Sarah Clarity Studios

Chapter 3: Behavior Trees & Telegraph Coverage Plans

Created by Sarah Choi (prompt writer using ChatGPT)

Behavior Trees & Telegraph Coverage Plans for Mecha

A mecha design is not just a shape—it is a set of behaviors the player learns to read. That reading happens through “telegraphs”: the visual, audio, and motion cues that announce what the mech is about to do, what it is currently doing, and what just happened. Behavior trees (and related AI logic systems) decide when the mech chooses those actions. A telegraph coverage plan ensures that every action the behavior system can trigger has clear, consistent cues that match the mech’s metrics and fantasy.

This article is written equally for concept artists on the concepting side and on the production side. If you’re concepting-side, you can design better silhouettes, mechanisms, and surface language when you understand the AI’s action set. If you’re production-side, you can package and coordinate telegraph assets—animations, VFX hooks, audio events, and readable materials—so the game is fair, legible, and satisfying.

What behavior trees are (in plain concept-art language)

A behavior tree is a structured decision map that helps AI choose actions. It’s typically built from conditions (“Is the player in range?” “Is the mech damaged?” “Is the weapon cooled down?”) and actions (“Aim,” “Shoot,” “Charge,” “Retreat,” “Call backup”). It organizes priorities, fallbacks, and sequences.

You don’t need to code to benefit from this. What matters for concept art is that behavior trees define the “action vocabulary” a mech will perform repeatedly. If an enemy mech can do eight attacks and three movement modes, those behaviors will show up hundreds of times in play. Your design needs to support that repetition with clear reads and durable cues.

What a telegraph coverage plan is

A telegraph coverage plan is a practical checklist that answers: “For every AI action and state, what does the player see and hear?” It connects behavior logic to a coverage set of cues.

A good plan covers three phases of an action: anticipation (warning), execution (the active window), and recovery (cooldown or vulnerability). It also covers state telegraphs: searching, alerted, engaged, low health, overheated, stunned, retreating, and so on.

For concept artists, this plan becomes a design tool. It tells you where you need moving parts, glow seams, venting, recoil mechanisms, audio-emitting components, or silhouette changes that the player can recognize at gameplay camera distance.

Why mecha telegraphs are special

Mecha telegraphs have a mechanical advantage: you can show intent through hardware. Armor plates can open. Vents can flare. Thrusters can spool. Capacitors can glow. Weapons can charge. Stabilizers can deploy. The player reads these as “machine behavior,” which feels grounded and satisfying.

The risk is that mecha can become noisy. If every part moves and every light blinks, the player won’t know what matters. Telegraph design is about hierarchy: big cues for big threats, small cues for minor actions, and consistent mapping so the player learns.

Metrics → behavior: how numbers become readable decisions

AI decisions are often based on metrics: distance to target, line of sight, angle, cooldown timers, heat, ammo, stagger thresholds, and navigation constraints. That means your telegraphs should reflect those metrics.

If the mech chooses a close-range grapple when within a certain distance, you need a close-range “tell” that reads as a lunge setup: posture drop, arm alignment, claw opening, tether spool. If the mech switches to ranged suppression beyond a distance, you need a ranged “tell”: shoulder brace, barrel tracking, muzzle charge.

A strong concept package makes these relationships visible. It helps design, AI, animation, VFX, and audio stay aligned, because the mech’s body language matches its decision logic.

Start with the action inventory: what the behavior tree can do

Before you can plan telegraphs, you need a clear list of actions and states. This is usually available from design and AI teams: move sets, weapon modes, special abilities, and state changes.

Concepting-side artists can ask for a high-level action list early. You don’t need the full tree; you need the verbs: “patrol,” “scan,” “aim,” “burst fire,” “rocket salvo,” “dash,” “stomp AOE,” “deploy shield,” “call drones,” “overheat vent,” “retreat.” Once you know the verbs, you can design the mech to support them.

