Adoption accelerates when AI is trusted in production: Anaptyss CDO

Enterprise operations are adding automation faster than they are rebuilding the execution systems underneath, leaving manual steps, fragile handoffs, and data gaps. Anaptyss works across BPO, knowledge services, and technology delivery, and frames its Digital Knowledge Operations (DKO) approach as the link between domain expertise, digital tools, and run-the-business execution. 
In a conversation with TechCircle, Rohit Gore, Chief Digital Officer at Anaptyss, discussed why the industry is shifting from improving user experience to fixing core execution, and how AI adoption is constrained by data quality and governance. Edited Excerpts: 

Banks have invested heavily in digital transformation, yet many core processes remain manual. Where do these efforts usually break down—architecture, data, operating model, or governance?

The core business problems in banking have not changed much. Processes are old, and risk is embedded in almost every function. Banks operate under shifting regulation, changing customer expectations, and a volatile geopolitical and political backdrop. At the same time, they have to respond quickly to current issues and still build something that can last five to 10 years. Cyber threats add pressure.
From my perspective, CIOs and CTOs tend to manage this through three tracks. The first track is tactical. Tools like RPA and earlier, limited AI were used to solve clear process issues and deliver fast cost reduction within the existing process landscape.

The second track is short- to mid-term change that is forward-looking but not deeply transformative. This includes business process management combined with OCR, voice and speech recognition, and chatbots. These efforts are often triggered by immediate operational pressures, including security threats, and are aimed at giving business users and customers a way to adjust processes and deploy changes quickly. This approach remains common because CIOs, CTOs, and CFOs value speed, accuracy improvements, and the ability to adapt over a two- to three-year horizon.

The third track focuses on longer-term problems and treats technology as a way to change business models, not only individual processes. A current example is regulatory complexity. AI-driven platforms are being used across areas such as AML, KYC, fraud prevention, model risk management, and cybersecurity. Fintech adoption has pushed this direction earlier than some traditional players, and banks are now pursuing it more actively.

Your company offers an AI-powered framework called DKO. Is it an orchestration layer, a decisioning framework, or something else—and where does it sit in the client stack?

DKO stands for Digital Knowledge Operations. There are three ways to understand it. First, it is a philosophy that guides how we engage with clients to help them meet KPIs and goals. In practical terms, it is a set of principles, including a “digital-first” approach. For example, if transaction volumes spike, one response is to add people to manage peaks. The DKO approach is to start with an AI-led, digital-first option instead of defaulting to older approaches.

Second, it is a set of capabilities that connects work across BPO, KPO, and technology. The “knowledge” comes from people and from systems and processes developed over time. That knowledge is applied to digital capabilities and to running operations. The intent is to avoid point solutions and instead combine digital and operations into a single approach.
Third, it shapes execution and delivery. It influences areas such as HR policies, training and enablement, operational and technology methodologies, audits, and client feedback mechanisms. Together, these elements form the Digital Knowledge Operations approach, which the company describes as trademarked.

Everyone talks about AI in banking, but data remains inconsistent and poorly governed. How much of your work is AI-driven versus foundational data engineering?

The rate of change in AI is faster than what I’ve seen in my career. The industry’s center of gravity shifts every two or three months, with new models and frameworks emerging from major technology providers and displacing what came before. The challenge is staying current in a way that’s practical. Internally, I push for teams to explain not only what a technology is, but how to evaluate it and apply it in a working environment.

On adoption, there is public reporting and analysis, including from large consulting firms, that AI adoption could be further along and that some projects have not delivered what was expected. One constraint is data quality. There is also a ceiling on how much data is available. Synthetic data is discussed as a path forward, but it depends on whether it can reach a level that is useful beyond production data. For now, the issue is reaching the ceiling on both data quality and quantity.

Data is also different from other resources because the same datasets can yield new information when better modelling techniques arrive. The industry is improving at extracting value from existing datasets.

Where is this technology shift heading in the near term, beyond the noise around AI and composable banking, and how will it reshape financial operations?

The most visible progress so far has been on the experience layer. I often refer to the idea that “AI is the new UI,” in the sense that interaction models are changing. Enterprise users now expect the same kind of intuitive, consumer-style experiences in business applications, and that expectation is driving the adoption of more AI-enabled interfaces.

Where the focus is shifting now is the execution layer: core processes, core commercial banking applications, and systems that manage sensitive customer data. Front-end experiences have improved faster than the internal processes that support them, creating an imbalance where expectations are high but back-end execution still breaks under pressure.

I see organisations moving from experience-led improvements toward strengthening execution. There is a growing number of proofs of concept and point solutions being developed and deployed to address specific problems. Once these solutions are orchestrated effectively, they can form a foundation that consistently supports customer-facing experiences with stronger core-process performance. Based on what I’ve seen in industry reports and discussions with CIOs, I expect meaningful progress on this execution layer within the next 18 months, and possibly closer to 12 months.