Five Trends That Will Make This A Business Break-Out Year For GenAI

Usman Javaid is the Chief Products and Marketing Officer at Orange Business.

A new era of generative AI (GenAI) is rapidly unfolding in front of us, and clearly, we know only a fraction of its enormous potential. As a global chief products and marketing officer at a network and digital integration company, I see firsthand how AI is transforming industries across the globe and accelerating the pace of innovation in ways previously unimaginable. While we’re only scratching the surface of what this technology can achieve, understanding what’s coming down the pipe in GenAI is pivotal to plotting a course through this new frontier.

The technologies that underpin GenAI are developing at an unprecedented rate. This is primarily due to significant investment from tech giants and research communities, alongside well-funded startups, according to Gartner researchers. For example, Gartner’s analysts expect that “by 2027, more than 50% of the GenAI models that enterprises use will be specific to either an industry or business function—up from approximately 1% in 2023.” These models tend to be smaller, less computationally intensive and have lower hallucination risks than general-purpose models.

Five GenAI Trends To Prepare For

From my recent interactions with enterprise customers and technology partners, I’ve singled out five trends that underscore the need for organizations to adapt their overarching strategies, infrastructure and skill set to ensure they remain creative and competitive in the rapidly evolving AI landscape and maximize return on AI investments.

1. Real-time applications will drive inference to the edge.

The likes of GPT-4o and Google Gemini open new possibilities powered by multimodal GenAI that enables human-like interactions with digital assistants in real time. This will open a new era where every device will have models built in and ready to service low latency, highly interactive use cases such as contextual augmented reality navigation, real-time translation, human-centric therapies and interactive retail applications. Faster response times for real-time AI applications will require harvesting and processing data closer to the users, driving inference to the edge.

To this end, tech players and enterprises will invest in distributed AI inference services, making it possible to invent new “killer” AI applications. The distributed AI inference architecture will also be beneficial for reduced carbon footprint, data sovereignty, optimized total costs and end-to-end performance.

2. Shadow GenAI use will continue to rise.

Noncompliant and rogue GenAI usage is becoming dangerous because it expands the cyberattack surface as enterprise users experiment with this technology. This is the result of deploying GenAI tools and systems without the explicit approval of the organization. Research shows that over one-third of people are using GenAI daily, with 75% looking to use it for work communications and automating work tasks. This is causing a security headache for IT departments. Pandora’s box is open, so enterprises must adopt tools to ensure secure, transparent and auditable AI access. Such tools will provide visibility into the use of noncompliant GenAI and enable IT admins to provision compliant alternatives.

3. AI models will take an open approach, with SaaS solutions winning.

Major AI players are taking an open approach to designing and implementing AI systems to minimize dependency on any one technology platform or AI model. This gives enterprises the flexibility to adopt new AI processes without being tied to a model or disrupting operations. The value will be realized in the model-to-app chain, which integrates AI models into practical applications. Models can be offered as SaaS-based solutions, which solve specific business needs such as knowledge management, digital assistance, customer support or expert advice.

4. Dynamic system design will move toward increased modularity and higher abstraction.

Modular system designs mean components can be developed and maintained independently and integrated seamlessly. In parallel, there’s a move to higher levels of abstraction, which allows complex implementations to be hidden behind simpler interfaces. This enables developers to create more advanced applications without needing to understand underlying complexities, speeding up prototyping and innovation. Leading tech companies such as Amazon, Microsoft, Google and Salesforce are developing tools and platforms that exploit these trends. AWS has demonstrated the capabilities of Q Developer, which automates the abstraction layer. Developers can interact in natural language, and Q will generate code. And Salesforce leverages metadata as a foundational element for developing and deploying applications, particularly those involving GenAI.

5. Voice applications will continue to advance with GenAI.

Advanced multimodal LLMs can generate more natural and expressive speech, making voice assistants feel more human-like. This will enhance interactions with customer service bots and interactive voice response systems. It will also advance real-time translation tools. New use cases are emerging with GenAI in two-way voice applications (user to user), as demonstrated by Microsoft Teams Phone Mobile, including call assistance and fraudulent call identification. In the future, if the called party is unavailable, the GenAI-powered personal digital assistant can take over the call.

Conclusion

As GenAI continues to mature, it will undoubtedly reshape industries, setting new standards for creativity, productivity and business intelligence. Enterprises that lead the pack in GenAI adoption will achieve a careful balance between delivering business outcomes and addressing their infrastructure foundation, compliance, dynamic system design and new ways of human interactions.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?