AI That Works From Your Store's Real Data
The AI capabilities in QuantumOS X3 work from your own catalog, orders and customers rather than generic text. Kynetra Search keeps a search index in your own database; Kynetra Prime watches customer search activity and proposes relevance changes, with risky ones held for your approval. You choose which models run by registering your own provider keys, and you set the budgets and guardrail rules.
Jump to capability
Kynetra Prime and the Regents — Autonomous Agents Inside the Platform
Kynetra AI is the layer behind the autonomous parts of QuantumOS X3. Each regent is an agent that watches one area of your business, writes recommendations, and either waits for your approval or applies the low-risk ones within limits you set. Kynetra Prime does this for search relevance. Other regents cover repricing, demand generation and ad accounts, each with its own console in the admin.
Kynetra Search
NewA search index kept in your own database and updated whenever a product is saved, with fuzzy matching, synonyms, redirects and ranking you can tune.
Kynetra Search reads from a flattened product search table in your own database instead of joining raw catalog tables on every query. The database keeps that table current automatically, so saving a product updates its search row in the same step and no scheduled rebuild is needed. Queries are cleaned up, matched with fuzzy text matching, expanded with your synonyms and ranked by weighted signals: exact name matches, prefixes, matches on every word, brand, category, whether the product can be bought, and whether it has an image. In Search Studio, merchandisers pin, boost or bury products, set redirects and zero-result fallbacks, and adjust the ranking weights per business.
- Search table updated automatically on every save, with no scheduled rebuild
- Fuzzy matching, synonyms and redirects
- Pinned, boosted and buried products
- Zero-result fallbacks: choose what to show when a query finds nothing
- Ranking weights you can adjust without a deploy
- Last good results served if the database errors
Kynetra Prime — Search Auto-Training
NewWatches customer search activity and proposes relevance changes. Risky changes wait in a review queue until you approve them.
Kynetra Prime watches search activity such as queries that return nothing, clicks and add-to-carts from search, and proposes improvements: synonyms, redirects, query rewrites, zero-result recoveries and boosts. A discovery run also asks a language model to look at recent search events and suggest more. The engine applies only low-risk changes by itself. Anything riskier appears in a review queue, and it goes live only when an admin approves it. A kill switch and a percentage rollout control whether the feature runs at all.
- Learns from search events, not from a one-off import
- Risky changes need an admin's approval before they go live
- Low-risk changes are applied automatically, so review the training history
- On-demand Train now and Discover suggestions actions
- Quality dashboard with zero-result rate and click-through
- Kill switch and percentage rollout
Meaning-Based Search & Related Products
SearchVector search on the storefront finds relevant products when keywords match nothing, and powers related-product suggestions.
On the storefront, the shopper's words are turned into an embedding with a Cloudflare Workers AI model and matched to the nearest products in a vector index kept separately for each business. This finds relevant products even when no keyword matches. Repeated queries reuse a cached embedding. The lookup runs against a time limit behind a circuit breaker, and if it is slow or unavailable, shoppers get the keyword results instead. The same index powers a more-like-this suggestion for a product. Image search needs a vision and vector service to be configured for the business.
- Finds products by meaning when no keyword matches
- Per-business vector index
- Repeat queries reuse a cached embedding
- More-like-this suggestions from the same index
- Falls back to keyword results if the lookup is unavailable
- Image search when a vision and vector service is configured
Search Resilience
ArchitectureTime limits, circuit breakers and cached fallbacks keep search answering when a dependency is slow or failing.
Slower steps such as meaning-based lookups and AI calls run against a time limit, so one slow dependency cannot hang a search. A circuit breaker stops calling a dependency that has failed repeatedly, then lets one trial request through to see whether it has recovered. Performance budgets are stored per metric, and a budget breach can switch an expensive step off so search degrades gracefully. If the database errors, the last good result for that search is served from cache. Identical queries arriving together are answered from one lookup.
- Deadlines on meaning-based lookups and AI calls
- Circuit breaker with a single trial request to test recovery
- Performance budgets stored per metric, which can switch an expensive step off
- Cached last good results if the database errors
- Identical simultaneous queries share one lookup
- Kill switches for individual search features
Command Chat — the Admin Copilot
CoreAsk questions about orders, catalog and analytics in plain language. It reads live data as you, and cannot change anything yet.
