At Keppel, AI is a force multiplier and a strategic enabler that is changing how Keppel operates and makes decisions as an asset manager and operator. From accelerating deal evaluations to predicting operational risks, Keppel is developing and deploying AI in practical, high-impact ways that deliver real-world results, enable employees to do more, sharpen decision making and accelerate processes.

Real-world impact through practical use cases

Keppel’s AI strategy is built around three core goals: achieving efficiency gains, enabling smarter decision-making, and enhancing operational capabilities. Going beyond strategic aspirations, these goals are grounded in tangible use cases that align with Keppel’s dual role as both an asset manager and operator.

Tasks that once took hours now take minutes. Deal evaluations that previously required teams to take months are now completed in days. Research on companies and sectors has accelerated by 10 to 20 times. AI helps Keppel rapidly analyse large datasets, streamlining due diligence and competitive analysis while improving visibility into operational risks.

Smarter deal evaluation and research with Alpha and Duet

Investment teams can interact with Alpha through smart questions and answers, instantly retrieving insights from deal documents and market research.

Keppel launched two generative AI applications: Alpha and Duet, about 18 months ago to focus on deal evaluation and research respectively. Since then, the tools have analysed over 50,000 companies and answered more than 17,000 questions.

Alpha accelerates investment deal evaluation by up to four times. It processes a wide range of document formats—PDFs, spreadsheets, presentations, and notes—building a comprehensive knowledge base of transactions. By integrating with Duet, Alpha gains access to deep market analysis, sector insights, and company-level research, enabling more informed and timely evaluations by the investment teams.

Investment teams can interact with Alpha through smart questions and answers, instantly retrieving insights from deal documents and market research. Alpha can also auto-generate draft investment presentations, giving the investment teams a head start.

Unlike many in the industry, Keppel goes beyond mainstream sources of information, by layering on its own proprietary operating data and alternative datasets which are continuously produced through AI and cloud computing. This expanded data surface can help uncover high-value insights.

Building new operational capabilities with KAI, Athena and Agent9D

Keppel’s AI innovation goes beyond speed—it is about doing things differently. The company has developed KAI, a proprietary AI operating system that enhances workflows, collaboration and decision-making.

KAI integrates multiple AI models—including ChatGPT, Claude and Gemini—to improve research and avoid blind spots. It supports mobile-first workflows and voice-enabled collaboration. With thousands of modular AI engine configurations, KAI offers flexibility and scalability for developers.

KAI enables the creation of AI agents that think like experts, trained on proprietary data and subject matter expertise. These agents evaluate opportunities using both qualitative and quantitative reasoning, offer real-time recommendations, and collaborate like digital co-workers. KAI can enhance processes, such as Know-Your-Customer and client onboarding. KAI is currently in beta, which means that it is being tested by a select group of users for intensive feedback and refinements, and will be widely rolled out by end of 2025.

In Keppel’s Connectivity Division, the Athena system, another AI solution, is already transforming data centre operations. Currently deployed in one facility with more rollouts underway, Athena acts as a “black swan detector”, predicting rare but high-impact failures like cooling system breakdowns before they occur. Using signal processing and artificial intelligence—covering vibration, sound, and higher order harmonics—Athena identifies early warning signs and highlights them to data centre operators, helping extend maintenance cycles without increasing risk, reducing costs while enhancing reliability.

Deployed in existing and new facilities, Athena acts as a “black swan detector,” predicting rare but high-impact failures like cooling system breakdowns before they occur.

Another innovation, Agent9D, is designed to work alongside project managers to improve complex project execution. It identifies chokepoints and potential delays in different project workstreams before they escalate, and suggests preemptive action. Agent9D also optimises resources in real time, adapting to changes during project execution. The targeted results: up to 30% improvement in timely project delivery and a 50% reduction in site changes—translating to real savings in time and cost. Agent9D is currently under development.

Agent9D identifies chokepoints and potential delays in different project workstreams before they escalate.

Keppel’s AI strategy builds capabilities that matter by embedding intelligence into its core operations and is redefining what it means to be a smart asset manager and operator in the digital age.