Visual R&D / Human-led AI

Generation is one step. Direction is the system.

I use generative tools to accelerate visual development while keeping authorship, cultural judgement, continuity and production accountability with the creative team.

AI-assisted concept-development frame of a cultural wellness landscape at duskAI-assisted concept-development image, art-directed by Michelle Tasic

Method before imagery

Human judgement at every decision.

The value is not the volume of generated images. It is the ability to build a visual grammar, recognise what serves the brief, maintain continuity and give stakeholders material they can use.

AI Creative recruiting profile

Directed methodology

From brief to production conversation

  1. 01

    Brief

    Define the audience, decision, cultural context and production question the work must answer.

  2. 02

    Reference architecture

    Separate source references, spatial cues, material language, cultural research and non-negotiable constraints.

  3. 03

    Visual grammar

    Set camera logic, scale, atmosphere, palette, movement and continuity before generating volume.

  4. 04

    Controlled generation

    Use human-authored prompts and image direction to explore within a bounded concept territory.

  5. 05

    Critical selection

    Reject novelty without narrative purpose; select for legibility, cultural judgement and production relevance.

  6. 06

    Iteration and consistency

    Correct composition, architecture, people, material behaviour and sequence continuity across frames.

  7. 07

    Refinement

    Composite, retouch and annotate keyframes so stakeholders can evaluate the intended idea rather than generation artifacts.

  8. 08

    Executive presentation

    Sequence the material around decisions, status and open production questions.

  9. 09

    Production conversation

    Translate visual intent into a useful discussion with spatial, technical, lighting, motion and delivery teams.

Evidence boundaries

What these images are - and are not.

The displayed HQ Worldwide Shows material is AI-assisted concept development art-directed by Michelle Tasic. It demonstrates visual-development thinking and sequence control. It is not presented as photography of approved or realised production.

Process files, prompt logs and confidential client source material are not published as substitutes for cleared evidence. The case studies show only material and role context that can be stated responsibly.

Reference and integrity policy

Human accountability extends to the inputs.