# The Rise of Autonomous Stores: How AI Will Run 70% of Ecommerce Operations by 2028

Autonomous stores are self-governing commerce systems in which AI predicts, decides and acts across merchandising, marketing, pricing, customer service, logistics and storefronts. As of November 2025, analysts project they will handle 60–70% of ecommerce operations by 2028.

- Type: Insight
- Category: AI automation
- Author: Isia Lin, Commerce strategist, Matter
- Published: 2025-11-28
- Reading time: 6 minutes
- Web page: https://matterdesign.com.au/news/the-rise-of-autonomous-stores-how-ai-will-run-70-of-your-ecommerce-operations-by-2028

## What does this article argue?

- The article, published in November 2025, argues that the biggest change in ecommerce over the next five years will be stores becoming self-governing systems that make decisions, optimise experiences and run operations without human intervention.
- Leading analysts project that by 2028 autonomous store systems will handle 60–70% of ecommerce operations, from merchandising and marketing to customer experience (CX), logistics and revenue optimisation.

The shift is driven by the convergence of:

- Agentic AI
- Predictive analytics
- API-driven architectures
- Auto-optimising UX systems
- Autonomous workflow engines

## Why does the article call autonomous commerce inevitable by 2028?

Between 2020 and 2024, ecommerce operations became much more complex:

- Shopify's app market grew to over 10,000 apps.
- The average brand runs 250–400 recurring workflows per week.
- Global ecommerce data doubled every 12–18 months.
- 71% of consumers expect personalised experiences (McKinsey, 2024). Source: https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-value-of-getting-personalization-right-or-wrong-is-multiplying
- The average retailer manages 5–9 channels.

The article states that by 2025–2026 this complexity will exceed what teams can manage manually, and AI will become the only viable operator rather than a competitive advantage.

## What is an autonomous store?

An autonomous store is not an AI plugin or chatbot. It is a self-governing commerce system with four integrated layers.

### Layer 1: Predictive intelligence

AI models predict:

- SKU demand, with 92% accuracy
- Customer intent in real time
- Supply chain constraints 2–6 weeks ahead
- Lifetime value (LTV) and churn risks
- Price elasticity and buying probability

These models typically combine LSTMs, transformer-based models and time-series forecasting frameworks such as Prophet and DeepAR.

### Layer 2: Decision engines (agentic AI)

AI agents decide:

- The optimal product ordering
- Which creative to deploy
- How to allocate marketing budget
- What discount is needed to convert
- Which warehouse routes the order

Methods used:

- Policy-based reinforcement learning
- Constraint-aware optimisation algorithms
- Contextual bandits for experimentation

### Layer 3: Execution layer (autonomous workflows)

AI triggers actions through:

- API calls
- Webhook-triggered automation
- Auto-generated code snippets
- Dynamic schema updates
- Edge-side deployment of personalisation rules

The article states that this layer eliminates human execution.

### Layer 4: Governance and guardrails

- Human approval flows
- Audit logs
- Rollback mechanisms
- AI quality assurance pipelines
- Ethical scoring on decisions

This layer keeps the system aligned with brand and financial goals.

## Which ecommerce functions will AI automate by 2028?

Projected automation levels by function, as stated in the article:

- Merchandising: 90% automated
- Marketing operations: 60–80% automated
- Pricing and promotions: 65–90% automated
- CX and customer service: 80–95% automated
- Logistics and inventory: 50–80% automated
- UX and storefront personalisation: 70–90% automated

## How will AI automate merchandising?

- Merchandising is becoming the first major function to go autonomous.
- Gartner predicts that by 2027, 80% of product data management will be AI-generated.
- AI will manage product categorisation, attribute extraction, variant prediction, SEO metadata generation, collection curation, dynamic product sequencing, product detail page (PDP) content optimisation and automated badges (low stock, trending, high conversion).
- Large language models, computer vision and structured catalogue data make merchandising suited to automation.

## How will AI automate marketing operations?

- Early tests show AI-generated campaigns reduce cost per acquisition (CPA) by 23–32%.
- AI systems will generate creative assets, predict campaign return on ad spend (ROAS), adjust bids in real time, identify audience microsegments, run multi-variable tests, set budget allocation and schedule promotions.
- Technical drivers: generative model pipelines, multi-arm bandit testing, predictive budget modelling and closed-loop feedback between ads and onsite behaviour.

