The 2-Z of 2025 predictions: A is for AI, part 1

A is for artificial intelligence (AI)

Source: Salesforce. Barfield.

2025 is the year generative AI moves beyond excitement and experimentation.

"Despite the initial excitement, few organisations have moved beyond proofs of concept (PoCs) and limited trials to full-scale implementation. In some cases, generative AI has failed to deliver accurate and useful outputs due to incomplete data. In others, solutions are disconnected from workflows, making them clunky and inefficient. 

"Many applications of generative AI, such as copilots and chatbots, were created as 'solutions looking for problems', focusing on experimentation rather than solving actual business issues," said Gavin Barfield, VP & CTO, Solutions, Salesforce ASEAN.

Utkarsh Maheshwari, Chief Partner Officer and Head of Midmarket, SAP Asia Pacific Japan (APJ), also said that 2025 will be the time when AI goes beyond the flashy buzzwords. 

Source: SAP.
Maheshwari.

"While we’re seeing fatigue from business leaders on the AI hype over the past months, in 2025, they will begin to realise the true power and value it brings as AI technology gains more widespread familiarity across business functions and at the executive level. As perspectives on generative AI (gen AI) mature, we will see more practical applications of the technology that solves actual business problems," he said.

"This increased adoption will further break through traditional narrow AI capabilities to drive true automation within business processes. We can also expect a rise in novel use cases previously unimagined, particularly as agentic AI capabilities – that is, AI-powered systems that act as intelligent agents with a certain degree of autonomy – take shape." 

Mohan Krishnan, VP & GM, HPE GreenLake Cloud Services, APAC, HPE, agreed that AI is evolving significantly in 2025. "While the past two years have seen an explosion in AI innovation, much of it has remained at the testing and experimentation stage. Heading into 2025, we will see organisations in Asia Pacific (APAC) take a more practical, outcome-driven approach towards AI, as business leaders look to translate AI’s potential into business value and optimise the return on their AI investments," he said. 

Source: HPE. Krishnan.

"In fact, IDC has called 2025 the year of the AI Pivot, where organisations will move beyond AI experimentation to comprehensive AI integration into their operations to accelerate innovation, efficiency and growth."

Experimentation is fine, going live is another thing entirely. "The breakneck pace of innovation in generative AI also means while many are interested in adopting and starting POCs/pilots, there is a lot more caution and consideration in moving those workloads into production," said Wu Chun Wei, MD, Technology, Temus.

"The human in the loop still remains critical to ensure a right balance of safety and innovation...With the emotional intelligence and critical thinking of people, combined with the speed and performance of AI, we have unprecedented opportunities to create new solutions and processes that will benefit businesses, our community and society at large."

Source: Temus. Wu.

Enterprises will move toward a renewed focus on fundamental business values and practical AI, predicted Pure Storage. In a list of 2025 predictions, the company stated: "Paradoxically, in dollar terms, enterprise investments in AI will increase in 2025 while the total number of gen AI proof of concepts (PoCs) and pilots will decline. In 2024, the failure rate for PoCs was higher than anticipated as they failed to deliver on expectations, or were not economically viable when scaling from the training phase to the inference phase." 

First-mover advantage

The rapid adoption of AI by tech giants has created a bottleneck in the supply of graphical processing units (GPUs), the processors used for AI, observed Ying Shaowei, Chief Scientist, NCS. 

"These giants have the resources to secure these components, leaving smaller players struggling to keep up. Analysts are predicting the GPU market will grow at a CAGR of 34% over the next decade. With the current industry dynamic, it will not be surprising that this growth be accompanied by an increasingly uneven playing field where only the most resourceful companies can thrive," he noted. 

Analytics

"One of the key learnings from 2023/4 is that a less sophisticated algorithm powered by a large dataset will outperform a more sophisticated algorithm accessing a smaller dataset. Armed with this knowledge, enterprises in 2025 will undertake projects to free-up siloed and locked-up datasets in the quest to improve the output of their analytics and AI investments," said Pure Storage.

"This emphasis on data unification will also reflect a broader understanding of data's strategic importance in driving innovation and maintaining a competitive edge. As organisations aim to harness the full potential of AI and analytics, they will prioritise initiatives that enhance data quality, streamline access, and foster collaboration among teams. Ultimately, this focus on unifying internal datasets will pave the way for more informed decision-making, improved customer experiences, and sustainable growth.

"Despite repeated efforts to democratise access, analytics has rarely reached more than 25%-30% of users. Gen AI-powered conversational interfaces added to business intelligence tools can help reach the remaining 70-75%, giving more employees access to insights," forecast Qlik. 

"Other approaches will exist, but it will increasingly become the dominant way to interact with data. A well-designed interface hides complexity. It connects the tasks to the tools in the platform underneath: fetching answers from a knowledge base, getting a data point from an existing dashboard, and starting automation."

Diversity

Source: RackSpace. Banerjee.

"The Asia Pacific region spans numerous languages and cultures, requiring AI models to be trained on localised, culturally relevant data. CIOs should focus on creating language-specific datasets and ensuring that AI solutions are inclusive and adapted for local markets, especially for customer-facing applications," said Hemanta Banerjee, VP of Public Cloud Data Services, Rackspace Technology.

