At the International Conference of Bharat Research in AI and NextGen (BRAINCON) 2026, held at Management Development Institute of Singapore (MDIS), Kevin Shepherdson, CEO and Founder of Straits Interactive, took to the stage. Under the topic, ‘Reasoning and Frugal AI’, he reframed how we should think about the next phase of Artificial Intelligence.
His central message was clear: The future of AI will not be about bigger models – but better thinking.
The Hidden Problem: AI Is Already Shaping How We Think
Today’s Large Language Models (LLMs) don’t just give us answers — they shape how we think.
Most LLMs operate on probability-driven outputs, pattern matching from training data, and narrow reasoning pathways influenced by fine-tuning.
This creates an invisible constraint. When the AI decides the reasoning path, the user simply has to receive the output. As such, alternative reasoning approaches are rarely explored.
In other words, while AI appears intelligent, it often reflects a single dominant reasoning trajectory, limiting critical thinking and exploration.
Introducing Reasoning-Aware AI: A New Paradigm
To address this, Kevin shared insights from Straits Interactive’s provisional patent on Reasoning-Aware AI, a framework designed to restore control of reasoning back to the user and the system designer.
At its core, Reasoning-Aware AI introduces two key innovations:
1. Auto-Reasoning Logic (Patent-Pending)
Instead of relying on the LLM’s default reasoning, the system automatically structures the thinking process and breaks down problems into step-by-step reasoning flows. It then applies guided logic paths before generating answers.
This transforms AI from a black-box predictor into a structured reasoning engine.
The impact is significant: improved accuracy, greater explainability, and more consistent outputs. In fact, internal benchmarks showed ~14% performance improvement when Straits Interactive’s auto-reasoning was applied.
2. Choice of Reasoning Methods (User-Controlled Thinking)
The conference also raised the question: what if users could choose how the AI reasons?
Instead of a single hidden reasoning process, Reasoning-Aware AI allows the following:
1. Selection of different reasoning styles (such as logical, step-by-step, comparative, evaluative)
2. Customisation of thinking frameworks based on the task
3. Alignment of reasoning with domain needs (such as legal, compliance, HR)
Hence, people in different roles can benefit. For example, a compliance officer can enforce risk-based reasoning; a consultant can apply structured frameworks; a learner can explore step-by-step explanations.
This fundamentally shifts AI from ‘AI decides how to think’ to ‘humans guide how AI should think’.
Frugal AI: Intelligence Without Excess
Aligned with the conference theme, Kevin positioned Reasoning-Aware AI as a key enabler of Frugal AI.
Rather than scaling parameters, data, and compute, Frugal AI focuses on the following:
1. Better reasoning efficiency
2. Lower total cost of ownership
3. Deployment-ready models for enterprises
This is particularly relevant for SMEs, government agencies, and organisations requiring on-premise or sovereign AI. As demonstrated in his presentation, reasoning optimisation can deliver enterprise-grade performance without relying on frontier-scale models.
Democratising AI: Unlocking the 80%
A recurring theme in Kevin’s keynote was the untapped 80% of the workforce — non-technical professionals with deep domain expertise.
Through platforms like Capabara, combined with reasoning-aware capabilities, users can now build AI tools without coding. In addition, they can embed structured reasoning into workflows, as well as turn knowledge into scalable AI applications.
This aligns with Straits Interactive’s Capability-as-a-Service approach: enabling organisations to move from simply using AI to creating and governing AI capabilities.
Governance Meets Reasoning
Unlike many AI innovations that focus purely on performance, Kevin emphasised that reasoning must be governed.
Reasoning-Aware AI integrates:
1. Risk analysis and validation
2. Adversarial prompt safeguards
3. Transparent reasoning traces
4. Privacy-by-design principles
This reflects Straits Interactive’s long-standing leadership in AI governance and data protection, where responsible AI is embedded from the start instead of being added later.
Sovereign AI: Thinking Locally, Not Just Hosting Locally
Kevin also reframed Sovereign AI beyond infrastructure: Sovereignty is not just about where your data sits – it is about how your AI thinks.
It casts focus on enabling localised reasoning approaches, context-aware decision frameworks, and domain-specific logic. Reasoning-Aware AI supports true localisation; it does not zero in on just data residency.
From AI Tools to AI Factories
Closing the keynote, Kevin introduced a larger vision: the AI Factory. Organisations can continuously build AI tools, embed reasoning into workflows, govern outputs, as well as scale capabilities across teams.
This marks a shift from AI as a tool to AI as an organisational capability.
Tapping into reasoning-aware AI capabilities
Kevin Shepherdson’s keynote at BRAINCON 2026 delivered a timely and important insight: AI is not just answering our questions — it is shaping our thinking.
By introducing Reasoning-Aware AI, he challenges organisations to rethink not just what AI can do, but how it should think. In a world racing toward larger models, this may well be the more sustainable path forward; this technology will facilitate smarter reasoning, lower cost, greater control, and responsible outcomes.
Perhaps most importantly, it enables a future where humans remain in control of how intelligence is applied.
Capabara, our Next-Gen AI Capability-as-a-Service platform, is currently available on a limited trial basis. The model has been updated with in-built reasoning-aware capabilities. If your company is interested in developing Gen AI competencies and capabilities with us, write in to sales@straitsinteractive.com with your company email address. You can also stay tuned to Capabara’s latest developments by following CAPABARA on LinkedIn or heading over to capabara.com to find out more about how your organisation can be empowered through safe, secure and sustainable Gen AI adoption.
