The gap it closes
A forecast becomes useful at the moment somebody has to commit: how many stories, how many options, which colours get lab-dipped, what carries over, what gets developed from scratch. Most trend output stops one step before that, leaving the translation to a design team working from mood.
F-Predict produces the translation as part of the run, with the reasoning attached, so a range decision can be defended in a review rather than justified by taste alone.
Where it lands in the calendar
- Concept and direction
- Set the scope, run the evidence stages, and read the consumer map, street, material, narrative and catwalk direction as the season brief. This is where a season story is chosen rather than assumed.
- Colour card
- Take the generated themes, curate a season palette tray by hand from them, and export with Pantone TPX references for lab dips and supplier briefs.
- Material and trim
- The material stage names emerging textiles and trims for the scope and explains the driver behind each — consumer, trade or brand-led — so sourcing conversations start with a reason.
- Range architecture
- A season skeleton: stories, how many options each should carry, and an explicit split between carry-over and new development, laid out against an effort ladder so development capacity is allocated deliberately.
- Design
- Design intelligence expresses the direction as silhouettes, details, prints, trims and materials; the design board generates visual boards in the correct product framing for the category.
- Review and sign-off
- A shared workspace where signals and images are pinned, submitted for review, revised with visible diffs, compared and logged — so the decision trail survives the season.
Carry-over versus new development
Development capacity is the scarcest resource in a product calendar, and it is usually spent evenly rather than where it earns most. F-Predict separates the two explicitly: directions with strong, corroborated momentum justify new development; directions that are stable or decelerating are better served by carrying over an existing option in a refreshed colour.
The trend acceleration and adoption stage readings are what make that split arguable rather than arbitrary.
What it does not do
Working as a team
- Named, saved scopes so a team runs the same brief consistently instead of re-typing it.
- A cross-scope library of saved analyses that can be reopened in full.
- Pinning of any signal or image into a shared board, with review status.
- Revisions with visible diffs, comparison views, polls and an exportable decision log.
- Seat-based team plans so a whole product team works from one workspace.
Frequently asked
Can F-Predict tell me how many units to buy?
No. It produces direction, option counts and a carry-over versus new-development split, but unit quantities and size curves require your own sales data and a demand planning system.
Can the output go into a tech pack?
Colour output carries Pantone TPX references and exports directly. Material, trim and detail direction is written as product language a designer or tech team can act on.
Does it work for a small brand without a data team?
Yes. A run needs a scope, not a dataset. The evidence is retrieved externally, so no internal data engineering is required to get a first forecast.
Keep reading
What is F-Predict?
F-Predict — formally F-Trend Predict — is an AI fashion forecasting platform. A user defines a market scope (product category, region, gender, target seaso…
How does F-Predict forecast colours?
F-Predict forecasts colour by first collecting the colours a market is actually using — hex values sampled from campaign and street imagery plus the colour…
What is trend adoption?
Trend adoption is the extent to which a trend has spread through a consumer population, described as a position on a curve running from innovators through…
F-Predict vs traditional trend forecasting
Traditional trend forecasting is expert-led: a forecaster synthesises cultural observation into a seasonal direction, working top-down from a macro narrati…