Digital Design Days Milan 2026 · ~90 min

AI in Creative Production

Real Briefs. Real Deadlines. Real Solutions.

Gray Grandeza - Co-Founder & CTO, Cuadro Group
Davide Lancini - Software Engineer & Tech Lead, Cuadro Group

Block 1 - Intro

Gray Grandeza

Co-Founder & CTO, Cuadro Group

10 years as a 3D artist. Creative director of a 60-person studio. Projects for Netflix, LEGO, DC.

Now: CTO at Cuadro Group, managing a team of developers, engineers and tech leads.

Block 1 - Intro

Davide Lancini

Software Engineer & Tech Lead, Cuadro Group

AI Developer and Python expert with a teaching background. Specialized in building AI-powered applications, intelligent automation, and backend systems.

Helping teams turn complex technical ideas into reliable products and practical software solutions.

AI Does Not Have to Be Disruptive

There is a more honest conversation for creatives.
It is about one resource we are always running out of.

Time.

Deadlines do not flex. Clients want more for less.
The gap between what a project requires and what a pipeline can deliver is a problem every creative in this room has felt.

AI, at its best, is a way to close that gap.

Today: 4 Real Examples

Module 01

Digital Avatars

Digital double in 2 weeks
→ Full AI pipeline

Module 02

CGI Product Animation

Precise product video in 10 days
→ Hybrid 3D + AI

Module 03

AI 3D + MetaHuman

Custom character in hours
→ AI + Unreal Engine

Module 04

Automation (Davide)

Workflow orchestration
→ N8N

Each one: a different way AI fits into a creative workflow.

Block 2 - Theory & Concept

Every Generation Uses
Every Tool Available

That is not new behavior. The question has always been the same:

What is the right tool for this problem, right now?

The Toolkit Evolves. The Approach Stays.

Then

3ds Max, Maya, ZBrush, Nuke

Now

Weavy, ElevenLabs, HeyGen, Meshy, Seedance, Unreal Engine 5

Assess the problem. Pick the right tool. Execute.
AI is the newest addition. Not a replacement.

"Can AI do this?"

"Where in this workflow does AI save us time without compromising the result?"

That framing is what you will see in every example today.

For Each Module

The Problem → The Solution → The Results

What the brief was, what the constraints were.
What the workflow looked like, which tools, which steps.
What came out, what it means.

Module 01

Digital Avatars

Full AI Pipeline

The Problem

Create a digital double of a real person.
10-15 lipsync videos.
Deadline: 2 weeks.

The Traditional Pipeline for a Digital Double

Step	Specialization	Time
3D Modelling	Modeler	Days to weeks
Rigging	Rigger	Days
Texturing & Shading	Texture artist	Days
Lighting	Lighter	Days
Dynamics	TD	Days
Animation	Animator	Weeks
Rendering	Pipeline	Hours to days
Review cycles	Everyone	Ongoing

Multiple specialists. Massive render times. Weeks to months.

Even Hollywood
Struggles with This

Unlimited budgets. Hundreds of people. Still uncanny.

So what do you do with 2 weeks and a real deliverable?

You Rethink
the Pipeline.

The Solution

The Workflow: 3 Tools

Step	Tool	What It Replaces
Character creation	Weavy	Weeks of modelling + art direction
Voice	ElevenLabs	Recording sessions + audio editing
Animation + Lipsync	HeyGen Avatar IV	Rigging + animation + rendering

Real identity → Stylized character → Cloned voice → Animated output.
Days instead of months.

Step 1: Character Creation

Tool: Weavy

Node-based AI image generation workflow.
Take the person's real identity. Transform into a semi-realistic stylized character.

Step 2: Voice

Tool: ElevenLabs

Clone the client's real voice. One recording session.
Generate the full audio library from written scripts.

Step 3: Animation

Tool: HeyGen Avatar IV

Character + audio in. Lipsync animation out.
Natural movement. In sync. Ready to deliver.

The Results

Output Examples

The Results

Output Examples

credit: runway.ml

Tools change.
The approach does not.

