Origin Energy
AI-native customer operations platform + genAI drafting assistance
Origin Energy migrated 4.2 million Australian customers to Kraken, the Octopus group's AI-native platform including the genAI tool Magic Ink, for $170 million saved in cost-to-serve and an 87% satisfaction index.
Key points
- Migration of 4.2 million customers to an AI-native platform.
- Kraken platform with the genAI tool Magic Ink to summarize and draft.
- $170 million saved in cost-to-serve, 87% satisfaction.
- Evidence B, confirmed status and a platform in production.
Objective
Reduce the cost-to-serve and gain agility by replacing legacy systems with an AI-native platform, while improving the customer experience across more than 4 million accounts.
The deployment
Origin Energy migrated its residential and business customer base to Kraken, the Octopus group's AI-native operations platform. The switch covered 4.2 million customers over two and a half years. The platform integrates Magic Ink, the genAI tool that summarizes interactions and drafts replies for Origin's agents. Origin highlights a cost-to-serve reduced by $170 million, a Customer Happiness Index of 87 percent, and a Trustpilot rating raised to Excellent. The platform's speed is cited as a factor: a change that took three to four months on the old system ships in two to three weeks on Kraken.
Results Proof B
Quantified platform case study (Kraken) with migration volume, cost-to-serve, and satisfaction, corroborated by the trade press (iTnews) reproducing the savings announced in Origin's half-year results. The Magic Ink genAI metrics are group-level, not isolated for Origin.
How it works
Documented architectureThe stack in detail
- plateforme Kraken The Octopus group's AI-native customer operations platform (account, billing, interactions), to which Origin migrated 4.2 million customers.
- outil Magic Ink GenAI tool integrated into Kraken: it summarizes interactions and drafts replies for Origin's agents.
- llm LLM type GPT (modele exact non publie) GPT-type text-generation models used by Magic Ink; version and provider not published.
How it runs, concretely
For ops teams-
1Account migration data team
Origin transfers its customers from the old system to Kraken, in successive waves.
-
2Drafting assistance AI
Magic Ink summarizes interactions and proposes draft replies to agents.
-
3Customer handling customer service
The agent validates and replies; routine operations move to self-service.
-
4Product iteration agency
Journey changes are shipped in a few weeks rather than several months.
The cost-to-serve per account and the satisfaction index. The platform relies on a unified customer history; without that view, the genAI assistance loses its value.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- unified customer history on a single platform
- billing and contract data accessible in real time
- corpus of interactions for the genAI assistance
Org prerequisites
- willingness to replace legacy systems
- wave-based migration program
- customer service trained in AI assistance
Possible stack
- AI-native energy operations platform (Kraken)
- enterprise LLM for drafting assistance
- integrated CRM/billing
The plan, step by step
- Step 1Audit the legacy systems (CRM, billing, channels) and frame the migration: scope, account waves, success criteria.Deliverable: Wave-based migration plan with milestones
- Step 2Migrate a first wave of accounts to the target platform and stabilize operations on that cohort.Deliverable: Pilot cohort operated entirely on the new platform
- Step 3Chain the following waves and train customer service to supervise the genAI assistance (review, correct, send).Deliverable: Majority of the base migrated, agents trained
- Step 4Turn on drafting assistance and self-service across the whole base, with measurement of satisfaction and volume handled per agent.Deliverable: GenAI drafts in production, satisfaction dashboard
- Step 5Decommission the legacy systems and quantify the cost-to-serve per account against the pre-migration baseline.Deliverable: Documented economic assessment
First step: Unify the customer history on a single platform before turning on genAI drafting assistance.
The same pattern, proven in other industries
ADT
genAI customer agent
Air India
genAI customer agent (automated customer service)
AT&T
Governed enterprise genAI assistant: an LLM connected to internal knowledge, open to tens of thousands of employees and called at scale by automated systems
Sources
- S1 Case Study: Origin - Kraken Interested party archive pending
- S2 Origin Energy says Kraken platform will save $70m to $80m this year Established press archive pending
- S3 AI Adoption Case Study: Kraken's generative AI tool for customer service (Magic Ink) Interested party archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.