New system promises faster and more energy-efficient AI agents

Researchers have unveiled a new optimisation framework called Murakkab that could significantly improve the speed and energy efficiency of artificial intelligence agents. By redesigning the multistep workflows that increasingly power advanced AI applications, the system aims to reduce computing costs while maintaining high levels of performance, addressing one of the fastest-growing challenges facing the AI industry.

Optimising the workflow rather than the model
Much of today’s AI innovation has focused on building larger language models with greater computing power. Murakkab takes a different approach. Instead of making the underlying model larger or faster, it examines how AI agents organise and execute the sequence of tasks required to solve complex problems.

Modern AI agents rarely answer questions in a single step. They often retrieve information, analyse documents, generate code, verify results and interact with external tools before producing a final response. Each stage consumes computing resources, increasing latency, energy use and operating costs.

Murakkab analyses these workflows and automatically identifies more efficient execution paths, eliminating unnecessary processing while preserving the quality of the final output.

Reducing both cost and energy consumption
As AI adoption accelerates, energy consumption has become a growing concern for both technology companies and governments. Large-scale AI systems require substantial computing infrastructure, with data centres consuming increasing amounts of electricity worldwide.

By reducing redundant computational steps, Murakkab enables AI agents to complete complex tasks using fewer processor cycles. This translates into faster response times, lower operating costs and reduced energy demand.

For organisations deploying thousands or even millions of AI interactions each day, even relatively small efficiency improvements can generate significant financial savings while lowering the environmental footprint of AI services.

Applications across multiple industries
The optimisation framework is designed to work across a broad range of AI applications rather than being limited to a specific model or industry. Potential use cases include enterprise assistants, software development, scientific research, financial analysis, healthcare support and customer service.

As AI agents become increasingly capable of handling multistep reasoning and autonomous decision-making, workflow optimisation is expected to become an essential component of future AI infrastructure.

The approach also offers greater flexibility, allowing developers to adapt workflows dynamically depending on the complexity of individual tasks rather than relying on fixed execution pipelines.

The next stage of AI development
Industry experts increasingly believe that future advances in artificial intelligence will come not only from more powerful models but also from smarter system design. Improvements in orchestration, memory management, tool selection and workflow optimisation may deliver performance gains comparable to those achieved through larger models, but at a fraction of the computational cost.

Murakkab reflects this changing philosophy. Instead of asking AI systems to work harder, it focuses on helping them work more intelligently by identifying the most efficient path through increasingly sophisticated reasoning processes.

Efficiency becomes a competitive advantage
As organisations continue integrating AI into everyday operations, efficiency is emerging as a key competitive differentiator. Lower inference costs enable broader deployment, while reduced energy consumption supports both sustainability goals and expanding regulatory expectations surrounding digital infrastructure.

Murakkab demonstrates that the future of artificial intelligence may depend as much on architectural optimisation as on breakthroughs in model capability. In an industry where computational demand continues to rise rapidly, systems that can deliver faster responses using fewer resources may prove just as valuable as the next generation of foundation models.

If widely adopted, workflow optimisation technologies such as Murakkab could help make advanced AI agents more scalable, more affordable and more environmentally sustainable, bringing sophisticated artificial intelligence within reach of a far wider range of organisations.

Newshub Editorial – Global, 9 July 2026