Skip to main content
Skip to footer
We're taking you to another TCS website now.


 

 

HIGHLIGHTS

  • There is increasing demand for AI intervention in the retail business processes due to the massive disruption in customer behavior and purchasing patterns.
  • This has forced retailers to take digital transformation seriously and reshape their legacy business processes with AI.
  • The 3A (assess, analyze, advice) framework is a knowledge-based recommendation engine that helps retailers improve AI maturity and reap the benefits of digital transformation.

 

 

Tridib Halder

Enterprise Architect, TCS

 
SURGE IN AI DEMAND

The demand for AI intervention in the retail business processes has ramped up significantly due to the massive disruption in customer behavior and purchasing patterns, customer product affinity, e-com sales, and in-store operations.

This has encouraged retailers to invest in digital transformation to reshape their legacy business processes with AI. According to Meticulous Market Research®, a leading market research company, “Artificial Intelligence in Retail market is expected to grow at a CAGR of 34.4% from 2020 to 2027 to reach USD 19.9 billion by 2027.” However, many enterprises find incorporating AI challenging for various reasons.

 

 
CHALLENGES WITH AI ADOPTION

Widespread AI adoption faces multiple challenges.

Despite investments and its obvious benefits, AI’s adoption in most retail organizations has remained siloed and short-lived. Six key challenges prevent widespread adoption of AI:

  1. Skepticism about long-term ROI – The unavailability of business value engineering framework makes it difficult to measure, monitor, and amplify the impact of AI in business processes.
  2. Lack of awareness – Key decision-makers don’t have any framework to get a 360o view on AI maturity of their organization.
  3. One-solution-fits-all approach – Strategists usually define a fixed set of AI improvement at an enterprise level. However, product teams or individuals find it difficult to follow the requirement due to their varied (present) maturity levels.
  4. Lack of governance No standardized policies, processes, and frameworks to optimize AI investment and improve AI maturity over time. This is often caused by:
    • Lack of AI governance policies and standards enforced by AI office
    • Lack of processes to rationalize and re-use developed AI-ML assets
    • Unavailability of framework to assess current AI maturity and recommend improvements
  5. Finding right data – Identifying ‘quality’ data without any bias and drift often becomes a challenge for data scientists and analysts. This often degrades value delivery and the time to market of AI solutions, causing business sponsor dis-satisfaction.
  6. Insufficient talent – Many organizations are hampered by insufficient in-house talent required to embark on an AI-ML journey. Missing fully digitized and dynamic skill development programs often restricts AI-ML adoption within an organization.

 

 
THE FOUNDATION FOR AI OFFICE

Over the last few decades, the growing emphasis on data office has helped organizations to become more data-driven and enabled data foundation for AI experimentation.

However, AI maturity improvement and adoption remains too complex due to the absence of a strong AI maturity improvement framework driven by focused function like AI office, hampering the widespread adoption of AI by data-driven retailers.

These are the different levels of AI maturity, based on observations and market research.

3A (assess, analyze, advice) framework for AI office will improve the AI maturity of retailers. The framework is a knowledge-based recommendation engine to improve AI maturity. This module is powered by a knowledge repository, managed by the governance team, and enriched by ‘crowdsourced’ market research data. The repository has two key segments—survey questions and recommended actions—and supported by the following functions:

  • Best practices, trends derived (web-scraping and NLP module) from market research and retailer success stories. This acts as a reference for knowledge repository.
  • Feature store and model registry to capture re-usable AI-ML modules and features
  • Rule Engine to automatically calculate multi-level scoring (focus of module 1) and associate the next best action (focus of module 3) based on survey response.

The engine recommends AI maturity improvement actions based on survey response. It can help stakeholders at different levels, like CDO (chief data officer), CAIO (chief AI officer), PO (product owner), managers, and associates of organizations to understand AI maturity gaps and the improvement actions required. Stakeholders can run a survey using a survey screen and get persona-based recommendations driven by knowledge repository and recommendation engine of the 3A framework.

One sample outcome for a retail store operations team is shown below.

  • Org Maturity >> strategy – Gap in store staff awareness on benefit of AI-driven smart store operations. Share link for business value engineering outcome and process document.
  • Analytics maturity>> AI knowledge – Store staff needs training on new application. Share suggested training links and upcoming events.
  • Analytics maturity>> AI asset – The customer domain team has created a model registry and can be re-purposed. Share the model registry reference with a link of process document.
  • Analytics maturity>> Data foundation – Link of current data quality dashboard focusing on store operations and related CDE (critical data elements) 

 

 
3A FRAMEWORK FOR AI MATURITY

The 3A framework for AI office is a combination of three modules (assess, analyze, advice) to help organizations define the next-best action in AI maturity improvement, depending on the measured gap between current and aspired state.

  1. Assess Module – This AI maturity assessment module captures user survey responses and creates a multi-level AI maturity scoring based on knowledge repository and hierarchy as depicted below. Questions will vary based on user persona and are driven by knowledge base.
  2. Analyze module – This module is powered by a knowledge repository, managed by the governance team, and enriched by ‘crowdsourced’ market research data. This knowledge repository has two key segments—survey questions and recommended actions—and supported by the following functions:
    • Best practices, trends derived (web-scraping and NLP module) from market research and retailer success stories. This acts as a reference for knowledge repository.
    • Feature store and model registry to capture re-usable AI-ML modules and features
    • Rule Engine to automatically calculate multi-level scoring (focus of module 1) and associate the next best action (focus of module 3) based on survey response.
  3. Advice module – This provides insights on the next-best action to achieve desired AI maturity. The recommendations will be personalized and based on the state of AI maturity aspired for. Improvement actions are prescribed based on maturity gaps calculated (current vs aspired) and will vary based on user persona (individual, product owner, CXOs). Recommendations will be categorized in six segments.

 



AI OFFICE ECOSYSTEM CRITICAL FOR RETAILERS’ GROWTH

The presence of artificial intelligence in the retail industry is bound to grow in the years to come.

This makes it increasingly important for retailers to build a sustainable ecosystem for AI maturity improvement. An AI office ecosystem powered by the 3A framework can play a critical role in fulfilling the aspiration of retailers in accelerating AI adoption across organization and business processes.