PRODUCT
Everything Rynex does today — labeled live, beta, or roadmap.
One page, no asterisks: the six capabilities that ship, the GEO Layer that works upstream of your store, the 28-protocol readiness scan, and the short list of things we have not built yet.
AI SALES AGENT
The infrastructure for autonomous sales.
Six core capabilities power Rynex today: cart execution, proactive triggers, decision control, experimentation, developer integration, and operating workflows. GEO extends the same loop upstream.
Rynex is built around product recommendation, cart execution, and revenue attribution.
Teams can inspect confidence levels, filters applied, ranking logic, and alternatives considered, so AI decisions are auditable instead of opaque.
Something warm for a reading corner?
Two picks under $150 — both in stock, ships in 2 days:
Havre Wool Throw
wool · 130×180 · in stock
why this? Matches “warm”, fits your budget filter, 4.8★ from 212 reviews — confidence 0.84
Lumo Floor Lamp
warm LED · 3 heights
illustrative data
the shopper-facing surface
Sell
Direct Cart Actions
Execute cart actions through URL, JavaScript, or API integration so shoppers move from question to checkout without a broken handoff.
- URL, JavaScript, or API integration paths
- The primary action type Rynex uses to close sales autonomously
- Turns answers into checkout movement
Proactive Trigger Engine
Use exit intent, time on page, scroll depth, cart value, idle time, visit count, and UTM campaign signals to intervene at the right moment.
- Exit intent
- Time on page
- Scroll depth
- Cart value
- Idle time
- Visit count
- UTM campaign
Prove
Decision Center + Explainability
Inspect confidence thresholds, ranking logic, alternatives considered, and human handover reasons with full attribution context.
- Confidence thresholds — under 0.65 low, under 0.85 medium
- Ranking logic and filters applied
- Alternatives considered
- Human handover rules with full attribution context
- Handover from low confidence, revenue-risk patterns, repeated no-result responses, or explicit user requests, with SLA timing tracked
Experimentation & Eval Sets
Run A/B experiments, allocate traffic, define evaluation sets, and measure uplift or latency impact before broad rollout.
- A/B experiments with traffic allocation
- Eval set definition
- Uplift measurement
- Latency impact, measured before broad rollout
Operate
Developers & Webhooks
Manage webhook endpoints, test deliveries, and monitor timestamped execution history for systems integration.
- Webhook endpoint management
- Test event deliveries
- Delivery outcome monitoring
- Timestamped execution history
Support Inbox Operations
Track support tickets, threaded replies, attachments, and SLA handover moments when a human should step in.
- Support ticket tracking
- Threaded replies
- Attachments
- A dedicated inbox workflow
- SLA handover moments when a human should step in
- Human takeover of any conversation, with order actions — cancel, refund, address change — run by the teammate who takes it over
- Routing and canned responses built in
the operator view
Dashboard — go-live checklist, revenue, funnel
Analytics — funnel, AI quality, handoff, segments
GEO LAYER
Be visible in AI. Let Rynex take over when they land.
Shoppers now start product discovery inside ChatGPT, Perplexity, and Gemini. GEO Layer prepares the catalog for those answer surfaces, then carries the landing context into Rynex on-site.
Catalog readiness
Full catalogStart from the same full product catalog that powers Rynex recommendations, availability checks, and storefront context.
GEO job runner
Automated refreshGenerate AI-readable product answers, FAQ blocks, and crawl-ready content on weekly or daily refresh cadence.
Structured snippet
No extra installUse the existing Rynex snippet to publish FAQ and JSON-LD markup without adding another site implementation.
AI traffic attribution
Revenue loopConnect AI referrers, landing sessions, shown products, and downstream conversions once the shopper arrives.
Full catalog, automatically
No product caps. The job runner works through the entire catalog, generates GEO-ready output for each product, and refreshes it on a weekly or daily schedule.
Zero setup with agent.js
Rynex is already installed on-site. GEO Layer uses that same snippet to inject FAQ blocks and JSON-LD schema into product pages automatically.
Visibility across AI platforms
ChatGPT, Perplexity, and Gemini can crawl the product page. Structured content improves how often the right SKU can surface in AI recommendations.
Competitor gap monitor
See where rivals are winning AI answers. Track prompt groups where your catalog is missing and prioritize the pages most likely to recover demand before the click.
AI-referrer attribution
Visits from an AI answer surface are tagged after landing, so source, query context, and the first Rynex session signals stay tied together.
Free on every plan: 50 tracked products · 3 competitors · 50 saved prompts · 100 prompt simulations / month with GEO Launch
AI visibility
last crawled 4h ago · 60 prompts per engine
Agent readiness
23 of 28 protocol checks passing.
1 warning
- llms.txtpass
- product JSON-LDpass
- markdown negotiationpass
- agent actions manifestwarn
AI answer share
8w trendChatGPT8 weeks · 52% → 68%
68% of sampled answers · 41 of 60 prompts
Competitor gap
“best linen sheets”Illustrative demo-store data · 1,912 agent answers sampled
AI Visibility analytics and Agent Readiness are free on every plan. See pricing.
AGENT READINESS
Can AI agents actually read your store? We check 28 protocols.
The scanner runs the classic technical checks — robots.txt, sitemap, HTTPS, structured data, AI-crawler access — plus an agent-protocol layer: llms.txt, Content Signals, markdown content negotiation, api-catalog (RFC 9727), Link headers, agent cards, MCP discovery. All of it graded with one honesty rule: a missing optional protocol is a warning, never a failure, and anything we can't verify passively gets no invented status.
No account needed.
Rynex customers get the same scan inside the console under AI Visibility → Agent Readiness, with copy-paste artifacts prefilled for the store and a weekly re-scan so regressions surface themselves.
- llms.txtpass
- robots.txt + AI crawler accesspass
- Sitemap, HTTPS, crawlabilitypass
- Content Signalspass
- Markdown negotiationpass
- llms-full.txtpass
- auth.mdpass
- api-catalog (RFC 9727)pass
- Link headerspass
- A2A agent card / MCP card / OAuth discoverywarn
Our own rynex.io scan, grouped — including the warning we keep on purpose. Read the write-up
Connect in an afternoon
Six live connectors, one catalog.
Most stores live in ~2 hours.
- Shopify
- WooCommerce
- XML/JSON feed
- REST API
- Webhooks
- File import
Source marketplace and catalog health — the same connectors listed above, in the console.
HONEST ROADMAP
What's live, and what is only roadmap.
Live is production-ready now, beta is controlled rollout, and roadmap items are planned scope only. We label each scope explicitly to avoid overclaiming.
Live
Production-ready now.- Direct Cart Actions
- Proactive Trigger Engine
- Decision Center + Explainability
- Experimentation & Eval Sets
- Developers & Webhooks
- Support Inbox Operations
- AI Visibility analytics
- Agent Readiness
- Shopify
- WooCommerce
- XML/JSON feed
- REST API
- Webhooks
- File import
Roadmap
Planned scope only — not built yet.- Visual flow builder
Do we support WhatsApp, Instagram, and a visual flow builder today? Not today. These are roadmap priorities and are labeled as planned scope so teams can plan with realistic rollout expectations.