What problem does it solve?
A store associate gets questions on thousands of products, projects and store procedures, and the answer sits in a product catalog, a process guide, an inventory system or another colleague's head. Lowe's says associate expertise is uniquely valuable for home improvement retailers because customers ask daily about a variety of complex projects and thousands of different products. A new hire cannot know all of it, and an experienced associate asked about another department cannot either.
The cost shows up in two places. The customer waits while the associate looks for a manager or a terminal, and a new or seasonal hire takes longer to become useful. Process questions, such as how to restart a register after a power outage or how to take back an item without a receipt, are examples that can hold up a queue.
How does it work?
- Connect the sources. The assistant is grounded in the retailer's own material: product data, project guides, process documents and store procedures, and live inventory and schedule lookups through the systems that hold them.
- Put it where the associate works. It runs as an app on the handheld devices or in the associate app already in use, and accepts typed or spoken questions, as in Lowe's voice to text for hands free use.
- Answer from the sources. The associate asks a question in plain language, for example how to sign a guest up for a loyalty card. The assistant retrieves the relevant guide or record and returns steps or facts in seconds, rather than inventing them.
- Turn long guides into steps. In June 2025 Walmart announced an upgrade to its associate assistant, due to launch in the coming months, that would use generative AI to turn lengthy process guides into clear step by step instructions, for questions such as processing a return without a receipt. At that date it was an announcement, not a delivered capability.
- Coach and escalate. The assistant can act as a coach for new and seasonal hires, as Target describes for Store Companion. A well designed one says when the sources do not cover a question and points the associate to a manager or a specialist.
- Audience
- Employee facing
- Autonomy
- Assist
- Adoption
- Early adopters
- Channels
- Internal tools, Mobile app
What is it worth?
Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Interactions handled | Not pooled | at least 3 million | 1 | 1 organization |
| Satisfaction uplift | Too few to pool | about 2% | 1 | 1 organization |
| Users served | Not pooled | at least 900,000 | 1 | 1 organization |
Value drivers: Employee productivity, Customer experience, Lower cost to serve.
Indicative value
A home improvement or grocery chain with 300 stores
USD 875,000 to USD 10.8 million
Value of associate time saved on lookups per year
How this is calculated
Formula: stores * associatesPerStore * queriesPerAssociateDay * minutesSavedPerQuestion / 60 * workingDays * hourlyCost. The low scenario uses every low input, the high scenario every high input.
| Input | Low | High | Basis |
|---|---|---|---|
| Stores in the chain stores, stores | 250 | 350 | Illustrative chain size, well below the 1,700 Lowe's stores and the nearly 2,000 Target stores that Target's release names. |
| Associates who use the assistant per store associatesPerStore, associates per store | 50 | 70 | Editorial assumption, replace with your own headcount of floor associates with a device. |
| Questions an associate asks the assistant per working day queriesPerAssociateDay, questions per associate per day | 2.5 | 3.5 | Derived from Walmart's reported 3 million queries per day and 900,000 weekly users, about 3.3 per weekly user per day if every user asked daily. Walmart does not state the ratio itself, so the range is kept around it. |
| Minutes saved per question compared with finding the answer another way minutesSavedPerQuestion, minutes | 0.5 | 1.5 | Editorial assumption, replace with your own time study of how long a lookup, a call to a colleague or a walk to a terminal takes today. |
| Working days per associate per year workingDays, days | 210 | 230 | Editorial assumption for a mix of full time and part time associates. |
| Fully loaded cost of an associate hour hourlyCost, USD per hour | 16 | 22 | Editorial assumption, replace with your own fully loaded wage cost. |
What it leaves out: The result is time that associates can spend with customers, not cash that leaves the payroll. It leaves out the cost of building and running the assistant, any change in sales or satisfaction (Lowe's CEO is reported to have said customer satisfaction is about 2% higher when associates use its tool, without describing the comparison), and the faster onboarding of new hires, which needs its own estimate.
Who already uses it?
3 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.
