CASE STUDY
Southeast Asian CSP cuts CPQ order time by over 80% with the Totogi Ontology
Key Facts
- 01 Customer
-
Multi-play CSP
Southeast Asia - 02 Product
-
Totogi Ontology
- 03 Use case
-
CPQ order automation
- 04 Systems
-
CloudSense CPQ
Salesforce - 05 Time frame
-
4 weeks
In brief
A major multi-play communications service provider in Southeast Asia, serving millions of mobile, broadband, and TV customers, ran its sales operation on CloudSense CPQ and Salesforce. Creating one order took more than 50 clicks and over 5 minutes, and only a few specialists could handle it.
Working with Totogi, the operator cut CPQ order creation time by more than 80%. The Totogi Ontology integrated CloudSense CPQ and Salesforce, and an embedded AI agent let any rep create a full, validated order from one prompt in under 50 seconds. Totogi deployed it in four weeks, no rip-and-replace.
Impact
Over 80% reduction in CPQ order creation time
<50 sec to create a full, validated order, down from over 5 minutes
One prompt replaces the 50-plus clicks each order once required
4 weeks from kickoff to a working AI solution in production
The Totogi Ontology dramatically simplified order creation, cut processing time, and reduced reliance on specialized knowledge.
Challenge
The operator already ran CloudSense CPQ and Salesforce to manage complex product catalogs and sales workflows. Even so, the day-to-day order creation journey was slow and error-prone. Creating a single new order took more than 50 clicks and over 5 minutes in a best-case scenario.
Validation and error handling were only partial, so orders failed often and needed rework. The process was also hard enough that only a handful of trained specialists could navigate it, which created bottlenecks and slowed the whole sales cycle.
These inefficiencies delayed sales operations and capped how far the operator could scale order processing. It needed a faster, simpler way to create orders without giving up accuracy, and a foundation it could build future AI use cases on.
Solution
The operator wanted to automate order creation inside its existing systems, with no full rip-and-replace, and to build a foundation for future AI work rather than a one-off fix.
Totogi started with a telco-specific ontology layer that integrated both CloudSense CPQ and Salesforce. The layer understands telecom data structures and the relationships between them, including standard product hierarchies, commercial constructs, and configuration logic. The Totogi Ontology was then trained on the operator’s own product catalog, business rules, and validation requirements.
On top of that layer, Totogi built an AI agent for single-prompt order generation, embedded in the Salesforce workspace. A rep now describes the order in plain language, and the AI agent creates the full validated order in seconds, filling in missing details, enforcing business rules, and executing it across Salesforce and CloudSense. Totogi delivered this in four weeks, where legacy vendors need months of customization.
Business
impact
The result was a reduction of more than 80% in CPQ order creation time. What once took over 50 clicks and more than 5 minutes, in a best case, now takes one prompt and under 50 seconds. Automated validation and intelligent field completion replaced manual data entry and the errors that came with it.
Order creation is also no longer the preserve of a few specialists. Any rep can now generate a complex, multi-step order, so the operator can scale order processing without adding headcount, and the team spends more time selling and less on rework.
The engagement also left the operator with a reusable AI framework. The same ontology layer and AI approach can power the next wave of use cases across sales operations, turning one order-automation win into a repeatable capability.
About the customer
The customer is a major multi-play communications service provider in Southeast Asia, serving millions of mobile, broadband, and TV customers. It runs a complex product catalog and sales operation on CloudSense CPQ and Salesforce. The operator is anonymized in this case study.