Production-side concept artists can help formalize this into a telegraph coverage table for art: action name, trigger condition, anticipation cue, execution cue, recovery cue, and associated VFX/audio hooks.

Designing telegraphs: the three-channel model

Telegraphs usually land through three channels: silhouette/motion, materials/light, and sound.

Silhouette and motion is the strongest at gameplay distance. A posture change, a shoulder lift, a torso twist, a weapon arm raising—these are readable even in low detail.

Materials and light is the next strongest. Glow seams, heat bloom, vent coloration, emissive pulses—these can clarify what is charging or overheating.

Sound is the final reinforcement. Servo whines, hydraulic hiss, capacitor hum, warning beeps, clacks at hard stops—these make the cues feel physical and also help players who are not staring directly at the mech.

A good coverage plan assigns primary and secondary channels per action. For a lethal attack, silhouette/motion should be primary. For a subtle state change like low ammo, materials/light and sound can carry it.

Telegraph hierarchy: what deserves the biggest read

Not every action should be equally loud. If everything has a huge wind-up, combat becomes slow and cluttered. Prioritize telegraphs based on threat and novelty.

High-threat, high-damage actions should have big anticipation cues and clear execution cues. Examples include grapples, charged beams, stomp AOEs, missile salvos, and rush attacks.

Moderate-threat actions can have smaller cues: a quick barrel alignment for burst fire, a short shoulder twitch for a micro-dash.

Low-threat or ambient actions can be mostly audio and subtle motion: scanning head movements, idle venting, minor posture shifts.

This hierarchy should connect to metrics. If an attack has a long range and high damage, the warning needs to be clear at that range.

Behavior states: making AI “thought” visible

Players read not just attacks, but intent. The AI state machine—searching, alert, engaged, retreating—needs telegraphs too.

Searching can be a slow head sweep, sensor ping light, antenna twitch, or scanning beam. Alerted can be a sharper posture change, a quick target lock animation, or a rising warning tone. Engaged can be a stable combat stance with weapon tracking and active vents. Retreating can include backward thruster bursts, smoke screens, or shield deployment.

These state cues are where concept art can shine. You can design sensor arrays, scan emitters, and posture silhouettes that communicate “attention” like a creature’s ears and eyes.

Rigging and physics constraints: telegraphs must be buildable

Telegraphs often require moving parts. That means they must be riggable and must survive deformation checks. If a shield generator opens in anticipation, you need clearance for that door. If a cannon recoils, you need travel space. If a thruster spools, you need vent shutters or glow seams that can animate.

Concepting-side artists can help by designing telegraph mechanisms as simple, robust motions—big plates sliding, clear pivots, obvious shutters—rather than micro-machinery that is hard to rig.

Production-side concept artists can provide rig notes specifically for telegraphs: which parts move for which actions, how far they move, and which movements are “must-have” for readability.

Physics is also a partner. If you want cable whip or cloth flutter as a telegraph, decide whether it is simulated or driven. Simulation can be noisy; driven animation is clearer but costs animator time. Your plan should choose intentionally.

VFX coverage: where effects live and how they scale

VFX is often the clearest telegraph channel in chaotic scenes. A coverage plan should define VFX hooks for anticipation and execution.

For example, a beam weapon might have a charge-up glow in the emitter, arcing electricity along coils, a focusing lens flare, then a beam with a distinct sound and impact effect. A stomp AOE might have dust pre-lift, a foot slam burst, a ground ring shockwave, and debris.

Concept artists can support this by providing surfaces that “host” VFX: vents, seams, coils, capacitors, emitter rings, and impact plates. If the mech has no clear VFX host points, effects feel pasted on.

Also consider scaling: at long range, you may need exaggerated effects or stronger color contrast so the telegraph is readable.

Audio coverage: mechanical truth and timing

Audio telegraphs work best when they are tied to mechanics the player can see. A servo whine that rises as a cannon charges feels believable if the cannon has visible coils or shutters. A hydraulic hiss works when pistons are compressing. A warning beep works when a HUD light or panel is flashing.

