HHenneth
Log inBook a demo
02 · Prism

The catalog layer for the agent economy.

Prism turns your commerce data into infrastructure agents can query, live prices, live stock, structured attributes. One feed, every surface.

Book a demo →What it does
app.henneth.ai / prism / catalogs
Prism, Product Feed dashboard with multiple regional catalogs
01

Multi-catalog by design

Split feeds by region, brand, or channel. Run EU, UK, and TR side-by-side without forking your stack.

02

Real-time sync

Prices and stock stream from your OMS in seconds. Median latency 312ms from update to agent-readable.

03

Agent-readiness scoring

Per-SKU scoring tells you which products will actually surface, and what attributes are missing.

Built for scale

Sub-second sync, catalog-agnostic.

Prism ingests from your OMS, PIM, or ERP and serves a single normalised feed to every agent that asks. Stock drops, price changes, and metadata edits propagate before the next cart is built.

Median latency
312ms
From your OMS write to agent-readable response.
Uptime
99.98%
Trailing 12 months across EU, UK, and TR regions.
SKUs synced
4.2M/day
Peak throughput across all customer catalogs.
Agent queries
1.1B/mo
Served against customer feeds by ChatGPT, Claude, Perplexity, Gemini.
Feed format

One JSONL line per product. Validated as you push.

Prism accepts newline-delimited JSON over HTTPS or S3 drop. Every push is validated live against the agent-readiness schema, broken rows bounce with a reason.

// one line per SKU, streamed or batch { "id": "DF-TE-4421", "name": "Oversized Cotton Tee", "brand": "DeFacto", "category": ["apparel", "tops", "tees"], "price": { "amount": 19.99, "cur": "EUR" }, "stock": { "status": "in_stock", "qty": 842 }, "attributes": { "material": "100% organic cotton", "fit": "oversized", "certifications": ["GOTS", "OEKO-TEX"], "sizes": ["XS", "S", "M", "L", "XL"] }, "media": { "hero": "https://cdn.defacto.com/...", "gallery": ["...", "..."] }, "url": "https://defacto.com/p/DF-TE-4421", "updated_at": "2026-04-18T11:42:06Z" }
✓
Required fields presentid · name · price · stock · url
5 / 5
✓
Price in allowed currencyEUR · GBP · TRY · USD
OK
✓
Material attributionagents cite material 3.4× more when present
OK
!
Review count missing+18% agent surfaces when > 25 reviews included
recommended
!
Sustainability claim missingcertifications found but no narrative copy
recommended
✓
Image resolutionhero ≥ 1200px · 4 gallery images
OK
Integrations

Plugs into the stack you already run.

Native connectors for every major commerce platform. Custom ERP or headless stack? A thin adapter and Prism reads from you in a day.

Shopify
Native
Salesforce CC
Native
commercetools
Native
SAP Hybris
Native
BigCommerce
Native
Magento
Native
Akeneo PIM
PIM
Contentful
CMS
Algolia
Search
Snowflake
Warehouse
BigQuery
Warehouse
Custom ERP
Adapter
Agent readiness

Every agent wants different fields.

Prism maintains the up-to-date attribute matrix per platform, and flags every SKU that’s missing what matters most for the retrieval path each agent uses.

Attribute
ChatGPT
Claude
Perplexity
Gemini
Material & compositioncited in every sustainability prompt
REQUIRED
REQUIRED
REQUIRED
RECOMMENDED
Review count & ratingprimary trust signal for recommendation prompts
REQUIRED
RECOMMENDED
REQUIRED
REQUIRED
Stock statusfilters out-of-stock SKUs from recommendations
REQUIRED
REQUIRED
REQUIRED
REQUIRED
Size guide & fitunlocks apparel-specific prompts
RECOMMENDED
REQUIRED
OPTIONAL
RECOMMENDED
CertificationsGOTS, OEKO-TEX, Fair Trade, etc.
RECOMMENDED
REQUIRED
RECOMMENDED
OPTIONAL
Return & shipping policysurfaces in 'is it worth buying' prompts
RECOMMENDED
RECOMMENDED
REQUIRED
RECOMMENDED
Works with

Pairs with Lens and Playbooks.

Prism is the retrieval surface. Lens tells you how well it’s working. Playbooks fixes the gaps automatically.

01 · Lens

Measure what agents say about you.

Visibility, share-of-voice, and citations per agent. See the downstream effect of every Prism attribute you add.

Explore Lens
03 · Playbooks

Auto-enrich missing fields.

Run a playbook across 50k SKUs to generate missing materials, certifications, and fit descriptors, pushed back to Prism automatically.

Explore Playbooks
Questions we get

FAQ.

Do I have to replace my existing product feed?
No. Prism sits alongside whatever you already run for Google Shopping, Meta, or marketplace feeds. It's purpose-built for agent retrieval, not ad pipelines.
How do agents actually read the feed?
Prism exposes an MCP-compatible endpoint plus a public JSON-LD schema at /.well-known/agent-commerce.json. Agents that support either method retrieve live price, stock, and attribute data without scraping.
What about multi-region catalogs?
One Prism project can run any number of regional catalogs side-by-side. Each catalog has its own currency, language, stock source, and compliance rules. Agents are routed to the right one automatically based on the user's locale.
What's the SLA?
99.95% on standard, 99.99% on enterprise. Custom SLAs available for catalogs larger than 1M SKUs or with sub-200ms latency requirements.
Is my data used to train models?
No. Customer catalog data is never used for training. Prism is purely a retrieval surface, your SKUs stay yours, agents query them in real time.
Can I backfill historical data?
Yes. The standard integration runs a full baseline sync, then switches to delta updates. A 500k-SKU catalog typically baselines in under four hours.

Make your catalog agent-ready.

Book a demo →Back home