How AI-Powered Data Ecosystems Improve Retail Resilience in Times of Crisis

Published in

Geek Culture

8 min readOct 16, 2021

Back in the day, retail companies tapped into operational data to identify customer behavior trends and evaluate employee performance — and leveraged these insights to apply corrective actions.

Today, the reactive approach is starting to hit its limits:

The amount of information produced by an average company’s IT systems, employees, partners, and customers is growing exponentially — and almost half of this data goes unused for analytics
Up to 90% of information that comprises our digital universe is unstructured, meaning it comes in the form of videos, images, log files, documents, presentations, social media posts, etc.
95% of businesses are struggling to organize unstructured data and act on it
Poor data quality costs your business anything between $9.7 million and $14.2 million annually

But it’s not only the amount, quality, and heterogeneity of data that prevent companies from capitalizing on their most valuable asset. Until recently, the tools that allow enterprises to process large volumes of data and predict events rather than deal with their consequences were unavailable or too complex to build and deploy.

It all changed with the introduction of AI-driven data ecosystems — a powerful combination of enterprise software, databases, and intelligent tools/algorithms that provides a 360-degree view of a company’s present state.

To learn how such systems could help retailers effectively address current and future challenges while keeping operating costs down and serving clients better, I turned to Vitali Likhadzed and Kirill Stashevsky, Founders of ITRex Group. Their technology consulting and software engineering company specializes in data analytics and artificial intelligence solutions for retail, healthcare, and education.

Meet AI-Driven Data Ecosystems — the Next Step in Business Intelligence and Data Analytics Evolution

A data ecosystem is a set of interconnected business applications, analytics tools, and IT infrastructure solutions that help aggregate, cleanse, store, process, and visualize information produced by a company’s units and technology systems and perform actions automatically.

AI-based data ecosystems may rely solely on internal data (i.e., closed ecosystems) or augment corporate data with insights from other companies and third-party services (i.e., strategic partnerships).

According to Vitali, artificial intelligence and data democratization make modern data ecosystems a significant improvement over traditional BI and analytics tools:

By applying machine learning algorithms and deep neural networks to both structured and unstructured data, companies gain an opportunity to identify recurring patterns, envision scenarios that might impact their operations, automate time-consuming tasks, and empower IT systems to make business-critical decisions autonomously.
By granting access to the insights drawn from corporate data to every employee regardless of their technical background, businesses manage to reduce bottlenecks in their workflows and help workers make better-informed decisions based on machine-driven recommendations.

Vitali singles out three major types of data solutions:

Software solutions like Tableau, Power BI, RabbitMQ, or Kafka that solve a particular business or technical task — e.g., visualizing operational data, ensuring data exchange between IT systems, preventing transaction fraud, or locating misplaced inventory using RFID tags.
Industrial SaaS platforms like SAP, Snowflake, Salesforce, and Kissflow, which provide a single platform for managing customer, sales, and supply chain data. These platforms might lack customization capabilities and become too costly to operate as your business (and the amount of data it generates!) grows.
Integrated enterprise-wide data ecosystems built as a combination of custom-made, industrial SaaS, and open-source software solutions that use data from all of the systems deployed by a company and are best-suited to meet your unique business needs.

How Data Ecosystems Could Future-Proof Your Retail Business

With a whopping 10.5% decline in sales and 12,200 brick-and-mortar stores closed in the USA alone, the retail industry was among the hardest-hit sectors in 2020. Amid the national lockdowns, retailers had to diversify their sales channels, streamline the hiring process to fill labor shortages at essential stores and warehouses, and trim expenses without compromising customer service.

Technology emerged as a viable solution to these daunting problems, with data ecosystems often serving as a basis for company-wide IT system deployments.

State of Retail Tech H2'20 Report from CB Insights indicates that retail technology funding reached a total of $22.8 billion last year, with in-store tech, eCommerce, and supply chain and logistics systems taking a steady lead. And it’s just the tip of the retail transformation iceberg as the novel virus is not going away anytime soon.

The more retail companies jump on the digital transformation bandwagon, the more information they need to handle. For example, every transaction, support ticket, or email from a supplier leaves data trails — and retailers could aggregate this information to reduce operating costs, dynamically change prices to attract more customers, and unlock new growth opportunities.

Top 3 Goals Retail Companies Could Accomplish with AI-Powered Data Ecosystems

Pinpoint Areas for Cost Reduction

A retail company that employs over three million workers worldwide turned to ITRex to create a data ecosystem that breaks down the silos between the company’s units, automatically modifies or deletes inaccurate information, and allows non-technical users to generate insights on customer behavior, operations, and financial efficiency without seeking an IT department’s assistance. For instance, the ML-driven data solution helped the company reduce operating expenses by streamlining the decision-making process for facility managers faced with the dilemma of whether to repair or replace pieces of equipment and other assets.

