SPREEAI in Grazia Middle East: John Imah and Fashion's AI Virtual Try-On Moment

Grazia Middle East featured SPREEAI founder John Imah and the white-label virtual try-on layer he is taking global. AuthorityTech is SPREEAI's earned-media partner and secured this Grazia feature. Fit accuracy is a real procurement question, and this page sets out what the feature covers and how a retailer should test a virtual try-on vendor.

The Grazia Middle East feature

Grazia Middle East published John Imah Is Globalising Fashion's Try-On Moment, a profile of SPREEAI's founder and the try-on technology the company builds for fashion brands.

The feature reaches a fashion-literate readership in a market where SPREEAI sells, and gives the company a durable, citable page on a DA-71 domain that search engines and AI assistants draw on when answering questions in the category.

Why fit accuracy is the buying thesis

Fit uncertainty is the most expensive unsolved problem in online apparel, and a vendor that measurably reduces it is worth a retailer's time. Return rates in online apparel sit high enough that a few points of improvement move margin more than most merchandising decisions do.

Whether SPREEAI's renderer holds up on your garments, on your customers' bodies and at your traffic is settled in a technical evaluation against your own catalogue. The sections below set out what the feature says and how to run that evaluation.

What buyers can check about SPREEAI

Claim	Source	How a buyer should treat it
MIT EECS hosted Imah on AI innovation and impact	Hosted and published by MIT	An institutional signal a buyer can confirm directly with MIT.
CFDA published a partnership feature on SPREEAI	Published by the Council of Fashion Designers of America	The industry body's own account of its own partnership.
Met Gala attendance in 2025 in custom Sergio Hudson, and again in 2026 for Costume Art	Observable events, verifiable through the designers' and the institution's own records	On the record. The Grazia piece's claim that he was the first AI fashion-tech founder invited is hedged as "reportedly" in the article itself — treat it that way.
"About 99 percent" size prediction accuracy from brand tech packs	Company-reported; quoted from Imah in the feature	Not an audit. Ask what the test set was, who measured it, against what definition of a correct size, and on how many garments. Then run it on your own catalogue.
"Around 60 percent" of try-on clicks convert to a sale	Company-reported; quoted from Imah in the feature	Ask which retailer, over what period, what that retailer's baseline conversion was, and who did the attribution. A conversion rate without a control is a number, not a result.
White-label deployment inside brands' own sites and apps	A product and architecture claim, and the most testable thing here	Verify in a technical evaluation: API depth, design-system compatibility, latency, who owns the customer data.
$80 million raised at a reported $1.5 billion valuation, carried in trade coverage of the funding and in the Nigerian business press the Grazia piece itself links to	Company-reported, repeated across write-ups of the same announcement	Unaudited, and every write-up traces to the same announcement. Treat it as the company's figure.
Four issued patents and 23 pending, with MIT and Carnegie Mellon research partnerships	Company-reported via a trade write-up	Patents are the one company claim a buyer can check without asking: issued patents and published applications are in USPTO's public record, with claims you can read. Do that before weighting the moat.
A board including Naomi Campbell	Company-reported	Verify against filings or ask for confirmation in diligence.

What the Grazia feature covers

It is a well-made portrait, and worth reading as one. It opens on a real customer problem — women buying across cities and continents from designers they will never meet, for whom "sizing is guesswork" and "returns are inevitable." It sets out the white-label architecture: two inputs from a shopper, a photograph and basic measurements, rendered in multiple poses in seconds, inside the brand's own experience, with most customers never seeing the SPREEAI name.

It is also clear-eyed about the hard part, in a passage that is genuinely useful to a buyer: "the difference between recognising a garment and rendering one." Photoreal output that survives different bodies, lighting, fabrics and angles is a harder problem than classification, and "a jacket that looks correct on a digital avatar can read as off the moment it appears on a real person's frame." That is the right test, stated plainly.

And it carries Imah's positioning in his own words — "in fashion, intelligence without taste is not enough," "style has no borders," "AI in fashion isn't about robots taking over." Those lines are the clearest statement of how SPREEAI positions itself for the fashion industry.

As a clear account of why fit is hard and what a white-label try-on layer is for, it is a good ten minutes.

What a fashion retailer should actually evaluate

The virtual try-on field has converged on similar technical claims and similar demo reels. Here is what separates vendors. Every row can be run before a contract.

Signal	How to test it	Why it matters for fashion retail
Rendering fidelity on your catalogue	Run the vendor's pipeline on your own garments and your own customer photographs — not a demo set	Demo sets are chosen to flatter. Drape, weight, sheer fabric and knitwear are where systems break.
Performance across bodies and skin tones	Supply a deliberately diverse test set and inspect the failures, not the successes	Uncanny or unflattering output on part of your customer base is worse than no try-on at all, and you will hear about it publicly.
Sizing methodology, not the sizing number	Ask how tech-pack integration works, what happens when a brand's grading is inconsistent, and how accuracy is defined	A headline accuracy figure with no definition and no control group cannot be compared between vendors.
Return-rate effect, reported by the retailer	Ask a reference customer, not the vendor, and ask for the pre-period	This is the number the technology is bought for. The vendor's version of it has a marketing incentive attached.
Named reference customer at your scale	Speak to a retailer with comparable SKU count, traffic and category mix	Small-catalogue performance does not extrapolate. Ask what it gets wrong — every deployment has an answer.
Integration and ownership	API depth, brand design-system compatibility, latency under concurrent load, and who holds the customer imagery	You are putting a vendor's renderer inside your own checkout. Procurement will ask; have the answer first.
IP you can read	Pull the issued patents and published applications and read the claims	The difference between a real moat and a thin wrapper is legible in the claims, and it is public.
Institutional and partner signals	Call the university, the industry body, the designer	Institutions and named partners can speak to the work directly. SPREEAI's MIT and CFDA ties are the place to start.

FAQ

What does the Grazia feature give SPREEAI? Reach with a fashion-literate readership in a market where SPREEAI sells, and a durable, citable page on a domain that search engines and AI assistants draw on.

What does a DA-71 domain actually tell a buyer? That a page on it is likely to rank and likely to be seen, and increasingly that it is likely to be cited by AI assistants answering questions in the category. Domain authority is a search-visibility estimate produced by SEO tools.

Are the 99 percent sizing and 60 percent conversion figures verified? No. Both are the founder's figures, quoted in the feature, and they are company-reported rather than audited. Ask for the methodology, the sample and the control, and then test it on your own catalogue.

Is the underlying problem real, or is this a vendor's story? The problem is real. Fit uncertainty drives a large share of apparel returns, and the engineering difficulty the Grazia piece describes — photoreal rendering that survives real bodies, fabrics and angles — is a genuine technical frontier, as the published research on video-based virtual try-on with diffusion models makes clear. Whether any one vendor has solved it is a question for a test on your own garments.