Lowe's
United States · Retail and ecommerce · 2025
Lowe's launched Mylow Companion on 5 May 2025 to every associate across its more than 1,700 stores. The generative AI app runs on the handheld devices associates already carry on the sales floor and answers questions about product details, project advice and inventory in natural language, including by voice. Lowe's built it with OpenAI on the same foundation as Mylow, its customer facing advisor, and positions it as a way to speed up onboarding of new hires. On the third quarter 2025 earnings call, CEO Marvin Ellison said customer satisfaction scores were about 2% higher when employees use Mylow Companion in stores, a comparison Lowe's did not describe further.
- Satisfaction uplift: about 2%, fiscal Q3 2025 earnings call of 19 November 2025, as reported by Digital Commerce 360 on 20 November 2025
"Additionally, customer satisfaction scores are about 2% higher when employees use Mylow Companion to help consumers shop in stores, Ellison said."
Claimed by: organization
Walmart
United States · Retail and ecommerce · 2025
Walmart associates in the United States have used a conversational AI assistant in the Walmart associate app for about five years to ask where products are located, check their schedule and look up how to do tasks. On 24 June 2025 Walmart said the assistant has more than 900,000 weekly users and handles over 3 million queries per day, and announced a generative AI upgrade that turns long process guides into step by step instructions, for example how to process a return without a receipt. In the same announcement Walmart described a real time translation tool for associates and customers in 44 languages. Walmart gives no accuracy or time saving figure for the assistant itself.
- Users served: at least 900,000, per week, as of 24 June 2025
"With more than 900,000 weekly users and over 3 million queries per day, Walmart’s conversational AI has become an essential resource for associates."
Claimed by: organization - Interactions handled: at least 3 million, per day, as of 24 June 2025
"With more than 900,000 weekly users and over 3 million queries per day, Walmart’s conversational AI has become an essential resource for associates."
Claimed by: organization
Target Corporation
United States · Retail and ecommerce · 2024
Target announced Store Companion on 20 June 2024, a generative AI chatbot built by its own technology team that runs as an app on the handheld devices of store team members. It answers process and procedure questions, for example how to restart a cash register after a power outage or how to sign a guest up for a Target Circle Card, and acts as a coach for new and seasonal team members.
Target's release names nearly 2,000 stores for the rollout. The team trained it on real frequently asked questions and process documents from stores, went from first tests to planned rollout in six months and piloted it in about 400 stores before the chainwide launch planned for August 2024. A trade report in June 2025 states that Target rolled out Store Companion the year before. Target disclosed no usage or outcome figures.
No outcome disclosed.
How do you implement it?
A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.
Data you need
- Current process guides, store procedures and training material in a form that can be indexed
- Product and project content, including regional or banner differences
- Read access to inventory, location and schedule data for the associate's own store
Systems to integrate
- Associate app or handheld device software, with single sign on
- Inventory and product information systems
- Workforce or schedule system, if schedule questions are in scope
- Document or content management system that holds the process guides
Complexity: Medium
The build starts with the content: product data, project and process documents, each with an owner who keeps it current, and read access to inventory and schedule systems. Store handhelds or an associate app need single sign on and a place for a chat panel, and the retailer needs a feedback loop from associates to correct wrong answers.
- 1
Start with one question type and one department
Pick the questions that hold up the floor most often, such as register procedures or returns rules, and launch in a few stores. Target piloted Store Companion in about 400 stores before its chainwide rollout.
- 2
Ground every answer in a named source
Show the guide or record behind each answer so the associate can check it and a content owner can fix it. An answer without a source on a process question is a risk at the register.
- 3
Give content owners a correction path
Let associates flag a wrong or stale answer from the answer itself, and route the flag to the person who owns that document, with a deadline.
- 4
Add live data once static answers hold
Add inventory, location and schedule lookups after the document answers are reliable, because a wrong stock count costs more trust than a missing one.
- 5
Measure use before claiming value
Track weekly users, questions, answers marked helpful and the share that end in an escalation, then compare customer satisfaction between stores or shifts with and without it.