Understand what the pipeline actually requires.
Find where AI steps in without losing quality where it matters.
Deliver.

Module 02

CGI Product Animation

Hybrid 3D + AI Pipeline

The Problem

CGI product animation. High quality.
Specific camera movements.
Deadline: 1-2 weeks.

"Just Prompt It"
Does Not Work Here

AI video generation is hit-or-miss. Sometimes 20 iterations for one usable shot.

Fine when you have time. Not fine when the client has approved a specific look and you have 10 days.

Control.

Full AI prompt

3D + AI workflow

The Solution

The Workflow: Precision + AI

Step	Tool	Purpose
1. Animate	Blender	Precise keyframes, approved camera movement
2. Playblast	Blender	Fast unrendered preview (3D animatic)
3. Key frames	Nano Banana 2	AI-rendered look from first/last frames
4. Final render	Seedance 2.0 (via Weavy)	Motion from animatic + look from key frames

The motion is locked. The look is generated. Combined.

Step 1-2: Animate + Playblast

Tool: Blender

Animate the product. Lock the camera movement. Export a playblast.
Think of it as a 3D animatic.

Step 3: AI-Rendered Key Frames

Tool: Nano Banana 2

Take the first and last frame from the animatic. Generate fully rendered key frames.

In minutes. Not days.

Step 4: Final Rendered Output

Tool: Seedance 2.0 (via Weavy)

Two inputs: Animatic → motion reference. Key frames → visual reference.
Final output.

The Results

Output Examples

credit: instagram.com/amirr_motion

AI is not always the full solution.
Sometimes it is one precise layer
in a larger system.

Knowing where to place it is the skill.

Module 03

AI 3D + MetaHuman
+ LiveLink

From Prompt to Real-Time Performance

The Problem

Can we create a fully custom, animatable digital character - one that moves and performs in real time - without the traditional pipeline?

The Traditional Pipeline

Step	What It Takes
Model the character	Weeks
Rig it	Days
Paint weights	Days
Build facial blend shapes	Days
Set up capture or keyframe	Days

Total: weeks to months.
What if you could compress that to hours?

The Solution

The Workflow: 3 Steps

Step	Tool	Output
1. Generate 3D model	Meshy 5	PBR textured, game-ready mesh
2. Convert to MetaHuman	Unreal Engine 5 plugin	Fully rigged, animatable character
3. Perform	Live Link Face (iPhone)	Real-time facial animation

Text prompt → 3D model → Rigged character → Live performance.

Step 1: AI 3D Generation

Tool: Meshy 5

Describe what you want. Out comes a 3D mesh with proper textures. PBR maps. Game-ready.
Not a sketch. An actual usable asset.

Step 2: Mesh to MetaHuman

Tool: Unreal Engine 5

The plugin conforms it to the MetaHuman rig. Fully rigged, fully animatable.
No manual rigging. No weight painting.

Step 3: Live Link Face

Your iPhone just became a motion capture studio.

iPhone 12+. TrueDepth camera + ARKit. Real-time facial tracking into Unreal Engine.

The Results

credit: youtube.com/@xanaduBlu

Where Is This Going?

Digital Human Market

	2025	2030
Market size	$47B	$258B
Annual growth		~40%

$47B

$66B

$92B

$129B

$181B

$258B

202520262027202820292030

Who Is Using This?

Brand & Marketing

Virtual spokespersons. Any language, any market. No talent fee, no scheduling conflict.

Virtual Influencers

Already a real business category with real revenue. Now within reach of any small studio.

Virtual Production

What major studios do with multimillion dollar stages. Now: Unreal Engine + an iPhone.

Education & Training

Interactive virtual instructors. Healthcare, corporate, anywhere live trainer hours are expensive.

The cost of a performing
digital character is
approaching zero.

That has real implications for what a studio of any size can now offer.

Block 6

If You Have Ever Used a Node Editor, You Already Know How This Works.

Blender's Geometry Nodes. Maya's Hypershade. Unreal's Blueprints.

Inputs flow in. Nodes process them. Outputs go to the next stage.

Same Structure. Same Logic.

The tool changes. The abstraction does not.