"Collaborate with regional data providers to source diverse datasets and prioritise AI models with multilingual capabilities. Invest in data annotation and translation services to train models in languages relevant to the Asia Pacific markets," he suggested.

"Localisation is becoming increasingly important in AI deployment," agreed Ying Shaowei, Chief Scientist, NCS. 

"To be truly effective, AI systems must consider the specific needs and conditions of local markets. This has led to the rise of new AI players who offer industry-focused solutions tailored to local contexts, providing businesses with more choices in navigating AI adoption."  

NCS is working on a conversational AI system that can understand Singaporean-accented speech, Ying shared. The Singapore variant of English contains borrowed words from other languages and many references to local places and events, making it more challenging for speech recognition systems to understand.

Ethics

Mizutani.

"Businesses and regulators will collaborate to establish clear AI policies. In Asia-Pacific, Singapore’s regulatory sandbox and Japan’s ethical guidelines are paving the way for responsible AI deployment. These frameworks will balance innovation with societal safeguards, providing enterprises with clarity as they integrate AI into their operations," said Yasutaka Mizutani, APAC President, Colt Technology Services.

Banerjee also said that the public and government sentiment favour data sovereignty and ethical AI. "CIOs should prioritise transparency in AI-driven decisions, especially in sectors like finance, healthcare, and public services. Establishing robust ethical guidelines and enhancing transparency in AI operations will build trust with stakeholders," he noted.

"Develop a regional ethics board or governance framework specifically for AI and data initiatives. Engage with local stakeholders to communicate data handling practices clearly, particularly for AI applications involving sensitive or personally-identifiable information." 

Anthony Spiteri, Regional CTO APJ, Veeam Software, added that AI middleware can help to maintain ethical standards without the need for in-house AI specialists. "This is significant as governments are imposing stricter regulations around the responsible and ethical use of AI. With the rise of AI middleware solutions, businesses will see a marked increase in the volume and complexity of data they need to handle. This surge will drive a greater need for robust data management practices, ensuring that critical AI datasets are well-protected and retained securely. As companies scale AI applications, the ability to efficiently manage and safeguard this expanding data pool will become essential to building a resilient AI strategy," he said. 

Lenovo predicted that leaders will not only be considering whether their AI tools align with their brand and business model, but also how regulatory compliance is supported. "With governments in markets such as Australia beginning to propose regulations and guidelines on ethical and responsible AI use, we will see more businesses look to AI solutions that align with these requirements from the start, or engage third parties to provide best practice guidance in AI implementation," the company said.

Governance

"In 2025, we expect the potential for regulations to increasingly address the use of AI itself. Governments and regulatory bodies around the world are racing to assess the implications of AI and develop governance for its use. The trend toward sovereign AI – a nation’s control or influence over the development, deployment and regulation of AI and regulatory frameworks aimed at governing AI – is a focus of The European Union’s Artificial Intelligence Act and China’s Cybersecurity Law (CSL) and AI Safety Governance Framework," noted Vertiv in a list of 2025 predictions. 

"Some form of guidance is inevitable, and restrictions are possible, if not likely."

Vertiv also added that governance will continue to be local or regional rather than global in 2025, and observed that the consistency and stringency of enforcement will vary widely.

"As AI systems become more integrated into everyday life, discussions will focus on the responsible use, compliance, data protection, and anti-discrimination laws and AI quality standards," said Tay Bee Kheng, President, ASEAN, Cisco.

"Collaboration between public and private sectors will be crucial to establish standards and baseline regulations that promote innovation and enhance AI safety. Global leaders will face mounting pressure to implement frameworks that improve the accountability of AI systems and address ethical and misinformation issues arising out of the use of AI, without inhibiting innovation. Companies will need to adopt responsible AI frameworks, conduct regular privacy assessments, and develop and implement a robust incident management plan to ensure judicious AI use.

Optimisation

Optimising AI workloads will be crucial in unlocking cost savings and performance gains, said Jay Jenkins, CTO, Akamai Technologies APJ. "As businesses in the APJ region face soaring AI workload costs, 2025 will mark a critical turning point. Leaders will prioritise optimising the inference phase—where AI generates actionable insights—to streamline operations and boost speed and accuracy," he said. 

"This focus on optimisation will not only cut computational expenses but also enhance performance, allowing organisations to redirect resources toward growth and innovation. The result will be a powerful cycle where smarter AI translates into improved profitability and continuous advancement in AI capabilities." 

Krishnan said a shift towards AI integration at scale will bring to the forefront the urgent need for AI-optimised IT infrastructure and data management. "To make the most out their AI, which is an inherently data-intensive and hybrid workload, more organisations will need to invest in native AI systems that optimise everything across the AI lifecycle, regardless of whether the workload is on-premises, in a colocation facility, the public cloud, or at the edge," he said. 

Source: NVIDIA. Pette.

