← Experience

Client implementation / Pega Customer Decision Hub

One decisioning foundation.
Two connected channels.

An implemented next-best-action experience connecting assisted call-center conversations and outbound email to Pega CDH. Both channels delivered relevant actions and returned interaction outcomes to the decision engine.

Customer contextRank & selectDeliver in-channel Outcomes shape future decisions

This describes a real client implementation. Client and environment identifiers are omitted.

01 / Shared foundation

Configure once.
Make decisions available in context.

Decision catalogueIssues → Groups → Actions → Treatments

Taxonomy, action content and channel treatments provide the structure for the decisioning experience.

Decision serviceEligibility → Applicability → Suitability → Contact policy → Arbitration

Configured strategies evaluate candidate actions against customer and business context, then return ranked next-best actions for a channel.

Integration patternRequest or batch → Deliver → Capture response

Real-time and outbound integrations used the same CDH decisioning foundation while recording channel-specific outcomes.

02 / Customer experience principles

One customer strategy
across the lifecycle.

Decisioning connects customer relevance with business priorities, then makes that choice usable in each channel.

Customer strategy and lifecycle marketing provide a shared foundation for these principles.

Improved personalisation

Use customer context and interaction history to make an offer more relevant, and keep its message appropriate to the touchpoint.

Increased collaboration

Give marketing, product and channel teams a shared action catalogue and decision framework to shape joined-up customer journeys.

Better decisions

Balance predicted response, current context and business value; connect decision outcomes to the measures used to refine future choices.

A maturity lens

From disconnected campaigns toward more consistent, individually relevant offers.

  1. 01DisconnectedSiloed communication
  2. 02PersonalisedTargeted segments
  3. 03ConsistentShared decisions across touchpoints
  4. 04OptimisedOffers adapt to customer and business signalsSelected maturity level

03 / Before delivery

What goes into
a next-best-action decision?

Each channel request draws on the same decision foundation. Policies narrow the candidate set; arbitration ranks what remains.

01
Customer Action Library

Define the actions and treatments

Organise offers under the business taxonomy—such as sales, acquisition, retention and mandatory actions. Each action can carry treatments that provide the message and content for a particular channel.

Issues→Groups→Actions→Treatments
02
Engagement policies

Check whether an action fits

Configured policy conditions assess the customer–action match before an offer is considered for delivery.

EligibilityDoes the customer qualify?
ApplicabilityIs the action relevant now?
SuitabilityIs it appropriate for this customer?
03
Constraints

Respect contact rules and channel limits

Contact policies use recorded exposure and response history to suppress over-contacted actions. Channel limits add delivery-specific controls, especially for outbound communication.

Prior outcomes can affect which actions are eligible to surface again.
04
Arbitration

Rank the remaining choices

Arbitration balances the predicted likelihood of customer response with contextual relevance and business priorities.

Propensity to accept×Value×Business levers×Context
A conceptual view of the factors shown for this solution; configured strategies can extend or adjust the model.
05
Next Best Action

Return an action ready for the channel

CDH returns ranked action data. The integration selects a suitable treatment and presents the offer in the channel context—for example, an offer tile for an agent or content in an outbound email.

The right message · the right treatment · the preferred channel

04 / Assisted call-center channel

Bring a relevant offer
into the live conversation.

The interaction portal requested offers through the Real-Time Container, rendered the returned action data, and captured presentation and agent feedback.

A
Real-time request

Agent-assisted next-best actions

Interaction Portal · RTC · CDH
UML sequence: customer call, agent opens profile, portal requests Call_Center_Offers through the Real-Time Container, CDH returns ranked action data, the agent discusses an offer, and the portal captures an impression or a rejection reason.
Call-center offer retrieval, presentation and agent-assisted feedback capture.
Read the call-center sequence in text
  1. 01Customer / Agent
    Start the interaction

    The customer calls; the agent opens the customer profile in the interaction portal.

    Customer context
  2. 02Interaction Portal
    Request offers for the call-center container

    The portal requests next-best actions for Call_Center_Offers through the Real-Time Container.

    NBA request
  3. 03RTC ↔ CDH
    Fetch and return ranked action data

    The container asks CDH to evaluate the customer and returns action data such as title, price and rank.

    Decision response
  4. 04RTC / Portal
    Map and render offer tiles

    The container maps the response to the offer-card template; the portal displays personalized tiles to the agent.

    Presentation
  5. 05CDH / Portal
    Record presentation and impression

    Presentation was implicitly recorded. When the agent selected a tile to discuss, the portal sent an Impression outcome through the Capture_Response API.

    Presentation · Impression
  6. 06Agent / Portal / CDH
    Capture a rejection reason

    If the customer declined, the agent selected Rejected. The portal opened a reason list; choosing “Expensive” sent a Rejected outcome with subtype Expensive to CDH.

    Rejected · Expensive

05 / Outbound email channel

Carry the decision into
the customer’s inbox.

A batch process generated NBA email records for the ESP. Delivery, open, click and negative feedback events were reported back to CDH.

B
Batch delivery + response capture

Outbound next-best-action email

Pega Batch · ESP · CDH
UML sequence: Pega batch creates and sends NBA email records through an email provider, then Capture_Response records Sent, Opened/Impression and Clicked events, plus unsubscribe or low-rating rejection feedback.
Outbound generation, email engagement and negative feedback capture.
Read the email sequence in text
  1. 01Pega Batch Engine
    Generate email NBA records

    The batch engine generated the customer list and email content; Pending was implicitly recorded for the outbound action.

    Pending
  2. 02Batch Engine → ESP
    Send the batch file

    The customer list and content were passed to the third-party email service provider.

    Outbound file
  3. 03ESP → CDH
    Confirm the email left the provider

    The ESP sent a Sent outcome through Capture_Response, confirming the email had successfully left the server.

    Sent
  4. 04Customer → ESP / CDH
    Capture open and click activity

    Opening the email produced an Opened / Impression outcome. Following its call-to-action link produced a Clicked outcome; both were sent through Capture_Response.

    Opened · Clicked
  5. 05Customer → ESP / CDH
    Capture negative feedback

    Unsubscribing produced Rejected with subtype Unsubscribe. Selecting a low email rating (for example, two stars) loaded a feedback confirmation page and recorded Rejected with subtype Two_stars.

    Rejected · reason

06 / Close the loop

Make channel outcomes
useful to future decisions.

Reliable response capture

Linking outcomes to the decision and action gives CDH a usable record of what happened after delivery.

Customer history & contact policies

Recorded interaction history can inform future context and contact-policy suppression according to the configured solution.

Adaptive learning

Responses can train adaptive models when the relevant models and response mappings are configured. Updated predictions can then contribute to future arbitration.

The channel events above describe the implemented integrations. How captured responses affect contact policies and adaptive models depends on the corresponding CDH configuration.

Experience

Thoughtful decisioning.
Built into real channels.

Discuss a decisioning challenge