Your storefront runs on knowledge. Too much of it lives in a few people's heads.
Retail and eCommerce businesses run on large catalogs, frequent pricing changes, seasonal peaks, and systems added one channel at a time. The knowledge that ties them together is hard to find, hard to hand off, and unreliable as context for the AI tools your customers and employees now use.
Why Retail Knowledge Is Hard to Keep Straight
Retail knowledge is hard to keep straight because the same facts have to stay consistent in many places at once. A single product's description, price, availability, and return terms may live in an ERP, an eCommerce platform, a product information system, marketplace listings, supplier files, and the answers customer service gives on the phone. Each system was usually added to serve a new channel, and the rules that keep them aligned often live with the people who set them up. When those people move on, or a replatforming project begins, the business finds out how much of its operation depended on knowledge no one had written down.
Core Issues
- Product knowledge in too many places. Descriptions, specifications, and availability are maintained in separate systems, and no one owns keeping them consistent across the website, marketplaces, and customer service.
- Pricing updates that depend on one person. Supplier cost changes, promotions, and special account terms run through manual steps that only one or two employees know how to perform.
- Reports no one can fully explain. eCommerce, finance, and marketing count new customers, orders, and margin differently, because the business logic behind each report lives in queries and in the people who wrote them.
- Answers that depend on who picks up. Customer service staff apply policies, exceptions, and account-specific terms from memory, so customers can get different answers to the same question.
- Seasonal hiring with little time to train. Peak periods bring in temporary staff who need to learn products, systems, and policies in days, while experienced employees are at their busiest.
- Customer insight that stays on the front line. Account managers and service representatives hear why customers are frustrated or leaving, with no established way to get that knowledge to the people who can fix the cause.
Knowledge Gaps Show Up as Returns, Margin Leaks, and Rework
In retail, a missing or outdated piece of knowledge rarely stays contained. It reaches a customer as a listing that doesn't match the product, reaches the P&L as a price that was never updated, or reaches a migration project as a custom rule nobody can explain.
The listing that doesn't match
A product's dimensions, compatibility, or included items are correct in one system and wrong in another. The customer buys from the wrong version, and the business pays for the return, the replacement, and the customer's lost confidence.
The price file only one person can run
Supplier cost updates arrive as spreadsheets that someone matches to the product database by hand. When that person is out, or the backlog grows, the business keeps selling at old prices while paying new costs.
The replatforming decision
Moving to a new eCommerce platform, ERP, or order management system forces decisions about which custom fields, pricing rules, and workarounds to carry forward, often before anyone has documented what they do.
AI Answers With Whatever Your Business Has Made Usable
Shoppers are using AI tools to research products and decide where to buy, and retailers are putting AI assistants in front of customers, store associates, and service teams. Those assistants are only as reliable as the product data, policies, and procedures behind them. When Searchable tested more than 72,000 questions about UK high-street businesses in ChatGPT, Gemini, and Perplexity, 64% of the businesses had at least one false fact returned about them.
An internal assistant drawing on conflicting product records, outdated return policies, or undocumented exceptions has the same problem. It doesn't resolve the conflict. It repeats it to every customer and employee who asks.
The retailers below share one thing: the product and operating knowledge was put in order before AI or self-service could make use of it.
Wayfair cut listing curation time by two-thirds by cleaning up its product catalog.
Wayfair used Google's Gemini models to categorize products, tag attributes such as color and style, and detect errors in product dimensions across a catalog of roughly 30 million items. The company reported 67% less time to curate product listings, savings in the hundreds of thousands of dollars, and a 2% improvement in some conversion rates.
Wayfair and Google Cloud, January 2025uBreakiFix halved onboarding across 685 stores with a single, certified source of answers.
New technicians at uBreakiFix by Asurion spent much of their training memorizing procedures, and answers depended on tribal knowledge. The company built a certified knowledge base as its single source of truth and embedded it in daily workflow. Onboarding time was cut in half, employees found answers 25–35% faster, and service call volume held steady despite rapid store growth.
Bloomfire and uBreakiFix by Asurion, Customer Contact Week 2026The Home Depot grounded its AI assistant in decades of project expertise and product details.
