Agent driven operation of an enterprise ecommerce with more than 15 brands at Tienda Newsan

Tienda Newsan sells the portfolio of more than fifteen official brands, among them Noblex, Philco, Atma, Motorola, Braun and JVC, across tens of thousands of active references.

An ecommerce this size does not fail for lack of tools: it fails because catalog, content, price and advertising live in separate circuits that nobody synchronizes at the same pace.

The operation was fragmented, and fragmentation has a cost that shows up in no report.

• Fifteen different brand manuals: each manufacturer has its own tone, and a sheet written in the wrong voice erodes the brand selling it.
• Tens of thousands of references: technical attributes, specifications and taxonomies that are not standardized by hand.
• Data that contradicts itself: the catalog price, the transactional price and the ad price can say three different things with nothing flagging it.
• Peaks that do not forgive: Hot Sale and CyberMonday are not a monthly average, they are a single day.
• Regulatory and brand risk: an invented specification on an appliance is not a copywriting error.

At this scale manual management is not slow: it is impossible to keep consistent.

We built a composable architecture where each layer does one thing, and a network of specialized agents that runs it end to end under contracts and gates.

The underlying decision is that master data has a single owner and the other layers consume it, instead of maintaining copies that drift apart.

Multi brand catalog ingestion and normalization

• Integration with the central PIM for the more than fifteen official brands in the portfolio.
• Strict taxonomy validation and deterministic mapping of technical attributes and critical product specifications.
• Deduplication and bulk standardization, so master data stays consistent without manual intervention.

Editorial governance on Magnolia

• Magnolia operates as the unified presentation and experience hub, decoupled from the transactional backend.
• Reusable modular components, brand landing templates, contextual banners and dynamic promotions.
• Multisite management with real time responsive preview and content versioning.

High scale transactional integration with Magento

• Continuous bidirectional sync of stock, list and offer prices, commercial rules and cross promotions.
• Payment gateways and optimized checkout.
• Tolerance for massive traffic peaks through microservices and asynchronous queues, which is what separates a Hot Sale from an outage.

The agent network, each with its role and contract

• Writing and enrichment: technical sheets aligned to each official brand tone and style manual.
• Organic visibility: metadata, structured product, offer and review data, citable summaries for generative engines and proactive detection of semantic cannibalization.
• Competitive intelligence: market price monitoring.
• Paid media: dynamic ad copy, enriched feeds and audience segmentation for the advertising platforms.
• None of these roles invades another: each has a written boundary, just like in a human team.

Gates, scoring and human review

• Cross validation between agents auditing generated attributes against the manufacturer standardized specifications before propagating them.
• Automatic reliability scoring: unattended publishing when the score is high, and routing to human curation on outliers.
• That is what makes scaling possible without giving up control: the person does not review everything, they review what the system flags as doubtful.
• With that, regulatory compliance, technical accuracy and brand protection hold at the same time.

Observability and the improvement loop

• Full action auditing and structured logging of every inference.
• Impact measurement on organic conversion and on paid media, feeding back into operational processes.
• Without that record there is no way to answer what an agent did and why, which is the question that appears when something goes wrong.

Architecture and technology

• Central PIM as master data, headless Magnolia as the experience layer and Magento as the transactional engine.
• Microservices and asynchronous queues for demand peaks.
• A multi-agent network with bounded roles, cross validation gates, reliability scoring and selective human review.

How generative AI is integrated and governed

• The solution integrates generative artificial intelligence services: language models that write, converse, extract and structure inside the workflow, rather than a rules engine with prewritten answers.
• Consumption is measured and billed in tokens, separating input, output and cache write and read, which are different line items with different prices and are accounted for separately.
• Telemetry attributes consumption per agent and per client, so the token budget is bounded and controlled per project instead of being read as a single invoice.
• Limits live in the record, not in the model: a record declares roles and caps, and which provider each role maps to is configuration that changes without touching the agent.
• The architecture is multi provider: the provider is fixed per agent and per session, never per request, and automatic retry against a different provider is deliberately switched off. If the provider fails, the agent stops and says so instead of silently continuing on another one.
• Context caching is the main cost lever, because rereading context accounts for most of the token volume in any agent operation. Sustaining the cache is what keeps the same task from costing several times more.
• And the minimum context window is declared per agent: a provider that does not cover it is ruled out by architecture, not by price.

Learn more about AI agents, ecommerce development, software development and Magnolia CMS.

The operation moved from fragmented management to a continuous circuit where data enters once and reaches every channel consistently.

What is running

• More than 15 official brands and tens of thousands of active references governed from a central PIM.
• Magnolia as the unified experience layer, with reusable components, brand landings and multisite management.
• Continuous bidirectional sync of stock, prices and commercial rules with the transactional engine.
• An architecture ready for Hot Sale and CyberMonday peaks through microservices and asynchronous queues.
• Agents specialized in writing, organic visibility, competitive intelligence and paid media.
• Cross validation gates with reliability scoring and selective human curation.
• Action auditing and structured logging of every inference.

What changes in control

• Human review stops being a bottleneck and concentrates on the cases the system flags as doubtful.
• Each brand tone stops depending on who wrote the sheet.
• And every action of every agent is recorded, so the question of what happened has an answer.

Why it matters

• In enterprise ecommerce, automation without gates does not reduce work: it moves it to the team that corrects afterwards.
• Reliability scoring is what makes the automation operable: it defines when the system publishes on its own and when it asks for help.