Release Features                                Date: June 30, 2021

This release introduces aiRA, user group-based loyalty, and customer status for CDP.

New Features

The following are the new releases.

1. AIRA models dashboard

2. User group-based loyalty: Build business engagements

3. Customer status: Make statuses more meaningful with custom labels

1. AIRA models dashboard

AIRA models dashboard is Capillary’s end-to-end observability platform to view, initiate, train, and manage Capillary’s machine learning propensity models.

Problem statement

Capillary has built several propensity models over the last two years in order to help clients to run personalized marketing campaigns. For example, Engage+ has a lapsation propensity filter to empower orgs to target customers who are going to leave in the next 30 days. This filter is driven by the lapsation propensity model and built by Capillary’s customer intelligence (CI) team. Similarly, there are other models and corresponding audience group filters for running campaigns based on offer propensity, transaction propensity, product propensity, campaign propensity, and other inferred customer behavior patterns.

  • Until now, there was no visibility into the list of models available for usage by an org. Customer success managers (CSMs) had to discuss with the product team or the customer intelligence team to understand available models and select the suitable model for their business objectives.
  • After model selection and initiation, there was a lot of effort involved to ensure sufficient data to train the model and get desired results. 
  • After training, when the model is in production, the customer intelligence team had to manually check accuracy parameters and take decisions on whether models need to be re-trained. All of these tasks led to productivity losses and delivery delays. 


Capillary’s AIRA models dashboard overcomes the above issues by automating the entire machine learning model lifecycle. It allows users (CSMs, Data Analysts, and Customer Intelligence team members) to view, initiate, validate, train, and manage the entire gamut of machine learning models built by the customer intelligence team in a single place. 

The following are few sample images.

Landing page

New model page

Model details page

Key Highlights

The following are the key highlights of AIRA.

  • Capillary’s machine learning operations (ML-Ops) platform gets models into production quickly and effortlessly.

  • The centralized dashboard helps preview and leverage Capillary’s out-of-the-box AI solutions available for an organization.

  • The entire machine learning model lifecycle is automated and handled on the platform - No more follow-ups through tickets and meetings.

  • Complete visibility on the status of models initiated – Identify and take action on models that are falling below accuracy thresholds.

  • Propensity models are available on Engage+ filters that are already available for initiation. There are more than 25 models in the pipeline for upcoming releases. 

2. User group-based loyalty: Build business engagements

For any brand to achieve the ultimate aim of benefiting its customers, it is important to motivate or encourage the contributors involved in this process. With the addition of this feature, the businesses also can launch B2B loyalty programs conveniently and systematically benefiting various levels of hierarchy.

Currently, the usability of Loyalty+ for any business or a brand is limited to benefit its end customers. The current addition facilitates orgs to create personalized loyalty programs not just to the end customer group, but also to the contributors in the supply chain. 

Use Cases

  • Fleet Loyalty: Fleet Drivers can earn points by purchasing fuel at retail outlets but points redemption can happen by fleet owners in the same group. 
  • Affiliate Marketing programs: The affiliate enrolled in the program can make customers buy products. With a product purchased by the customer, the rewards are accrued to the affiliate's account. 
  • Multi-Level Marketing Programs: A sales structure where existing distributors can recruit new distributors to sell the products. Distributors make money through a percentage of their recruits' sales and also direct sales of products to customers.
  • Corporate Loyalty Programs: Employees making transactions and rewards are awarded and redeemed at a company level.

About User Group Loyalty

Group Loyalty programs are customer retention solutions with custom group structures and loyalty mechanisms designed to help brands establish brand loyalty. These programs typically treat and incentivize individuals who are part of the same account (as a single group of users). The end-customers benefit from these programs because it allows them to have accelerated rewards as a group. For example, if a B2B account is a “Platinum” account then all members of the group enjoy better benefits, although to attain the “Platinum” account status the previous members might have done loyalty related activities to achieve that status. Orgs benefit from programs like this because it allows them to capture a larger share of mind by involving the community rather than individuals and drive collective behaviour in the desired manner.

Difference between B2B and B2C loyalty groups

Existing user groups are primarily aimed at B2C member groups. In the B2C model, with the help of the Engage+ platform, orgs are able to directly benefit the end customer, ignoring the various levels involved in the chain. 

B2C model has only one primary member with multiple secondary associated. Due to the lack of flexibility in the B2C model, several use cases such as accessing points earned lie mainly with the primary user. 

Whereas, in the B2B model the user can effectively incentivize various groups in the hierarchy. By studying the industry landscape and the B2C user groups requirements, we are targeting to make new user group entities that can cater to a wider variety of use cases, eventually fulfilling the B2B requirement.

Features of B2B User Groups

  1. Segmentation of groups in a hierarchical manner.
  2. Customizable hierarchy and user role depending on the org structure.
  3. Ability to have a user in multiple groups.
  4. Distinct loyalty actions depending on the user role.
  5. Hierarchy Definition to capture client company’s organisational structure.
  6. Ability to capture a customer’s association with other customers.
  7. Auto creation of groups and addition of members based on user roles.
  8. Loyalty permissions in a group based on user role.
  9. Granular control of rewards to the group and individual accounts.


The following are the new terminologies of user group loyalty.

  • Client: A business entity that is enrolled in the org’s User Group Loyalty. A client can have sub-clients under it (smaller business entities).
  • Role: Part (function) played by an individual associated with the business. For example, a company enrolled in a fleet loyalty program can have various user roles such as fleet driver, fleet owner and group fleet owner.
  • Customer Hierarchy: It is the ranking of members of org/clients according to the relative status, or position. Hierarchy helps define the association between two customer roles. In the previous example, a fleet driver can report to a fleet owner, whereas a fleet owner can report to a group fleet owner. In this, the customer hierarchy captures each role and their reporting role.
  • Hierarchy definition stores attribute required to create a hierarchy (such as customer roles), validation rules for each role, group automation and role-based loyalty permissions.
    • You can define multiple hierarchy definitions for an org and the default hierarchy definition.
    • Each client can have only 1 hierarchy. Once the hierarchy is set up, all the attributes and configurations created in the hierarchy will apply to the customers associated with the company.
  • Validation rules: Used to define the parent-child relationship with other roles. You can limit the max. no. of child roles for each role. Configs for skip-level parent-child mapping available.
  • Group Automations: Groups can be implicitly created when a user is associated with a specific role. The user becomes the owner of the group and other users can be added to the group based on the association with the owner.
  • Group Permissions: Allows setting up Loyalty permissions such as points redemption and points transfer for each member of the group based on the role. 

Front end APIs, see User Group Loyalty APIs

3. Customer status: Make statuses more meaningful with custom labels

As a starting point for brands to develop and launch impactful marketing campaigns it is really important to categorize their customers according to their engagement with the brand. You can now create different status labels against each standard status making it more meaningful. An active customer is not just an antonym to inactive, you can have meaningful labels that make sense to your brand like registered, transacted, regular. You can restrict specific status from redeeming points or coupons.