize — say, tackling the message with the least time left first — so the session keeps a reasonable pace even if the human is thinking things through slowly.

What a bot never does in a knowledge game is make the decision that belongs to a human who's present. That line is what separates "a teammate that helps" from "a machine that takes the exam for you."

Three levels, the same rule

Like the rest of the Experiencia RPG catalog, knowledge-game bots come in three levels — Rookie, Skilled and Expert — that change how fast they react and how well they prioritize among several possible actions. A rookie bot investigates slowly and sometimes forgets to share what it found; an expert one anticipates, prioritizes the most urgent message, and explains its reasoning the moment it acts. But at no level, not even Expert, does a bot cross the line into deciding something you could decide. The level changes how good a teammate the bot is, not how much of the game it solves for you.

Why this matters beyond the game

At its core, this rule is the same one that separates an assistive tool from one that replaces a person's judgment. An email filter that automatically blocks everything suspicious also "decides for you" — and that's exactly why, when it fails (and every filter fails eventually), the person on the other end doesn't know how to spot the signal the filter missed. The skill of noticing that something's off, even when every automated check says it's fine, doesn't get trained by letting a machine exercise it in your place.

That's why playing Firewall Humano solo isn't "watching three bots handle a shift": it's taking the station you most want to practice, with two artificial teammates that give you information and an opinion, but never take the message out of your hands. If the final result says you passed a level, it's because you — not a bot — made those calls.