Magic Apron draws on five decades of home improvement project expertise and details on millions of products, and now adds localized knowledge for each U.S. store so customers and associates can find products and ask compatibility questions in the aisle. It is live in more than 2,000 U.S. stores and fields millions of questions a month.
The Home Depot, August 2026Glymr client
J.W. Pepper: Making Complex Retail Data and Processes Usable
J.W. Pepper is the world's largest sheet music retailer, serving music educators, choir directors, band leaders, churches, and individual musicians, with eCommerce at the center of the business. Glymr's work began with the company's Business Intelligence team, helping it better understand the company's complex data, and has since extended into other departments, where documenting business processes is surfacing opportunities to automate or optimize how work gets done.
"Our partnership with Glymr has had a positive impact on our organization. Glymr's expertise has boosted the productivity of our Business Intelligence (BI) team. This has enabled us to better understand our complex data and begin to unlock its full potential. Their methodical approach to knowledge management and analytics has allowed us to gain deeper insights into our operations. Beyond the BI team, Glymr's impact is now extending into other departments within our enterprise. Their efforts to document business processes have been key to finding opportunities to automate or optimize business functions. The standardization and optimization of processes, coupled with our enhanced reporting, is leading to measurable value, reinforcing our strategic goals and optimizing operational functions."
Two Decades of eCommerce Experience Behind the Work
Glymr's founder, Jeff Greenhouse, spent two decades building and marketing eCommerce sites as an agency leader. As Vice President of Subscriber Growth at AMC Networks, he led growth and analytics work focused on acquiring, engaging, and retaining direct-to-consumer streaming subscribers. Glymr's knowledge and process work starts from that working understanding of how product data, merchandising, analytics, and customer operations fit together in a business that sells online.
How Glymr Helps
Glymr works with operations, eCommerce, merchandising, and technology leaders to make product, process, and customer knowledge usable for the employees and AI tools that depend on it. Work can start with a single team, a process such as pricing updates or returns, a replatforming project, or the run-up to a peak season.
Make systems and processes visible. Data Landscape Mapping identifies the systems, databases, and spreadsheets that hold product, pricing, customer, and order data, including ones leadership doesn't know are still in use, which is especially valuable before an eCommerce platform or ERP migration. Process Mapping documents how work such as supplier price updates, product onboarding, order exceptions, and returns actually runs, including the exceptions and special account terms that determine what customers experience.
Protect continuity. Critical Knowledge Capture documents what long-tenured buyers, merchandisers, customer service leads, and systems specialists know about the processes, exceptions, and history that no procedure covers. Knowledge Exit Interviews give departing employees a structured way to hand that knowledge over before their last day.
Keep knowledge current and findable. Knowledge Management Programs put ownership, review cycles, and a clear structure behind product knowledge, policies, customer service answers, and report definitions, so they stay consistent across channels and give AI tools reliable context to work from.
Where to Start
The two assessments answer different questions.
A Knowledge Friction Assessment is the right starting point when the concern is everyday work and continuity: teams searching for product or policy information, repeated questions to the same experts, uneven onboarding for new and seasonal staff, and critical knowledge concentrated in too few people. It measures the friction that affects how your business runs today.
An AI Effectiveness Assessment fits retailers where employees are already using AI regularly, whether for product content, customer service, analysis, or everyday tasks. Based on your own survey results, it shows whether that use is turning into useful work and where it breaks down, whether in the knowledge AI draws on, in access and governance, or in the time employees spend checking and correcting output.
Who This Is For
This is most relevant for CEOs, COOs, CIOs, and eCommerce and operations leaders at retailers and eCommerce businesses facing a platform or ERP migration, the departure of long-tenured employees, a seasonal hiring cycle, or a rollout of AI tools for customers or staff that depend on product and policy knowledge few people hold.
Related Reading
See Operational Inconsistency for a closer look at why the same work gets done differently depending on who does it.
Find out where your product, pricing, and customer knowledge depends on too few people.
A Knowledge Friction Assessment shows where teams lose time and where critical knowledge is concentrated. If AI is already in wide use across your merchandising, eCommerce, and customer service teams, an AI Effectiveness Assessment shows whether that use is turning into useful work.