Guardrails
- Answers on pricing, returns, age restricted sales and safety procedures come only from approved documents
- The assistant says when it has no source and points to a manager, instead of guessing
- Question logs are kept for quality review under a stated retention period and are not used to rank individual associates
- New or changed process documents are tested against a fixed set of questions before they go live
KPIs to instrument
- Weekly active associates as a share of associates with a device
- Questions per day, by topic, and the share answered from a source
- Share of answers marked helpful or corrected by associates
- Customer satisfaction in stores or shifts that use the assistant compared with those that do not
Human in the loop
The associate stays the decision maker: the assistant informs, and the associate talks to the customer and carries out the task. Content owners review flagged answers, and store managers decide which procedures the assistant may explain.
Common failure modes
- Stale process guides
- A returns or promotion rule changes and the assistant keeps quoting the old guide. Give each document an owner and a review date, and retire guides that nobody maintains.
- Confident answers without a source
- A fluent wrong answer on a procedure is more damaging than a slow right one. Require a retrieved source for every procedural answer and show it.
- Slow response on the floor
- An answer that takes longer than asking a colleague will not be used. Test latency on the actual handheld and store network, and keep answers short.
- Usage read as value
- High query counts show that associates open the tool, not that customers are served better. Pair usage with a satisfaction or conversion comparison, as Lowe's does with its satisfaction figure.
What are the risks and rules?
EU AI Act
Depends on design
An assistant that only answers product, process and inventory questions for employees is not listed in Annex III and is not a high risk system, with AI literacy duties for the deployer under Article 4. Article 50(1) transparency duties also reach systems that interact directly with people, employees included, unless it is obvious from the circumstances that they are dealing with AI.
Annex III point 4(b) covers systems intended to allocate tasks based on individual behavior or personal traits or to monitor and evaluate worker performance and behavior. An associate app that also assigns tasks to individual associates or scores them falls to be assessed under that point, so the design decides the tier.
Walmart's separate AI task tool, which recommends and prioritizes tasks for overnight stocking, shows how close the two categories sit, and the source does not say whether it allocates by individual behavior.
Guidance
- Regulation (EU) 2024/1689 (EU AI Act), Article 4 and Annex III point 4 (European Union, Europe). Article 4 requires AI literacy for staff who use AI systems; Annex III point 4 lists employment and worker management systems, which applies if the assistant is extended to task allocation or performance monitoring.
- Regulation (EU) 2016/679 (GDPR) (European Union, Europe). Question logs identify associates and can reveal their knowledge gaps, so purpose limitation, retention and transparency duties for employee data apply.
Controls to put in place
- A named owner and review date for every document the assistant can quote
- Logging of questions, retrieved sources and answers, with a retention period agreed with employee representatives
- A visible way for associates to flag a wrong answer, with a tracked fix
- A test set of procedure questions that must pass before a content change goes live
Frequently asked questions
- How is a store associate assistant different from a customer shopping assistant?
The associate assistant is employee facing, so the associate can judge an answer and correct it before passing it on, which makes errors easier to catch than with a customer facing bot. It does not guarantee they are caught, because an associate can still pass a wrong answer to the customer. Lowe's built Mylow Companion on the same foundation as its customer facing Mylow advisor, which shows the two can share models and content while serving different users. Assistants for telecom shops, including customer facing kiosks and digital humans, are covered on the retail store and kiosk assistant page.
- What results have retailers reported?
Walmart reports more than 900,000 weekly users and over 3 million queries per day on its associate conversational AI. That assistant has been in use for about five years and is not the generative upgrade Walmart announced in June 2025, so the figures do not measure generative AI. Digital Commerce 360 reports that Lowe's CEO said customer satisfaction scores are about 2% higher when employees use Mylow Companion, without describing the comparison. Target has disclosed no outcome figures for Store Companion. The figures are the retailers' own statements, not audits.
- How do we know whether a store associate assistant is working?
Usage shows that associates open the tool, not that customers are served better. Walmart reports weekly users and queries per day, and Lowe's reported a satisfaction difference without describing the comparison. Track weekly active associates, answers marked helpful or corrected, and the share that end in an escalation, then compare customer satisfaction between stores or shifts with and without the assistant.
How to cite this page
Blits.ai AI Use Case Library, "AI assistant for retail store associates", last verified 14 October 2026, https://www.blits.ai/ai-use-cases/store-associate-assistant. Licensed under CC BY 4.0. Method: how we verify use cases.
Changelog
- 14 October 2026: First published