Same Structure. Same Logic.

3D Node Graph

→Inputs flow in

→Nodes transform data

→Outputs pass downstream

→You chain logic

AI Workflow

→Inputs flow in

→Nodes transform data

→Outputs pass downstream

→You chain logic

Creatives Underestimate Themselves

The lateral thinking. The critical thinking. The ability to break a problem into connected steps.

That is already in you.

Now: Davide Shows You the Same Thinking Applied to Automation.

Same nodes. Same logic. Different domain.

Tool Reference - The Complete Stack

Module 01 - Digital Avatars

Weavy	Node-based AI image generation	weavy.ai
ElevenLabs	Voice cloning + generation	elevenlabs.io
HeyGen	Avatar animation + lipsync	heygen.com

Module 02 - CGI Product Animation

Blender	3D animation + playblast	blender.org
Nano Banana 2	AI-rendered key frames
Seedance 2.0	Motion + visual → final render	(via Weavy)

Module 03 - AI 3D + MetaHuman

Meshy 5	AI 3D model generation	meshy.ai
Unreal Engine 5	MetaHuman conversion + runtime	unrealengine.com
Live Link Face	iPhone facial mocap	Free (Epic Games)

Module 04 - Automation (Davide)

N8N

Workflow automation

n8n.io

Digital Design Days Milan 2026

Gray Grandeza - Co-Founder & CTO
Davide Lancini - Software Engineer & Tech Lead

Cuadro Group
cuadrogroup.com

AI in Creative Production

Gray Grandeza

Davide Lancini

AI Does Not Have to Be Disruptive

Time.

Digital Avatars

CGI Product Animation

AI 3D + MetaHuman

Automation (Davide)

Every Generation UsesEvery Tool Available

Then

Now

"Can AI do this?"

"Where in this workflow does AI save us time without compromising the result?"

The Problem → The Solution → The Results

Digital Avatars

Create a digital double of a real person.10-15 lipsync videos.Deadline: 2 weeks.

Even HollywoodStruggles with This

You Rethinkthe Pipeline.

Tool: Weavy

Tool: ElevenLabs

Tool: HeyGen Avatar IV

Output Examples

Output Examples

Tools change.The approach does not.

CGI Product Animation

CGI product animation. High quality.Specific camera movements.Deadline: 1-2 weeks.

"Just Prompt It"Does Not Work Here

Control.

Tool: Blender

Tool: Nano Banana 2

Tool: Seedance 2.0 (via Weavy)

Output Examples

AI is not always the full solution.Sometimes it is one precise layerin a larger system.

AI 3D + MetaHuman+ LiveLink

Can we create a fully custom, animatable digital character - one that moves and performs in real time - without the traditional pipeline?

Tool: Meshy 5

Tool: Unreal Engine 5

Your iPhone just became a motion capture studio.

Digital Human Market

Brand & Marketing

Virtual Influencers

Virtual Production

Education & Training

The cost of a performingdigital character isapproaching zero.

If You Have Ever Used a Node Editor, You Already Know How This Works.

3D Node Graph

AI Workflow

Creatives Underestimate Themselves

Now: Davide Shows You the Same Thinking Applied to Automation.

Module 01 - Digital Avatars

Module 02 - CGI Product Animation

Module 03 - AI 3D + MetaHuman

Module 04 - Automation (Davide)

Gray Grandeza - Co-Founder & CTODavide Lancini - Software Engineer & Tech Lead

Every Generation Uses
Every Tool Available

Create a digital double of a real person.
10-15 lipsync videos.
Deadline: 2 weeks.

Even Hollywood
Struggles with This

You Rethink
the Pipeline.

Tools change.
The approach does not.

CGI product animation. High quality.
Specific camera movements.
Deadline: 1-2 weeks.

"Just Prompt It"
Does Not Work Here

AI is not always the full solution.
Sometimes it is one precise layer
in a larger system.

AI 3D + MetaHuman
+ LiveLink

The cost of a performing
digital character is
approaching zero.

Gray Grandeza - Co-Founder & CTO
Davide Lancini - Software Engineer & Tech Lead