Ones to watch
The next frontier in intelligent reasoning
Neurosymbolic AI blends the pattern recognition power of neural networks with the logic and structure of symbolic reasoning. It’s designed to tackle complex, multi-step questions with clarity, traceability, and precision. This makes it ideal for high-stakes domains like healthcare, finance, and compliance. Unlike traditional AI, which often struggles with explainability, Neurosymbolic systems can connect the dots across fragmented data, deliver transparent answers, and adapt to specialised knowledge domains.
With frameworks like KnowGL and GenRL already showing impressive accuracy and scalability, Neurosymbolic AI is emerging as a strategic capability—not just a technical innovation.
It’s one to watch for anyone building AI that needs to reason, explain, and earn trust.
How can AI tackle complex, multi-step questions with clarity and confidence?
Vaishnavi R explores how Neurosymbolic AI, through the KnowGL framework, is reshaping question-answering in healthcare, law, and compliance. By combining neural networks with symbolic reasoning, KnowGL delivers accurate, scalable, and explainable results.
With 98% accuracy and human-level confidence, it’s a strategic leap for enterprise AI.
- Solves complex, multi-hop questions across large datasets
- Combines neural and symbolic methods for accuracy and explainability
- 98% accuracy, scalable and cost-effective
What if your chatbot could reason like an expert—not just guess like a parrot?
Vaishnavi R shares a hands-on journey from Neuro-Symbolic AI to Neuro-Conceptual AI (NCAI)—a method using structured domain knowledge to boost accuracy and explainability in finance, healthcare, and compliance.
With precision scores over 86%, NCAI offers a practical path to logic-driven, trustworthy AI.
- Uses structured domain knowledge for logic-driven answers
- Ideal for finance, healthcare, compliance
- Scores: Precision 86%, Recall 92%, F1 86.5%
What if your AI could answer complex questions with logic—not just pattern matching?
At Version 1 AI Labs, we built a demo using GenRL, fine-tuning BART to map natural language questions to structured knowledge base relations—merging neural flexibility with symbolic precision.
The result? Clear, accurate answers for domains where trust matters.
- Live demo of Neurosymbolic AI for Q&A
- Fine-tuned BART maps questions to knowledge base
- Delivers clear, accurate answers for trust-critical domain
AI trend radar
We're seeing AI categories like agents, multi-modal systems, and privacy-preserving techniques hold steady, while the subtopics within them evolve rapidly.
Our view is that this reflects a shift from broad innovation to deeper specialisation.
Techniques like federated learning and neurosymbolic AI offer compelling implications for privacy and reasoning, though they come with trade-offs in complexity and maturity. Multi-modal models are increasingly practical, while liquid neural networks remain exploratory.
Consider this a signal: the next phase of AI progress may lie in refining what we already trust, not reinventing it.
From agentic systems and frontier models to global governance and ethical considerations, this report has explored the full spectrum of AI’s evolution and impact.
As the landscape matures, the true differentiator is how organisations turn these emerging trends into practical, responsible solutions.
By focusing on trusted technologies, operational excellence, and ethical leadership, we help clients move confidently from experimentation to enterprise-scale impact. The next phase of AI goes beyond innovation into delivering measurable value, securely and sustainably, across every sector we serve.
Research by: Rosemary J Thomas, PhD
Senior Technical Researcher, Version 1