Retailers face immense pressure to adapt pricing and product assortments to shifting market demands. Manually analyzing competitor data and customer preferences was unsustainable for a mid-sized lighting retailer.
Effective assortment optimization requires retailers to follow several key analytical steps:
evaluating product performance across different dimensions (including sales volume, profitability, and turnover rate)
analyzing customer purchasing behavior to understand preferences
examining market trends to identify opportunities
and optimizing inventory levels to balance availability with investment.
Without proper tools, these complex analyses become overwhelming and time-consuming, leading to missed opportunities and inefficient resource allocation.
Teams spent weeks auditing spreadsheets, struggling to keep pace with pricing changes or spot gaps in their product lineup. The retailer partnered with Akaike to deploy BYOB, a full-stack Agentic AI platform that combines text-to-SQL analytics with contextual insight generation, enabling more efficient insights gathering. Through the conversational assortment intelligence capabilities of BYOB, the retailer was able to make well-informed decisions to optimize their product portfolio and pricing.
The Challenge:
The retailer's analysts faced three core issues:
Delayed Pricing Insights: Competitor price analyses took days to complete manually, leading to missed competitive positioning opportunities.
Hidden Assortment Gaps: Without automated tools, identifying missing products in their catalog relied heavily on anecdotal feedback rather than data-driven decisions.
Data Overload: Teams spent approximately 40 hours weekly cleaning and organizing data from multiple sources, including competitor catalogs.
Existing SaaS tools struggled to interpret industry-specific terms like "lumens" or "IP65-rated," and generic models failed to adapt to the retailer's unique product taxonomy.
The Solution:
BYOB Platform – A Privacy-First AI Analyst
Akaike's BYOB platform provided a solution with Agentic AI architecture that transformed the traditional ETL (Extract, Transform, Load) process into an ETLC (Extract, Transform, Load, Context) paradigm. By adding contextual understanding to data processing, the platform enabled more meaningful interactions with data.
Introducing the ETLC Process for Assortment Intelligence
Extract: BYOB collected data from various sources, primarily competitor catalogs datasets, and integrated them into a unified view. Unlike traditional data pipelines that require rigid schema definitions, BYOB's flexible architecture adapts to different data formats seamlessly.
Transform: The platform automatically cleaned and standardized industry-specific terminology across different sources, creating a consistent taxonomy that bridged the gap between technical database fields and natural human questions.
Load: Processed data was organized in a queryable format that supported natural language interactions, eliminating the need for SQL expertise among retail analysts.
Context: The critical innovation of BYOB was its ability to add contextual understanding to retail data. By recognizing relationships between products, pricing bands, and competitive positioning, the platform enabled analysts to ask questions like "How is competitor X performing in the $20-$50 pricing band?" and receive meaningful insights rather than raw data.
BYOB Key Features and Implementation
Unified Data Access with Natural Language Queries
Multi-Source Integration: Integrated multiple siloed data sources into BYOB through custom automated pipelines.
Visualization & Dashboarding: Results were visualized as dynamic charts and saved to shared dashboards, enabling faster collaboration.
Competitive Pricing Analysis
Product Matching & Comparison: Aggregated pricing data from major competitors. The system used fuzzy matching to reconcile product variations (e.g., "Philips Hue Ambiance" vs "Hue Smart Bulb").
Privacy-First Execution: All data processing occurred within the retailer's secure environment, ensuring sensitive information remained protected.
AI-Powered Assortment Gap Detection
Decision Tree Analysis: Analyzed competitor catalogs to identify potential product gaps in the retailer's assortment, using a structured decision-making approach rather than complex statistical methods.
Self-Hosted & Adaptable: The system was designed to recognize industry-specific terminology and adapt to the retailer's unique product categorization.
How It Worked: Agentic AI for Retail Agility
Leveraging BYOB eliminated data engineering bottlenecks:
Conversational Interface: Analysts interacted with data through a conversational bot, making complex queries more accessible to non-technical team members.
Guided Feedback Loops: Teams could correct misinterpretations (e.g., conflating "smart bulbs" with "smart switches"), and BYOB adapted to these corrections.
Automated Data Organization: The platform suggested improvements to help standardize product attributes across catalogs.
Results:
Why BYOB Stands Out
Transparent AI: The retailer retained full ownership of models and data.
Enterprise-Ready Privacy: Self-hostable with robust security features.
Adaptable System: Capable of learning from feedback to improve performance over time.
Akaike's BYOB platform transformed raw data into an actionable assortment strategy. By combining conversational intelligence with enterprise-grade security, the retailer shifted from reactive data cleaning to more proactive assortment planning—whether optimizing product selections or identifying new market opportunities.
Ready to Deploy Your AI Analyst?
BYOB is not merely a tool; it is your strategic partner in navigating the complexities of the business intelligence landscape. Join the ranks of innovative business leaders who are redefining their approach with BYOB.
%20(1).png)
.png)
%20(1).png)
