Skip to content
Book a Demo

Turn product and policy knowledge into better customer decisions.

A wide range, promotions that change weekly, a returns policy that depends on channel and date, and a workforce that turns over every season.

Book a Demo Explore ClearMash Brain

The answer changes by channel, by date and by store

A retail customer asks simple questions with complicated answers. Can I return this? Does the promotion apply to my basket? Will this work with the model I already own? What happens if it was a gift, bought online, and I want to bring it into a store?

Each of those depends on the purchase channel, the purchase date, the promotion running at the time, the product category, the payment method and occasionally the specific store. The rules exist, and they are rarely written in the order a customer or a colleague on the shop floor needs them.

Staffing makes it harder. Seasonal peaks bring in temporary colleagues who are trained quickly and expected to handle the same questions as a supervisor with five years of service. Franchise and concession arrangements add legitimate local variation on top.

And every answer is compared. Customers check the website while standing at the counter, and any difference between what the page says and what a colleague says becomes the conversation.

Head office and stores run on different clocks. A policy decided centrally on Monday reaches colleagues through a system message, a briefing and a supervisor’s summary, and by Friday several versions of it are in circulation.

What a retailer keeps in its Brain:

  • Returns and exchange rules by channel and date
  • Promotion mechanics and their exclusions
  • Price matching and adjustment rules
  • Refund methods by payment type
  • Store returns for gifts and online orders
  • Gift card and voucher terms
  • Product specifications and compatibility
  • Care, warranty and aftercare terms
  • Delivery and collection rules
  • Loyalty program rules
  • Product recall notices
  • Store and franchise variations

What it costs at the counter

The immediate cost is margin. When a colleague is unsure, the safe answer is usually the generous one, and goodwill given by default is a policy that nobody approved and finance discovers later.

Then there is the escalation to a manager for decisions that have defined answers, during the hours when the manager is most needed on the floor.

Conversion follows. A customer who cannot get a confident answer about compatibility, availability or the returns position frequently postpones, which in retail generally means never.

Colleague confidence has a commercial effect. Someone who is unsure recommends less, adds less and defers more decisions, which is invisible in every report except the sales figure.

  • Unnecessary goodwill

    Refunds and exceptions given because the correct position was not to hand.

    Less unnecessary goodwill

    An exception is a decision, taken with the correct position in hand.

  • Manager escalation

    Supervisors pulled into decisions the policy already covers.

    Fewer manager escalations

    Colleagues settle what policy covers, and managers stay on the floor.

  • Lost sales

    Uncertainty at the point of decision, converted into a postponed purchase.

    Better conversion at the decision point

    Compatibility and availability questions answered while the customer is still deciding.

The site, the store and the assistant give one answer

Retail was early to customer-facing AI, and the failure mode is easy to picture: an assistant answering policy questions from a page written for a different channel, or from a promotion that ended last week.

Because the customer can check, inconsistency is immediately visible, and it undermines the assistant faster than any single wrong answer would.

Governed policy and product knowledge behind every channel is what makes the assistant useful. Same rules, same promotion logic, same returns position, whether the customer is on the site, in the app, on the phone or at the counter.

What a colleague has to settle at the counter

A colleague at the counter has seconds. These decide what they can say and do.

  • WHATWhat the returns position is on this item, bought on this date, in this channel.
  • WHICHWhich promotion was running when they purchased, and whether it stacks.
  • WHOWho can approve an exception here, and up to what value.
  • WHENWhen this policy change takes effect across stores.
  • WHYWhy the answer is this one, said at the counter without calling a manager.
  • HOWHow to explain a refusal so the customer accepts it.
A simple question with a complicated answer. The item carries its own facts - the channel it was bought in, the date, the promotion that was running - and together they select one cell of the returns grid, not one policy, so the colleague states the position at the counter and stays inside the discretion they hold.

Bought online, as a gift, on promotion, brought into a store.

The colleague answers, and does not give it away.

A simple question with a complicated answer. The item carries its own facts - the channel it was bought in, the date, the promotion that was running - and together they select one cell of the returns grid, not one policy, so the colleague states the position at the counter and stays inside the discretion they hold.

What Products · Promotions · Policy

  • Returns policy
  • Promo from Jun
  • Channel rule
  • Promo terms
  • Stacking rule
  • Gift receipt
  • Compatibility
  • Care guide
  • Price match
  • Warranty terms
  • Store variation

How At the counter · Approvals · Wording

  • At the counter
  • Check the date
  • Approval limit
  • Exception path
  • Refusal wording
  • Refund steps
  • Exchange steps
  • Escalation rule
  • AI agent skill
  • Record the case
  • Order lookup

Context Decides which What and which How to act on, for this basket

  • Who This customer
  • When Bought on
  • Where Online or store
  • Which Promo running
  • Why Why this answer
  • Store colleaguesanswer at the counter
  • Customersthe same on the app
  • Self-servicereturns without contact
  • The exception is visible now
Store colleague support Policy and product answers at the counter, in seconds. Sales Customer service One position across contact center, chat, app and store. Customer Service Product knowledge Specifications, compatibility and care information colleagues can use. Knowledge & Portal Seasonal onboarding Temporary colleagues productive within days. Organizational Learning Self-service Customers resolving returns and policy questions without contact. Self-Service Customer Support

Tangible impact

The changes a working Brain makes, in the order a shop floor meets them.

Business impact

  • Reduction in new employee onboarding 70%
  • Reduction in expert escalations 65%
  • Increase in customer satisfaction 15%
  • Increase in first contact resolution (FCR) 40%
  • Reduction in inquiry handling time 28%
  • Increase in self-service adoption 30%
  • Increase in employee satisfaction 62%
  • Reduction in compliance violations 14%

Technological impact

  • Increase in AI agent speed 25%
  • AI agent accuracy from 85% 97%
  • Reduction in AI costs 74%

Additional measures

And the measures a retailer already instruments across its channels:

  • Returns and exception rates against policy
  • manager escalation volume
  • contact rate per order
  • self-service completion
  • time to competence for seasonal staff
  • consistency between channels

Control across stores and franchises

One owner per policy area, approval before publication, and access that respects franchise, concession and regional boundaries. Orders, stock and customer accounts stay in the commerce platform. Where stores are franchised, ownership of local variation can sit with the franchisee while brand-level policy stays central.

See a return handled for its channel and date

Book a demo and we will show you a return settled from the channel, the date, the promotion and the payment method, alike at the counter, on the site and in the assistant.

Book a Demo Explore ClearMash Brain

Prefer to start with the numbers? The impact estimator works them out