"Validated reference architectures, which provide guidance on appropriate hardware and software platforms, will become crucial to optimise performance and accelerate AI deployments. These architectures will serve as essential tools for organisations navigating the complex terrain of AI implementation by helping ensure that their investments align with current needs and future technological advancements," said Bob Pette, VP of Enterprise Platforms, NVIDIA.

Privacy

"Companies adopting AI technologies face numerous security challenges, including gen AI privacy and data protection concerns, as well as vulnerabilities within the AI supply chain. To address these risks, the adoption of private AI platforms is set to grow. These platforms can empower companies with full control over their data, safeguarding their operations against increasing threats such as unauthorised data sharing, regulatory non-compliance, and the proliferation of 'shadow AI' usage," said Vishal Ghariwala, CTO for SUSE Asia Pacific. 

"Data privacy and security is another tenet of AI governance. As organisations increasingly operate across multiple jurisdictions, there will be increasing pressure on them to adopt measures that align data storage and processing with local data sovereignty laws. The next generation of privacy laws will continue to drive transparency, fairness, and accountability in spaces like data collection and use, cross-border data flows, and verifiable compliance. Companies will have to consider how their employees interact with AI systems and develop strategies to mitigate data breaches and associated risks through mock drills and continuous monitoring," said Tay from Cisco. 

Talent

Richard Yan, VP, People & Talent, Airwallex said: "By leveraging AI to automate manual tasks and free up valuable time, leaders can empower their teams to focus on critical thinking and culture-building initiatives. Those who prioritise meaningful connections with employees will stand out in the talent market, with their strong collaborative cultures as a differentiator. Ultimately, leaders who get this balance right will recruit and retain high-performing teams who unlock new opportunities for growth and innovation."

Source: Accenture. Govender.

“Heading into 2025, Accenture’s Pulse of Change Research found that 69% of companies plan to accelerate hiring to match pace with developments in generative AI. Generative AI (gen AI) is revolutionising even traditionally non-technical fields like HR, finance, and procurement, creating new opportunities for AI-focused roles like test analysts and data analysts - positions requiring functional expertise and foundational software engineering knowledge," said Nesan Govender, Talent & Organization Lead, Southeast Asia, Accenture.

"In the near future, the in-demand professional will be the techno-functional collaborator, coined in our latest Tech Talent report to be multidisciplinary expert blending technical skills, business acumen, and strong interpersonal abilities to integrate people, processes, and technology effectively. Similarly, businesses must become talent creators, proactively assessing existing employee skillsets, identifying reskilling opportunities, and investing in continuous learning and development to reinvent work and reshape the workforce in preparation for a gen AI world.” 

Wu from Temus, on the other hand, suggested that there will be a hiring boom for a diverse group of professionals "including clinicians, data scientists, legal, compliance, risk, and marketing professionals to oversee the development and deployment of AI-enabled technologies and monitor its impact on the organisation."

Work transformation

Jess O'Reilly, Area VP, Asia for UiPath, termed the inevitable reshuffle of work tasks across humans and AI the 'Great Work Reallocation'. "Agentic AI will drive enterprises to 'redesign and reassign' jobs and workflows to better leverage the unique strengths of both humans and machines," she explained.

"Starting in 2025 and continuing through the decade, enterprises will face the immense challenge of reinventing operating models, reshaping jobs, retraining workers, and redistributing tasks between human and virtual employees. The C-suite will lead this transformation, supported by a growing network of consultants and operations experts focused on designing new AI-driven operating models, managing large-scale change, and implementing cross- enterprise agentic systems." 

Source: UiPath. O'Reilly.

O'Reilly elaborated: "As enterprises face the growing challenge of managing handoffs between humans and machines amid workforce reallocation, orchestration capabilities will become crucial to ensure clear roles, systems, and processes. Without enabling infrastructure, orchestration, and controls, agentic AI will be unscalable and unsustainable. Meanwhile, a human workplace without defined roles, systems, and processes would also be chaotic, underperforming, and unproductive.

"The focus will shift towards setting in place a new AI and automation-infused workplace ecosystem designed to foster full collaboration among agents, robots, and people—while providing control, visibility, and active governance."

Ricky Kapur, Head of Asia Pacific, Zoom, said that new paradigms of work will emerge in 2025. "AI in the workplace will shift from one- size-fits-all solutions and surface-level automation to highly customised experiences that address the specific needs of businesses and employees. As organisations adopt increasingly complex tech stacks, critical information often becomes dispersed across various platforms," he said.

Source: Zoom. Kapur.

"Custom AI will play a pivotal role by integrating with internal systems and knowledge collections, quickly resurfacing relevant information, and enabling employees to focus on high- impact tasks. For example, AI can retrieve an employee’s relevant project updates, automatically summarise key discussions, and assign actionable tasks, seamlessly managing workflows across multiple platforms."

Explore

This is the 1st of a three-part series on AI predictions for 2025. Read part 2 here, and part 3 here. 

More AI-related predictions can be found in posts on agentic AI as well as AI cybersecurity, as well as throughout the 2-Z of 2025 predictions series. A separate post on AI skills digs deeper into what will be required in the era of AI.

Hashtag: #2025Predictions