Timeline: an AI campaign in 10 days, step by step | Mira Studio

An AI campaign typically delivers in one week: D0 brief, D1-2 moodboard, D3-5 production, D6 revisions, D7 delivery. This AI production schedule isn't theoretical—it's the one we apply at Mira Studio for most of our projects, and the one we compressed to 14 days for Jacob & Co in December 2024. But every day counts, and it's often the client who gains or loses 48 hours.

Real-world planning: campaign in 10 days vs traditional production

Stage	Traditional production	AI campaign (10 days)	AI campaign (accelerated, 7 days)
Brief / scoping	D0	D0	D0
Reference research	D1-3	D1	D1 morning
Moodboard / validation	D4-7	D1-2	D1 afternoon
Pre-production / casting	D8-15	—	—
Production (shoot / generation)	D16-17	D3-5	D2-4
Post-production	D18-25	D6	D5
Revisions	D26-30	D6-7	D6
Delivery	D30+	D7-10	D7

These figures correspond to a campaign of 15 to 30 final visuals, e-commerce or social media format. Video productions or multi-format outputs add 2 to 5 days. A traditional production mobilizes team, studio, models, external retouching; an AI production concentrates everything on art direction and iteration.

D0: Brief – 1 to 2 hours that determine everything

The brief must answer five questions: for whom, for what (channel, use case), what visual universe, what constraints (brand guidelines, physical product to integrate), what timeline. We always ask for precise visual references: Pinterest, existing campaigns, films, paintings. The more visual the brief, the less time lost in back-and-forth.

Where the client saves time: send 10 to 15 references and a brand guidelines PDF on D0. Where they lose it: "We'll see as we go" or "Give us three proposals."

D1-D2: Moodboard and direction validation – the strategic lock

We deliver a moodboard of 6 to 12 boards: composition, lighting, palette, typography if needed. This is the only stage where everything can tip. If the client validates within 24 hours, we start production D3. If they ask for three alternative versions, we slip three days.

At Mira Studio, we impose a single round of revisions on the moodboard. Why? Because AI production iterates quickly on the final image, but slowly on direction. Better to lock the intent before generating.

The smart client move: gather all decision-makers for D2 validation, not in cascade over three days. We delivered Jacob & Co in 14 days because every validation came within 12 hours.

D3-D5: Production and first images – 60 to 200 iterations per visual

We generate, we sort, we refine. Each final visual goes through 60 to 200 variations: framing, expression, lighting, product consistency. AI doesn't "output" a campaign in one click: it accelerates exploration, but art direction does the sorting.

Concretely, we deliver a first selection of 20 to 30 raw images at the end of D4 or start of D5. The client picks their favorites, flags what's wrong (product angle, tone, detail). We refine D5 and deliver final images in high resolution.

Where the client saves time: annotate images with visual comments (Figma, Frame.io, or even PowerPoint), not a three-page email. Where they lose it: ask for "a more premium version" without a visual reference.

D6: Post-production and revisions – the last 20%

AI post-production means light retouching (standardize lighting, adjust a reflection, clean a detail), multi-format exports (Instagram square, 16:9 web, 9:16 story), and sometimes compositing if we're integrating a scanned physical product.

We deliver D6 morning, the client reviews during the day, we correct D6 afternoon. One round of revisions, not three. If the moodboard was validated, adjustments are minor.

The classic trap: asking on D6 to "change the mood" or "test another model." That's going back to D3. If doubt remains, better to clear it at moodboard stage.

D7-D10: Delivery and additional assets

Final delivery D7 for validated formats. If the client wants variations (10 additional visuals in the same direction, video formats, animation), we spread over D8-D10. The AI advantage: reproducing a validated direction costs little, in time and budget.

We deliver in high resolution (300 dpi, TIFF or PNG depending on use), with metadata and clear numbering. No "final final v3" folder: a single export, clean, with the nomenclature agreed in the brief.

Where the AI campaign timeline compresses – and where it explodes

An AI campaign timeline rests on three levers: client responsiveness (moodboard validation, image feedback), brief clarity (visual references, product constraints), and project complexity (number of visuals, formats, 3D integration).

We delivered Jacob & Co in 14 days because the brief was sharp, validations quick, and we iterated in parallel on several series. Conversely, a 20-visual campaign can slip to three weeks if each client feedback takes 48 hours.

The AI production schedule isn't magic: it respects the same rules as traditional production (brief, validation, revision), but compresses the material stages (casting, studio, logistics, external retouching). What we gain in agility, we can lose in back-and-forth if direction isn't locked.

What we can't compress

Certain stages are non-negotiable. Initial scoping (D0) determines everything: without a precise visual brief, we fumble. Moodboard validation (D2) locks direction: without it, every image becomes a debate. And post-production (D6) requires a human eye to guarantee consistency, even in AI.

The frequent mistake: believing an AI campaign delivers "in 48 hours." It's possible for 3 visuals in an already proven direction (variation of an existing campaign, for example). But an original 20-visual campaign requires 7 to 10 days, half of which for art direction, not technique.

Who drives the timeline: agency or client?

Both. The agency sets milestones (moodboard D2, first images D5, delivery D7) and guarantees responsiveness in between. The client validates within agreed deadlines and centralizes feedback (one contact person, one comment file). As soon as we're waiting for validation or receiving contradictory feedback, the timeline slips.

At Mira Studio, we share a Notion or Google Sheet from D0: each stage, its deadline, its owner. No "we'll see," no "as soon as possible." An AI production schedule is only fast if it's explicit.

FAQ

Can you really deliver a campaign in 7 days?

Yes, if the brief is clear, validations quick (24 hours per stage), and scope bounded (15 to 30 visuals, one validated direction). We did it for Jacob & Co in 14 days, revisions included. Beyond 30 visuals or with several directions in parallel, we shift to 10 to 14 days.

What's the difference between 7 and 10 days for an AI campaign timeline?

Revision margin and number of formats. In 7 days, we deliver one direction, one main format, one round of revisions. In 10 days, we add variations (social formats, short video, color variants) and a second retouching pass. The creative core remains the same.

Where are the 48 hours usually lost?

Moodboard validation in several waves, non-centralized client feedback (three contacts with three different opinions), or request to change direction after D5. The smart move: gather all decision-makers from D0 and D2, and annotate images rather than describe in text.