Command Chat is the operations copilot inside the admin. It acts as the signed-in user on their business, so it can only reach what that user's role allows. It answers by calling the admin's own read tools, shows each tool call as it happens, and links the records it mentions. In this release it is read-only: if you ask for a change, it says so and points you to the screen where you can make it.
- Works as the signed-in user
- Calls the admin's own read tools and shows each call
- Read-only in this release
- Tool calls logged with user, session and result status
- Uses your model keys through the same AI gateway as every other AI feature
- Points you to the right screen when you ask for a change
Theme Copilot
CoreDescribe storefront theme changes in plain language and see them applied to the live preview, with undo.
In the theme editor, you describe a change such as a transparent header or bigger, rounder buttons. The copilot sees your current theme, returns a checked set of design-token changes with an explanation, and layers them over the theme in the editor. Because the edit goes through the editor's normal history, you can undo it, preview it and let autosave keep it. It works across several turns, so each request builds on what is already there.
- Reads the current theme, so each request is relative to what is there now
- Design-token changes are checked before they are applied
- Changes flow through undo, redo, preview and autosave
- Works over several turns
- Uses your model keys through the AI gateway
Conversation Engine
ConversationalA flow-based bot engine behind the WhatsApp and Telegram assistants. Language-model intent detection is built but not connected yet.
Conversations run on a session store that remembers where each customer is in a flow. A flow can search the catalog, add items to a cart, create a payment link and flag a conversation for a person to take over. A separate agent that reads a message and returns an intent (browse, add to cart, checkout, payment, complaint, return, negotiate or greeting), the items and quantities mentioned and whether to escalate has been built, but it is not part of the live flow yet, so today's bots follow the flows you design. See the conversational commerce page for the channels that are live.
- Sessions remember the customer's place in the flow
- Flows for catalog search, add to cart and payment links
- A handover flag for a person to take over
- Intent detection agent built, not yet connected
- Used by the WhatsApp and Telegram channels
Fraud Screening
SecurityWeighted rules score each new order and place it in one of four outcomes: allow, review, hold or block.
After an order is created, active fraud rules are checked against it and the weights of the rules that match are added up. A score under 30 is allowed, 30 to 60 goes to review, 60 to 90 is held, and above 90 is blocked; a rule set to block blocks the order on its own. Rules you can use include order velocity, large first orders, brand-new accounts paying with a chosen method, email address patterns, country mismatches and a blocked-payment-methods list. Held and blocked orders are put on hold with the reason, and the admin shows every triggered rule with its weight so a person can decide.
- Each rule has a weight and an action you set
- Four outcomes: allow, review, hold and block
- Held and blocked orders go on hold with a reason
- Every triggered rule is shown with its weight
- Score and triggered rules saved with the order
- Review queue and rule editor in the admin
Segmentation
MarketingBuild customer segments from nested AND/OR rules on lifetime value, order history, tier, points and location.
A segment is a set of rules joined by AND and OR, and groups of rules can be nested inside each other. Rules run against your live customer and order data when you ask for the segment, with no copy to a separate analytics store. Fields include email, name, phone, city, country, sign-up date, order count, lifetime value, last order date, loyalty points and loyalty tier. Segments can be used in marketing journeys and campaigns.
- Nested AND/OR rule groups
- Rules run on your live data
- Count and sample preview before saving
- Used by journeys and campaigns
- No separate analytics store
Marketing Agent
MarketingRecommends coupons, campaigns, segments, journeys, bundles and price changes, with budget and margin guardrails.
The marketing agent looks at your orders, stock and customers and fills an inbox of recommendations. Triggers include ageing inventory, lapsed customers, abandoned carts, festivals, new-customer welcomes, replenishment and weekend flash sales. A recommendation can be a coupon, promotion, campaign, segment, journey, price change, bundle or referral. You can approve, reject, edit or schedule each one. Guardrails cap discounts per day and per month and protect a minimum margin, and the agent can be set to apply recommendations itself within those limits.
