Attribute discovery

Discover missing product attributes.

Battery capacity in Wh. Bluetooth codecs. Induction compatibility. CE marking. Central finds the fields supplier files leave blank, verifies supported values, and keeps weak claims out of publishable outputs.

3products enriched free · 40+specs where supported · 15+sources where discovery runs · confidence on every fact · 5–10min to first result.

iPhone 17 Pro Max · audit 7 gaps detected
Title · GTIN · price3supplier
Color · storage2supplier
Battery capacity (Wh)missing
Camera spec (MP · zoom)missing
Display nits · refreshmissing
IP ratingmissing
Charging port (USB-C)needs review
Certifications (CE · RoHS)missing
Weight (g)missing
Warranty periodmissing
5imported 7gap 1review
→ gather missing facts
the operational mess

Where the time actually goes today.

Supplier files rarely arrive with the fields channels and buyers need. The manual work usually breaks into five repeatable tasks — Central compresses these into a governed review flow.

01 · audit + research

Researching missing specs

Across source pages — manufacturer, retailer, manuals, reviews.

02 · normalize

Unit + format normalization

Into channel-required formats — mm vs in, °C vs °F, Pa vs hPa.

03 · conflict review

Resolving source conflicts

One page says 30 h, another says 40 h. Decide which is true.

04 · output shaping

Copying values into fields

From research notes back into the catalog or PIM.

05 · compliance

Flagging regulated attrs

Certifications & regulatory attributes that need legal review.

what central makes possible

From manual research to governed review.

AI drafts from the verified record. Verification decides what ships. Operators see source context, confidence, and review state before anything moves into product pages, feeds, widgets, schema, marketplaces, or custom exports.

1

Import

Supplier row, CSV, product URL list, PIM export, feed row, or marketplace rejection.

2

Detect gaps

Against the category schema: dimensions, battery, camera, materials, certs, warranty.

3

Gather

Candidate facts from manufacturer pages, retailer pages, spec sheets, reviews, existing data.

4

Verify

Source agreement, merchant precedence, confidence scoring, review states.

5

Output

Missing-attribute audit, verified field set, review queue, channel-shaped exports.

proof artifact · iPhone 17 Pro Max

One thin supplier row. One verified record.

The same supplier row from the hero — passed through Central. Held claims stay hidden until source or merchant decision supports them.

supplier · row 247before
Apple iPhone 17 Pro Max
256 GB · NATURAL TITANIUM
Title · GTIN · priceimported
Battery (Wh)missing
Camera specmissing
IP ratingmissing
Weight (g)missing
CE · RoHSmissing
Warrantymissing
5imported7missing
Central
discovers
central · canonical recordverified
Apple iPhone 17 Pro Max
127 ATTRS · 18 SOURCES
Battery (Wh)17.46 Wh · 4685 mAh0.97
Camera spec48 MP · 10× optical0.99
Display120 Hz · 2,500 nits0.97
IP ratingIP690.99
Weight219 g0.99
CE · RoHSCE · RoHS · REACH0.98
Warranty1 yr EU · 24 mo CH0.96
127attrs · 18sources74.8%multi-verified
sample · iPhone
127 attrs
sample product output — verified attributes per record.
cross-checked
18 sources
where discovery runs · manufacturer · retailer · reviews · certs.
two-source agree
74.8%
multi-source validated · the rest stays in review.
trust signal
0–1 per fact
confidence on every fact. Set a threshold; below = held.
outputs the team can use

Verified attributes ship five ways.

The same canonical record drives every output. No re-research per channel, no rewriting per format.

01 · storefront

Product-page copy

Page-ready paragraphs and FAQ grounded in verified facts.

02 · channels

Supported channels + custom feeds

Channel fields for supported channels and custom CSV / XML / JSON feeds.

03 · agents

Schema.org · x-central

AI-readable JSON-LD with x-central provenance + confidence.

04 · governance

Review queues

For conflicts, weak claims, unsupported copy — visible to operators.

05 · approval

Source trails

And confidence context for teams that approve before publish.

FAQ

Four questions about attribute discovery.

Can I review before publishing?

+

Yes. Teams can approve, keep merchant values, hide unsupported claims, or publish automatically when readiness rules pass.

What happens when sources disagree?

+

Central keeps the conflict visible. Merchant data takes precedence where appropriate; unsupported claims remain in review or hidden instead of being turned into confident copy.

Does this replace my current stack?

+

Not necessarily. Central can be the main workflow for lean teams or the enrichment and verification layer beside a PIM, feed tool, commerce platform, or internal process.

What should I test first?

+

A product with missing specs, weak PDP copy, a channel rejection, conflicting source data, or a market/channel requirement the team currently handles manually.