Confluent says resale alone lacks the volume and margins for partner profitability

As real-time data adoption deepens in BFSI and digital‑native firms, the company outlines why partners must combine resale with services to build sustainable practices.

As Indian enterprises push real-time data into core production systems, Confluent is urging partners to rethink their business models, warning that resale alone no longer delivers the scale or margins required in today’s environment.

The shift is being driven by rising complexity in production deployments, particularly in regulated sectors, where customers now expect partners to take ownership of architecture, execution and ongoing operations, not just transactions.

Speaking to CRN India, Confluent’s senior vice president and general manager for APAC, Gregory Taylor, said, “The pure resale model is not sufficient, there isn’t enough volume, and margins are limited. Partners need to think about this holistically by combining resale with implementation services, managed services, and advisory offerings. That is where profitability comes from.”

“What we expect from partners is to engage deeply with customers, build capabilities, work closely on implementations, and over time, develop managed services that allow them to support and manage customer environments continuously.”

Taylor added, “We support partners in this journey through training and enablement, helping them build and scale their Confluent practice.”

Partner guidance becomes even more critical in India - Confluent’s number one APAC market, according to Taylor.

The market’s scale, depth of engineering talent and growing maturity in data‑driven architectures are accelerating production‑grade adoption, raising expectations from customers and increasing the stakes for partners delivering real‑time data platforms.

BFSI pushes real‑time data from pilots into production

Among Indian enterprises, Confluent sees BFSI and digital‑native companies converting fastest into production environments.

While digital‑native firms typically have the engineering maturity to adopt new architectures quickly, BFSI is emerging as the most partner‑dependent segment due to a combination of regulatory demands, competitive pressure and operational complexity.

Confluent’s area vice president for India and emerging markets, Rubal Sahni, said, “In BFSI, traditional financial institutions are being disrupted by fintech players. As a result, they need to build an environment where they can collaborate and compete by opening up their systems to partners while maintaining compliance, security, and governance.”

This includes use cases, including co-branded credit cards, where customer acquisition may happen through a fintech, while processing is handled by a bank.

Sahni said core banking systems are also under pressure as multiple microservices attempt to access the same data simultaneously for customer 360 programmes, real‑time personalisation and risk management.

Confluent’s approach allows banks to offload these workloads from core systems, keeping operational platforms stable while enabling real‑time processing elsewhere, he added.

As a result, banks are no longer running isolated experiments.

Sahni said BFSI customers are operating 30 to 40 real‑time use cases in parallel, typically delivered with support from a broad ecosystem of global system integrators, Indian SIs and specialist boutique partners, depending on the bank’s environment and compliance constraints.

Taylor said the surge in production adoption aligns with a broader architectural shift.

Enterprises that successfully move AI and real‑time analytics into production are increasingly adopting “shift‑left” architectures that process data closer to the source, rather than relying on delayed batch pipelines.

This shift, he said, is allowing use cases to expand beyond early retrieval‑augmented generation (RAG) deployments into areas including anti‑money laundering, customer 360 platforms and real‑time decisioning, where milliseconds matter.

Regulation, skills and legacy systems shape the partner opportunity

While demand for real‑time data platforms is rising, Taylor said the primary blockers to adoption in India vary sharply by segment, creating both constraints and opportunities for partners.

Digital‑native companies, he said, are typically cloud‑first and face fewer structural barriers. However, even these organisations require guidance on integrating complex technologies.

To address this, Taylor said, “We provide services, either directly or through partners, to help with initial architecture and implementation. Once that foundation is in place, it becomes easy to move forward quickly given their strong engineering capabilities.”

In BFSI, the challenges are more pronounced. Regulatory requirements around data residency and compliance limit how quickly banks can move fully to the cloud, forcing many to adopt hybrid or on‑premises models.

Skills shortages in regulated environments compound the issue, increasing reliance on partners that understand both modern data architectures and local compliance frameworks.

One of Confluent’s strengths, Taylor said, is the ability to meet customers where they are - supporting on‑premises, cloud and hybrid deployments and allowing banks to transition over time.

Sahni added that technical debt arising from legacy systems remains a persistent challenge across Indian enterprises.

In such environments, change management becomes as important as technology selection. Organisations that proactively modernise legacy systems and manage this transition carefully are better positioned to scale real‑time data architectures into production.

Where partners go wrong on streaming projects

Despite strong technical talent in India, Confluent sees recurring execution problems in partner‑led streaming deployments.

Taylor said, “From a business perspective, one of the most common mistakes partners make is being overconfident in their ability to deliver scalable, high-quality solutions.”

Partners assume that experience with open‑source Kafka or general cloud infrastructure is sufficient to deliver scalable, production‑grade environments.

This often leads to flawed initial architectures and sub‑optimal configurations, particularly in cloud deployments.

When systems are not optimised correctly from day one, Taylor said, customers experience inefficiencies and excessive consumption, eroding the economic benefits of real‑time architectures.

In such cases, Confluent is often required to step in later to re‑architect deployments, an avoidable outcome had the platform been designed correctly upfront.

Sahni said another frequent error is a narrow focus on short‑term integration needs.

“While customers may begin with a specific streaming use case, the platform should be designed to evolve into a central nervous system for the organisation, supporting multiple teams and functions over time,” he said.

Failing to plan for scale, future use cases and platform evolution limits long‑term value.

Architecture choices now determine cost, margins and AI outcomes

As enterprises expand real-time data usage, architectural decisions are increasingly tied to economic outcomes for both customers and partners.

Taylor said organisations across APAC, including India, are replacing self-managed open-source Kafka, legacy messaging systems and hyperscaler-managed services as they reassess cost, efficiency and operational complexity.

A key driver is the shift towards “shift-left” architectures, that is, processing data closer to the source and deciding early what needs long-term storage, rather than pushing everything into data lakes or lakehouses.

This reduces storage and compute costs while improving latency.

Traditional batch pipelines that move data through multiple layers are also being reconsidered. Processing data earlier and writing directly into high-value layers reduces infrastructure overhead and complexity.

In India, Taylor said infrastructure costs, particularly storage and compute on hyperscalers, are becoming more visible, even as engineering talent remains relatively accessible. However, organisations do not want skilled teams tied up in managing infrastructure.

Partners that can optimise architecture and improve engineering productivity deliver measurable value.

Sahni said Confluent’s platform supports different cost and performance requirements, allowing customers to use premium clusters for mission-critical workloads and more cost-effective options for internal use cases.

Multi-cloud support across AWS, Azure and GCP also gives enterprises flexibility and control over core data.

These architectural decisions become critical as AI moves into production. Taylor said AI initiatives succeed only when supported by real-time, governed data infrastructure.

Sahni added that AI projects often fail because models produce inaccurate outputs or become uneconomical due to uncontrolled token consumption.

To address this, enterprises need access to fresh, contextual, real-time data at scale, combining historical and real-time inputs across internal and external sources.

Equally important is governance. By setting guardrails on how AI systems access data and make decisions, organisations can control token usage, improve accuracy and ensure economic viability.

Together, these capabilities allow enterprises to move AI from experimentation into production at scale.