A coverage plan should include timing notes: when does the sound start (anticipation), peak (execution), and end (recovery). This helps audio designers sync with animation events.

Concept artists can help by giving audio “sources”: exposed actuators, turbine intakes, vent arrays, and mechanical locks. These are also opportunities for small design flourishes that improve believability.

The telegraph coverage sheet: a concept artist’s deliverable

A telegraph coverage sheet is a single page (or a small set of pages) that maps the action inventory to cues. It does not have to be a spreadsheet; it can be a designed sheet with thumbnails and callouts.

A useful layout includes: an action list grouped by range and role; three-phase cue thumbnails (anticipation/execution/recovery) for the key actions; and callouts indicating which parts move, where VFX spawns, and what audio cue family is expected.

For concepting-side work, you can keep it coarse: “beam charge glow + shutter open + rising hum.” For production-side work, you can be more specific: “coil emissive ramp,” “vent shutters open 30%,” “servo whine starts at 0.3s,” “impact spark burst.”

This deliverable is extremely valuable because it becomes a shared reference across disciplines. It also protects consistency across multiple artists and outsourcing.

Example telegraph patterns that map well to mecha

One reliable pattern is the “brace and lock” telegraph for heavy weapons. The mech plants feet wider, drops center of mass, stabilizers deploy, weapon aligns, and you hear locking clacks. This telegraph tells the player: big shot incoming, and the mech is committed.

Another pattern is the “vent and flare” telegraph for heat-based systems. Vents glow, shutters open, heat haze appears, and you hear a hiss. This communicates overheat or power release.

A third pattern is the “sensor ping and track” telegraph for targeting. The head or sensor block pivots, a scanning light sweeps, you hear a ping, and the weapon begins to track. This communicates that the mech has acquired you.

These patterns are design tools. You can embed them into your mech’s hardware so the game’s readability is built into the silhouette.

Common failure modes (and how to avoid them)

A common failure is telegraphs that rely only on small lights or subtle details. At gameplay distance, those cues disappear. Fix this by making silhouette and motion the primary channel for important actions.

Another failure is inconsistent mapping. If a red glow means “beam charge” on one mech and “overheat” on another, players get confused. Fix this with a shared telegraph language across factions and enemies.

A third failure is over-telegraphing. If every action has a long wind-up, combat becomes predictable and slow. Fix this by reserving big cues for big threats and using smaller cues for routine actions.

Finally, a failure is unbuildable telegraphs. If your telegraph requires a complex unfolding mechanism that cannot be rigged or that breaks collision, it will be cut. Fix this by designing telegraph mechanisms as robust, simple motions and providing rig notes.

How concepting-side and production-side artists can collaborate on telegraphs

Concepting-side artists can embed telegraph hardware early: vent shutters, stabilizers, emitter rings, sensor arrays, and bracing silhouettes. They can also propose the cue hierarchy: what needs to be most readable.

Production-side artists can turn that into a coverage plan that survives implementation. They can coordinate with design and AI to confirm the action inventory, with animation and rigging to confirm buildable motion, and with VFX/audio to confirm cue timing and readability.

When both sides work together, telegraphs remain consistent even as the mech evolves.

The core principle: fairness is a visual design problem

Telegraphs are not just “effects.” They are how the game communicates rules. A behavior tree chooses actions; telegraphs teach players how to respond. When those two systems align with the mech’s metrics and motion reality, the game feels fair and satisfying.

If you remember one takeaway, make it this: every behavior that matters needs a readable cue at the camera distance where it matters. Your job as a mecha concept artist is to design a body that can express those cues—through silhouette, motion mechanisms, material light, VFX hooks, and audio sources.

When you build a telegraph coverage plan, you are partnering with design, animation, rigging, physics, AI, VFX, and audio in the most direct way possible: you are designing the language the player reads.