How CRED is tapping AI to deliver premium customer experiences

TL;DR: India’s CRED, serving 15 million affluent members, has built three AI tools using OpenAI models (GPT-4o, GPT-5, o3) to deliver concierge-level customer experiences. Their AI companion Cleo achieves 98% resolution accuracy, whilst internal tools Thea and Stark support agents and operations teams, resulting in a 14 percentage point CSAT improvement and 31% reduction in session drop-offs.

Building AI-first for India’s most discerning users

CRED, India’s members-only club for creditworthy individuals, has built its brand since 2018 on delivering exceptional digital experiences to the country’s most affluent consumers. With over 15 million monthly active users, the company faced a critical challenge: how to maintain premium, concierge-like service quality whilst scaling rapidly across multiple products and teams.

Swamy Seetharaman of CRED describes the company’s transformation into an AI-first organisation, centred on a core question: “How do we make every single member in every function be 10X?” For CRED, artificial intelligence has become the primary unlock in that journey, enabling the company to “move fast and stay right, without sacrificing our core principles” of trust, transparency, security, reliability, and exceptional design.

Three AI tools for comprehensive service transformation

CRED’s AI implementation centres on three distinct tools, each addressing different aspects of the customer experience:

Cleo: AI conversational companion

Powered by OpenAI models including GPT-4o, GPT-5, and o3, Cleo handles three common query types:

• Informational queries: General product questions like “What is CRED Cash?”
• Contextual queries: User-specific questions such as “Am I eligible for CRED Cash?”
• Transactional queries: Action-oriented requests like refund processing

The system diagnoses issues, classifies intent, maps queries to the appropriate standard operating procedure, and frames contextual, accurate responses tailored to each member’s situation.

Thea: Agent support system

Designed for customer support agents, Thea summarises multi-format conversations across text, voice, and Hinglish (Hindi-English code-switching common in India), suggesting next steps and enabling more efficient, empathetic customer interactions.

Stark: Operations automation

Built for operations teams, Stark accelerates SOP creation and updates, reducing what previously took days into a minutes-long process. This enables faster adaptation to new products, policies, and customer needs.

Measurable impact: 98% accuracy and substantial CSAT gains

Three months after launch, CRED’s AI implementation has delivered significant measurable improvements:

• 98% resolution accuracy rate for Cleo across all query types
• 14 percentage point improvement in customer satisfaction (CSAT) scores
• 18% increase in successful multi-intent conversation resolution
• 31% reduction in session drop-offs
• Decreased average handling times across all three AI tools

Seetharaman notes that whilst these are early results, they’re “incredibly encouraging” and demonstrate progress towards CRED’s goal of creating “a true concierge experience built on trust, reliability, security, and exceptional design.”

Overcoming scepticism through measurable results

The development journey revealed an interesting organisational dynamic. Initial scepticism about AI capabilities existed within teams, but confidence grew rapidly once members experienced results through CRED’s internal evaluation framework (which also uses OpenAI models).

“The biggest surprise has been how fast people adapt once they experience real unlocks,” Seetharaman reflects. “They see that they can be exponentially more effective and efficient with AI.”

Looking forward: Real-time knowledge base enhancement

CRED’s next phase focuses on expanding Cleo across all business lines whilst building systems that detect “data dead-ends”—cases where user queries cannot be answered—and feed them back into the knowledge base to improve SOPs in real time.

The broader organisational goal extends beyond customer service: “Our broader goal is for every team member, across every function like engineering, QA, infra, compliance to become 10x more efficient,” Seetharaman explains.

He identifies the critical success factors in the evolving AI ecosystem: “Success will depend on how fast and how right we can act, identifying the right insights from noise and making accurate decisions at speed.”

Implementation advice: Align AI with core values

For companies exploring AI adoption but remaining hesitant, Seetharaman offers pragmatic guidance: “Every company must identify what’s most important whether it’s efficiency, effectiveness, or both, and then leverage AI in alignment with their values.”

For CRED, incorporating OpenAI’s technology has been “a true unlock” across two of the company’s core values: compounding growth and being “fast and right.” The early results validate this approach, with focus now on amplifying and scaling the impact across the entire organisation.

Source: OpenAI Original Article: How CRED is tapping AI to deliver premium customer experiences Published: 5 November 2025