‘AI-adjacent’ tech rules
business in 2025

The AI boom is no longer about what generative models can create. It’s about the tools, systems and safeguards being built around them, usually out of necessity, to make AI viable in the real world.

This is the underlying message of a new report from Forrester, a leading global research and advisory firm that tracks the business impact of technology. Its annual Top 10 Emerging Technologies list is closely watched for its focus on practical value over hype. The 2025 edition, released recently, draws a clear line: the centre of gravity in tech has shifted from spectacular AI outputs to the hidden infrastructure that makes them usable, secure and scalable.

“The winners will be those that build a secure, trusted foundation, enabling them to scale AI faster while reducing risk and inefficiency,” says Forrester.

At the centre of this shift are what can be described as “AI-adjacent” technologies: those that don’t generate headlines but do generate return on investment. At number 3, Synthetic data, for example, is solving a growing dilemma in AI deployment: the need for large, accurate datasets that don’t compromise user privacy.

By generating realistic but anonymised data, companies can train powerful models without touching sensitive information. Only 14% of organisations have been able to scale up this tech, yet more than half are actively pursuing it, spurred on by regulators who see it as a safer alternative to conventional datasets.

TuringBots, a term coined by Forrester, are a surprise entry at number 2. These are intelligent developer assistants that are revolutionising the software development lifecycle. Already, over 70% of enterprise developers are using them to streamline coding, testing and deployment. These tools are becoming a bridge between abstract AI capability and practical enterprise application.

The appearance of Generative AI for language, at number 1, makes the point that we are still in the midst of the revolution that was sparked by the emergence of ChatGPT at the end of 2022. However, it is now surrounded by emerging technologies that give it greater business relevance.

At the same time, generative AI is changing shape. While language remains its most mature use case, GenAI for visual content (GAIVC), in 9th place on the list, is catching up. Businesses are using it to create 3D prototypes, architectural visualisations, and photorealistic marketing assets. However, Forrester puts a sober spin on its more elaborate uses: “Many companies will realise the benefit of generating production-ready content in the next year. However, most will not realise GAIVC’s role in immersive experiences for two or more years.”

Security, predictably, is lagging behind the hype. As enterprises race to adopt AI, they expose themselves to new vulnerabilities, especially when connected devices and legacy infrastructure collide. IoT security is becoming a priority not out of innovation, but out of necessity. Old medical devices, warehouse sensors and industrial control systems weren’t designed with AI in mind. Yet these are now part of the digital ecosystem and, without tailored security, they become points of failure.

Quantum security, ranked 8th, presents a more abstract threat, but one with enormous implications. The rise of quantum computing could eventually render today’s encryption obsolete. No current quantum system can break existing cryptography standards, but the risk of future breakthroughs means that agility in this area is no longer optional. However, while “post-quantum encryption” is emerging, it is expensive and poorly standardised.

Agentic AI, the current flavour of the moment, enters the list at number 5 with its promise of pushing beyond the simple chatbot into systems that can plan, act, and adapt. The vision is software that understands goals and executes on them across tools and systems. In theory, it’s a step toward general-purpose automation. In reality, it’s a messy frontier, filled with security challenges and the need for close human supervision. The report cautions: “Testing and validating agentic systems requires significant investment in resources today.”

Propping up the bottom of the list is the humanoid robot. China has declared it a national priority, and hardware costs are falling fast. Such robots could eventually automate tasks in logistics, healthcare, customer service, and home assistance. But their path is strewn with hardware limitations, integration headaches, and thorny ethical questions.