Another example comes from Tesco. The British supermarket chain enhanced in-store refrigerators with temperature sensors that send telemetry data to a cloud-based data analytics platform every three seconds. By collecting the temperature data over the course of twelve months, Tesco engineers identified refrigerators that had been running at a lower temperature than necessary. The insights helped the company reduce cooling costs by up to 20%.

And New York’s Westside Market became one of the first non-major US retailers to deploy a scan-and-go technology system paired with a mobile app. The solution features a smart scale that automatically identifies goods in a customer’s basket and allows visitors to pay for items by scanning barcodes that pop up on the screen. AI-driven systems like this could help retail companies cut staffing costs while minimizing interactions in physical stores amid the pandemic.

Bolster Sales

Although 81% of retailers collect shopper data, just 16% of companies can analyze the information and use it proactively to anticipate customer behavior.

According to Vitali, ITRex CEO, “With AI-powered data analytics solutions, retailers can identify the optimal strategies for upselling and cross-selling to individuals and increase their revenue per customer. In addition, it may help determine which customers will churn and what would incentivize them to stay. Also, by analyzing sales and pricing data in combination with historical and current market trends, smart data ecosystems can compute ideal pricing models that attract more sales.”

An example of this would be advanced POS solutions that track and calculate various store performance metrics, such as conversion rate, sales per square foot, sales per employee, average transaction value, and online vs. offline sales. Data-driven retailers can harness this data to timely reorder popular items, make personalized offers to customers, and plan their staff rosters better.

Align Supply with Demand

In Q2 2020, Zara, a Spanish fast-fashion brand, accumulated an impressive $254 million in net sales with little to no advertising. Even though online sales comprise just 14% of its revenue, Zara managed to reopen 98% of its stores globally. Zara’s success can be primarily attributed to its end-to-end digital transformation strategy relying on customer intelligence, personalization, and supply chain management technologies. For example, the company only produces a limited amount of merchandise, distributes clothes between stores, tracks each item’s sales via SKUs, and leverages the insights to manufacture the following order. This approach helps Zara avoid overproduction and boost sustainability.

Similarly, H&M uses artificial intelligence to track purchases in each of their 5,000 stores, timely restock popular items, and predict market trends.

And Finesse, a technology-driven fashion brand that raised $4.5 million in funding earlier this year, uses artificial intelligence to parse user-generated content (that is, unstructured data) and discover styles and clothes Internet users are gushing over. Next, the company models clothing items in 3D and tries to predict how many items the new model could sell before submitting the designs to a manufacturer.

As information is becoming retailers’ critical asset, innovative data ecosystems have turned into the underlying technology for other IT solutions — and it’s only a matter of time until more retailers come to understand their benefits.

What Does It Take to Build an AI-Based Data Ecosystem, and Could Smaller Retailers Afford It?

Kirill Stashevsky, ITRex Group CTO, believes the size of an enterprise and its level of digitalization is not a key factor behind the decision to invest or not to invest in intelligent data ecosystems. Despite being around since the 1950s, artificial intelligence remains a terra incognita for most businesses, meaning every company should start their AI journey small and go through a trial-and-error phase until they get a full-featured solution that can be used at scale.

According to Kirill, the fundamentals of an AI-driven data ecosystem include:

A responsive and highly resilient data architecture
Built-in scalability
Intelligent data management driven by AI algorithms
AI-based self-educating insights and actions engine
Established and AI-driven data governance model

To start creating an intelligent data ecosystem, your company needs to:

Talk to internal and external stakeholders and identify business problems you’re aiming to solve
Design a blueprint, vendor-neutral, and technology/infrastructure-agnostic architecture that would allow any stakeholder, regardless of their background, to access data, ask questions, and produce reports that answer them
Lay out the core data governance principles covering security, management, and quality of your corporate data, as well as specifying user roles and access permissions
Identify and prioritize use cases using a product prioritization framework — for instance, the MoSCoW principle
Build data components for priority use cases, leaving an option to scale and expand innovative capabilities horizontally to support other scenarios. That’s how you get a minimum viable product (MVP) containing just enough features to justify the essential use cases and start benefiting from the data ecosystem immediately
Collect feedback from business stakeholders
Tweak the system accordingly
Repeat the steps enumerated above over and over again until you get a comprehensive data ecosystem that sources data from every IT system used by your organization

This approach would help you create a data ecosystem that brings value from day one, supports all types of applications and devices, and can be easily modernized to support new business requirements should the need arise.

Regardless of your company’s size, profitability, and technology readiness level, creating a comprehensive data ecosystem and automating analytics using artificial intelligence is key to prepare your business for change — whether we’re talking about innovation, major shifts in customer behavior, or a pandemic that shuts down brick-and-mortar stores across the globe.

About author: Andrei Klubnikin is a Content Management Team Lead at ITRex Group. With a digital marketing degree from CIM and 7+ years of experience in technology journalism, Andrei explores the transformative impact of AI, IoT, and cloud computing on business operations.