## How will AI automate pricing and promotions?

- Dynamic pricing becomes the baseline.
- Retailers using dynamic pricing see a 30% margin lift. Source: https://www.luckyorange.com/blog/posts/ecommerce-dynamic-pricing
- AI evaluates competitor prices, supply and demand ratios, inventory ageing, customer intent scoring, price elasticity and lifetime value projections.
- By 2028, autonomous systems will run flash sale optimisation, personalised pricing, inventory-based discounting, bundle structuring and B2B volume pricing rules.

## How will AI automate customer service?

- AI moves from responding to resolving, using action-oriented agents, API-enabled problem solvers and end-to-end workflow drivers.
- LLM-based agents already resolve 60–80% of ecommerce tickets without escalation, as of November 2025. Source: https://primotly.com/article/how-e-commerce-brands-use-ai-agents-to-handle-80-of-customer-inquiries-automatically
- AI will process refunds, create replacements, adjust subscriptions, manage warranties, handle complex inquiries and pass exceptions to humans.

## How will AI automate logistics and inventory?

- Predictive logistics systems can reduce shipping costs by 12–18% and out-of-stock events by 35%. Source: https://www.parcelperform.com/insights/how-ai-reduces-ecommerce-logistics-costs
- AI optimises routing, label printing, packaging decisions, warehouse selection, demand forecasting and inventory balancing across regions.

## How will AI personalise storefronts?

- The storefront will rebuild itself for each visitor in real time.
- Changes include adaptive homepages, AI-refactored navigation, personalised PDP layouts, intent-aware checkout flows, contextual cross-sells and session-based optimisation.
- Technical basis: edge-side rendering, dynamic middleware, vector embeddings for user context and streaming inference models.

## How are Shopify, BigCommerce and Adobe enabling autonomous commerce?

Shopify, described as an AI-native merchant operating system:

- Sidekick 2.0, an operations-level execution agent
- Predictive APIs for LTV, demand and probability of purchase
- Hydrogen and GenUI, for AI-generated storefront components
- AI-native B2B workflows, such as automated pricing and catalogue rules
- Shopify's long-term strategy is to be the central decision-making hub for SMB to mid-market merchants.

BigCommerce, described as composable AI orchestration:

- AI-driven catalogue management
- AI orchestration across PIM, ERP and search
- Enterprise-scale product classification
- Prediction-driven headless search
- BigCommerce is betting on open architecture plus AI coordination.

Adobe Commerce, with Experience Cloud and AI, differentiated by enterprise-grade first-party data and real-time AI:

- AI-powered segmentation
- Real-time CDP orchestration
- Genetic model-driven asset generation (as written in the article)
- Dynamic page adaptation
- B2B purchasing intelligence

## What is the architecture of a fully autonomous store?

- Data layer: unified CDP, PIM, CRM and OMS
- Intelligence layer: LLMs, predictive models and reinforcement learning systems
- Decision layer: agentic systems and action frameworks
- Execution layer: APIs, automations and integrations
- Governance layer: rules, QA and approval flows

The article expects this stack to become the industry norm by 2027–2028.

## What will retail teams look like in 2028?

Teams shift from operational to supervisory work:

- Marketers become growth architects.
- Developers become system orchestrators.
- Merchandisers become strategic planners.
- CX teams handle only high-complexity cases.
- Leaders focus on model oversight and brand positioning.

Accenture projects that AI-augmented teams will be 40% more productive with 30% less operational overhead.

## What should brands do to prepare?

Steps for 2024–2027:

1. Invest in clean, structured data: AI autonomy depends on clean PIM and CDP systems.
2. Move toward composable or flexible architectures: closed monoliths restrict automation.
3. Adopt AI-native tools with agentic execution capabilities, not AI app plugins.
4. Build governance policies early to control bias, accuracy and financial limits.
5. Start running automation pilots now.

## What will autonomous stores do by 2028?

The article expects autonomous stores to become the default model for ecommerce by 2028. They will:

- Predict customer intent
- Decide optimal pathways
- Deploy experiences in seconds
- Run campaigns without human input
- Optimise revenue automatically
- Route logistics without oversight
- Provide fully autonomous CX
- Learn and improve continuously

The article states that the advantage goes to teams that adopt autonomous systems, because they can operate at a scale, speed and precision that human-only systems cannot match.
