Skip to content
New: The Agent Solutions Hub is live. Explore real agent use cases

Bash Conversational Shopping Agent

Shoppers who engaged converted 35.2% higher over Black Friday weekend, with revenue per visit up 39.8%.

Request an intro

Three product pages viewed without progress, and the agent starts a conversation. Conversion up 35.2% on Black Friday.

What the agent does

Watches for hesitation, opens the conversation itself after three product pages without progress, then infers preferences and recommends.

Built by
Bloomreach, with no Bash engineering resource
Contributed by
Bloomreach (Loomi Conversational Agent)
Industry
Retail, Apparel & Footwear
Audience
B2C
Category
Search & DiscoveryMarketing & PersonalizationContent & Experience
01
The problem

Why this exists

Bash had a conversion problem with a specific shape: shoppers were browsing extensively and not buying. They viewed multiple products, showed clear intent, and dropped off without completing a transaction or signaling why.

Read more

Static banners, algorithmic product blocks, and generic recommendations were not reaching those shoppers. The team wanted personalization that felt helpful rather than intrusive, and existing approaches could not interpret the intent behind a query like an outfit for a garden wedding that is not too formal.

Scale made it worse. During Black Friday, thousands of high-intent shoppers hesitate at once. Live chat and contact center agents cannot work at that speed or volume, and rules-based automation cannot interpret an abstract multi-step query. Meanwhile fewer than 2% of visitors engaged with traditional chatbots, so the conventional tool was never going to reach the people who needed it.

02
Outcomes

What changes when this is in place

  • In a Black Friday weekend A/B test, shoppers who engaged with the agent saw a 35.2% higher conversion rate, 39.8% higher revenue per visit, and a 28.1% lower exit rate against the control group, all statistically significant.
  • These are engagement-cohort results, measured on top of Bash's existing Black Friday performance, which confirms the agent added value rather than re-routing demand that was already there.
  • The implementation required no engineering resource beyond specifying test criteria, so Bash can expand the agent to new funnel moments without engineering dependencies.
Read more
  • Bash is now extending the agent up the funnel to product listing pages, with pre-generated contextual questions under the add-to-cart button.
03
How it works

Architecture

The agent monitors browsing in real time and triggers on a behavioral condition: three or more product detail pages in one session without progressing toward purchase. That condition is defined in configuration rather than code, so Bash changes thresholds and triggers without re-implementing the agent.

Read more

Once engaged, it answers questions, infers preferences, and recommends from the live catalog. It also decides when not to continue, which is the part that keeps the experience helpful rather than intrusive.

The architecture around it is modular. The agent taps the product catalog and multiple FAQ and content sources, orchestrated through a customer data engine and AI stack drawing on more than 15 years of ecommerce behavioral data. The agent itself runs as a front-end experience layer tied into existing ecommerce and analytics systems, which is what allows fast experimentation and expansion to new page types.

No human reviews or approves individual conversations. Bash configures targeting conditions and success criteria; the model and data connections are tuned separately. Bloomreach performed the data integration without requiring Bash engineering resource. The production launch was August 2024.

A shopper browses bash.com. A behavioral trigger set in configuration, three or more product pages in one session without progress toward purchase, starts the Loomi Conversational Agent, which runs as a front-end experience layer. The agent answers questions, infers preferences, recommends from the live catalog and decides when not to continue. It draws on the Bloomreach customer data engine and AI stack, which connects the product catalog and FAQ and content sources, and ties into Bash's existing ecommerce and analytics systems. Bash configures the targeting conditions and success criteria.

A shopper browses bash.com. A behavioral trigger set in configuration, three or more product pages in one session without progress toward purchase, starts the Loomi Conversational Agent, which runs as a front-end experience layer. The agent answers questions, infers preferences, recommends from the live catalog and decides when not to continue. It draws on the Bloomreach customer data engine and AI stack, which connects the product catalog and FAQ and content sources, and ties into Bash's existing ecommerce and analytics systems. Bash configures the targeting conditions and success criteria.
Open standards used
REST APIs
Tech stack
Bloomreach Loomi Conversational AgentBloomreach customer data engineProduct catalogFAQ and content sources
Get hands-on

Resources

Published case study
Next step

Request an intro

Bloomreach is behind this solution. Ask about recovering shoppers who stall, and the MACH Alliance will put you in touch with the person who can answer.

The MACH Alliance passes your details to a named contact at Bloomreach who knows this work and can answer your questions. We aim to reply within 2 business days. Your details are handled under the MACH Alliance privacy policy.

Not ready to talk to anyone? MACH Insider is the Alliance mailing list. New solutions, program openings, and the research behind them, once a month.

Join MACH Insider

Everyone credited on this solution

Every organization named here has a page listing the rest of its work on the Agent Solutions Hub.

Keep reading

Related solutions

Verification

Published case study

The detail on this page is drawn from a case study published by the contributor or the MACH Alliance. Any figures on this page are the ones that case study states.

Where this came from

Published from the winner's case study and the 2026 MACH Impact Awards announcement.

Status: Live in productionAdded Apr 29, 2026A listing is an editorial decision, not a MACH Alliance certification.
Put a solution forwardBack to the Agent Solutions Hub

Built something worth showing?

Put your agentic solution in front of the enterprise leaders who use the Agent Solutions Hub to decide what to build next. Submissions are open to MACH Alliance Members, Collaborators, or Ambassadors.

Put a solution forward