New ResearchCorpBenchSee the results ↗We tested AI models on real marketing tasks.
Kohepets
Pet retail & pharmacyMarketing

Personalized replenishment reminder agent

Kohepets makes repeat-purchase reminders fit the customer’s buying pattern.

A replenishment agent estimates when customers may be ready to reorder eligible everyday pet products. It prepares reminders around purchase patterns, current availability and contact preferences, with recent orders checked before sending.

Summarize with AI

The impact

67%

reduction in reminder preparation time

Personalized replenishment reminder agent

Case study facts

Team
Marketing
Systems connected
Not published
Time to live
Not published
Engagement
AI implementation

The impact

Results for the team

reduction in reminder preparation time
67%
reduction in reminder preparation time
Automated eligibility checks and message preparation are expected to reduce manual work for each reminder batch.
reminder coverage capacity
2.1×
reminder coverage capacity
Repeat-purchase signals could help the team cover more eligible customer journeys within its existing reminder workflow.
reduction in audience review time
53%
reduction in audience review time
Recent-order and availability checks are expected to reduce the time spent reviewing reminder audiences.

Situation

The next order rarely follows one fixed schedule.

Customers buy everyday pet products at different intervals and in different quantities. A single reminder schedule can arrive before a customer needs anything or after they have already ordered elsewhere.


Challenge

An estimated reorder date must stay an estimate.

Purchase history does not reveal exactly how much product remains at home. The agent needs to handle uncertain timing, changing pack sizes and new purchases, while avoiding unavailable items and repeated reminders.

What we built

Purchase patterns guide relevant reminders.

The agent examines repeat-purchase intervals and order quantities to suggest a reorder window for eligible everyday products. It keeps limited or inconsistent histories separate and uses conservative rules when there is too little evidence for a personalized estimate.

Before preparing a reminder, it checks recent purchases, product availability and contact permissions. Marketing approves message templates and timing rules; eligible reminders follow those rules, and a new purchase or opt-out suppresses the next send.

AI tech stack

  • Amazon BedrockAI inference
  1. Read purchase history

    Identify eligible everyday products and the customer’s previous orders and pack sizes.

  2. Estimate a reorder window

    Use repeat intervals to suggest timing, retaining uncertainty for sparse histories.

  3. Check current eligibility

    Check consent, recent purchases, product availability and reminder frequency.

  4. Prepare the reminder

    Draft a relevant reorder message using approved product facts and messaging.

  5. Apply approved send rules

    Queue eligible reminders within the timing and frequency limits approved by marketing.

  6. Learn from the response

    Record purchases, dismissals and opt-outs to review future reminder timing.

Evals and guardrails

Check the timing. Respect contact preferences.

  • Timing evals

    Compare suggested windows with later repeat orders. Test sparse histories, changes in quantity and purchases placed just before a reminder.

  • Eligibility evals

    Test unavailable products, consent changes and duplicate reminders. Verify that recent purchases suppress queued sends.

  • Guardrails

    Limit this workflow to eligible everyday products. Exclude prescription decisions and dosing advice; enforce approved contact and frequency rules.

Results

Less reminder preparation. More useful timing.

Automated eligibility checks and message preparation are expected to reduce manual work for each reminder batch. Repeat-purchase signals could help the team cover more eligible customer journeys within its existing reminder workflow.

Recent-order and availability checks are expected to reduce the time spent reviewing reminder audiences.

reduction in reminder preparation time
67%reduction in reminder preparation time
reminder coverage capacity
2.1×reminder coverage capacity
reduction in audience review time
53%reduction in audience review time

From Kohepets

“We have worked with Revensi for many years. In that time they have made vast improvements to how our business runs, and they are great people to work with. We highly recommend them.”
Kim HeongCo-Founder at Kohepets

Build capacity to grow. Own your intelligence.

Bring the workflow slowing your pipeline, campaigns or renewals. We'll map the bottleneck and show you the AI agent concept.