- Recommendations in an inbox with a status for each
- Bulk approve and reject
- Daily and monthly discount budgets
- Minimum margin protection
- Optional auto-apply within the limits
- Role permissions on approve and apply
Description Enhancer
CatalogGenerate or improve product descriptions in bulk with Claude, GPT or Gemini, and keep the drafts for review.
Choose products and a model, and the enhancer writes descriptions in the tone you pick, using the product's own details and, if you allow it, the brand's website or Google Shopping snippets. Bulk runs are background jobs with progress and per-item status, so a large batch does not block the page. You can run it in draft mode and review each result before replacing the live description. Each model shows a credit cost in the settings, and requests are recorded in the AI gateway with the model, feature and token use.
- Claude, GPT and Gemini
- Five tone choices
- Draft mode with review
- Bulk runs as background jobs with per-item status
- Optional brand-site and Google Shopping sources
- Requests logged with model, feature and tokens
Your Own Keys & Guardrails
SecurityRegister your own provider keys, set budgets, and apply rules for personal data, risky prompts and expensive calls.
You can add your own provider keys in the admin. They are encrypted at rest with AES-256-GCM and only the last four characters are shown again. Calls use your key when it matches the model's provider. When a model fails, the gateway tries the fallback models you have configured. Guardrail rules run on prompts and responses: block, redact or warn on personal data patterns such as email, phone, card number, Aadhaar and PAN, flag prompt-injection patterns, set a maximum cost per call, allow or deny models, block output patterns and rate-limit calls. Spend budgets, set by scope and period, block calls once they are exceeded, and alerts fire as usage approaches the limit.
- Provider keys encrypted with AES-256-GCM
- Your key is used when it matches the model's provider
- Fallback models
- Rules can apply to chosen features only
- Spend budgets that block, with usage alerts
- A test panel to try a prompt against your rules
Know What Each AI Call Costs
When you register your own provider key, calls to that provider use your key and the provider bills you directly. AI usage on the platform's shared key is billed separately from the subscription, as the pricing page explains.
Spend budgets
Set budgets by scope and period. Once a budget is exceeded the gateway blocks further calls instead of only reporting them, and alerts fire as usage approaches the limit.
A record of every call
The AI gateway records each call with the feature that made it, the model and the token use. You can see which part of the product is spending, not one pooled line.
Rules per feature
Guardrail rules can apply to chosen features only, so a rule for the description enhancer does not have to apply to the admin copilot.
Built In Versus Assembling It Yourself
The alternative to a built-in capability is usually a service you choose, connect and keep in sync. The differences below are about what you have to build and run, not about any other platform's product.
| Dimension | QuantumOS X3 | Assemble it yourself |
|---|---|---|
| Search engine | Search index in your own database, updated on every product save, with Search Studio for synonyms, redirects and ranking | You choose, run and keep in sync a search service such as Algolia, Typesense or Elasticsearch |
| Auto-training | Kynetra Prime proposes synonyms, redirects and boosts from search activity. Risky changes wait for approval. | You build the pipeline that learns from search events and decide how changes are reviewed |
| Grounding | Works from your own catalog, orders and customers | Generic unless you build retrieval over your own data |
| Model choice | Claude, GPT and Gemini through your own keys, with fallback models | Any model you integrate yourself |
| AI cost | On your own key, the provider bills you directly. Usage on the platform's shared key is billed separately from the subscription. | You pay providers directly and carry the build cost |
| Fraud screening | Weighted rules, four outcomes and a review queue | Not included. You integrate a fraud service. |
| Meaning-based search | A vector index for each business, with keyword results as the fallback | You run and sync a vector database |
| Call log | The AI gateway records every call, and Command Chat records each tool call | You implement logging |
| Guardrails | Rules for personal data, prompt injection, cost per call, model lists, output patterns and rate limits | You implement every guardrail yourself |
AI With a Record of What It Did
Every AI call goes through one gateway that records it. The admin copilot acts as the signed-in user and is limited to their role. Guardrail rules can screen prompts for personal data and risky patterns before they are sent.
QuantumOS Dispatch
Insights for commerce operators
Operator stories, platform deep dives and notes on what ships.
We'll send the QuantumOS Dispatch by email. Unsubscribe any time. See our privacy notice.