Top 5 Business Trends in AI for 2025, from Sertis CCO, Randy McGraw: On behalf of the 140-strong Sertis team serving the AI and Data needs of the region and beyond, Happy New Year 2025 to our customers, industry colleagues, and Sertis alumni spread all around the world. Looking back on 2024 - a banner year for AI and Sertis - we would not be good AI professionals if we did not “review & extrapolate” from our position at the center of AI in the enterprise sphere.
Top 5 Business Trends in AI for 2025
Positive movement in the ROI case
Agentic AI
Unified AI Frameworks
Empathetic AI design
Rightsizing of AI models
1) Positive Movement in the ROI Case
ROI has been the hot topic in AI deployments for years, the sine qua non of adoption. Prohibitive cloud cost (running cost), project deployment risk, high vendor costs, legacy systems requiring more squeeze, ill-prepared data infrastructures, and high capital hurdle rates often made AI projects “nice but not necessary” risky bets, with alternatives to AI or project postponement more attractive. AI had become an experiment, a learning curve of ‘what ifs’ and preparations for a future that would surely come but was not “quite here yet.”
Moreover, most companies lacked a comprehensive, strategic, and enterprise-wide AI approach, causing fits-and-starts trying to build AI on top of antiquated ERP, CRM, and EDI systems or via palliative-type solutions, the end results being limited and disappointing. Without a solid foundation of data, modern production technologies and processes, and a very planned-for outcome, meager results were often followed by major disappointment. But heavy-up infra investments by hyperscalers in 2024, reduction in cloud costs, the emergence of GenAi, semi-productizations in the arena, and other important factors such as Sertis’ clear-cut case studies and reusable tools which it can “consult and deploy”, are conspiring to flip the ROI case attractively across verticals for a wider array of applications and enterprise use cases.
In 2024, Sertis began 25 POCs with large enterprises in Thailand alone, a jump of 20% from the previous year, and has at least 8 medium to large AI deployment deals in the bottom of the sales funnel as of December end. The ROI case is no longer as much of an impediment as it was. AI project deployments are picking up as the ROI case turns positive, more germane case studies are available, Agentic AI can further reduce risk and cost whilst staying inside enterprise security guidelines, and as companies like Sertis have a head start in knowledge, skills, and reusable parts to help the ROI case.
2) Agentic AI The rise of agentic AI is a key trend for the future of work, defined as integrated workflows and processes wherein AI Agents autonomously execute tasks, collaborate with human workers, and drive value across the business. Agentic AI has the potential to achieve significant gains in operational efficiency, customer experience, and decision-making. It also leans neatly into the ROI and CapEx case of any project, enabling a sharper understanding for all stakeholders including financial and budget managers, of overall value. It also reduces project risk overall by enabling a more reliable assessment of how operational flows are impacted, and how AI will be integrated into existing workflows by breaking tasks into digestible subcomponents.
In one example, the cost of LLM access for scaled repeating tasks, vast differences in output, and emerging security concerns are all conspiring to drive a new type of product category into the mix, one with broad applicability. Smaller, specialized Agentic AI learning models are emerging as powerful alternatives to massive public LLMs. Private LLM or SLM - Small Language Models - which are purpose-built or industry-built models designed for focused enterprise tasks, or simpler computational tasks that exist on-premises rather than in the cloud as many companies now have compliance requirements to be off public clouds.
Sertis has both Private LLMs and SLMs and highly specialized KMAI - Knowledge Management AI (enterprise-specific internal chatbot on steroids) that sit atop the private LLM/SLM - or on public clouds and LLMs - to deliver measurable value to any enterprise by enabling smaller morsels of tasks to be enhanced rather than changed outright. An entire internal KMAI is now in only the range of 5M THB, give or take, and Sertis has a Private LLM all-in at 20M THB (which we can also finance over 60 months).
3) Unified AI Frameworks To hit the gas pedal on AI, fear of misstep is exceeded only by the actual risks of misstepping, and we see the roadway behind us littered with premature and ill-begotten approaches. Sertis itself was involved in no fewer than 2 major rescue projects this year alone; 2 in 4 CTOs we surveyed outside of Thailand stated they had regret for going too quickly into certain AI projects in the last 5 years. Unified AI Frameworks help ameliorate this risk, and enable a much-needed North Star for decision-making. Enterprises with multiple AI projects across different departments need such a framework to mitigate risks related to compliance, data privacy, legal issues, intellectual property, and, equally as important, de-siloing an AI endeavor inside a company. “What should we do, under what conditions, what is the right scope and definition of outcomes, and how do we all need to contribute to an initiative’s success?” can no longer be an internal debate always starting from scratch.
Drafting a Unified AI Framework for an organization is a cross-functional exercise, one in which Sertis and its expertise in Data and AI are uniquely positioned to assist. Such a document, with multiple departmental thumbprints, serves multiple derisking masters, from helping eliminate risk from the use of unapproved AI tools, software, or platforms by employees or teams within an organization, to ensuring a comprehensive checklist of ROI approaches for the enterprise before a project is commissioned. If your enterprise does not have a documented Unified AI Framework, you are not alone. If you do not have one by the end of 2025, you will be.
4) Empathetic AI Design
The human-centric AI approach shifts AI from pure productivity tools to enhancing human experiences and capabilities, aligning with the rise of Agentic AI. This approach by enterprises will help employees automate routine tasks, boost creativity, and unlock new opportunities in a way that is emotionally and intellectually comfortable for people. By focusing on empathetic AI design, enterprises can build stronger teamwork, reduce stress and burnout effects, deepen customer relationships, and even deepen brand loyalty. We will also see common apps in the marketplace, ones we use every day in our work and leisure, obtain an AI layer that may be invisible to users but enhance reliability and results in a comfortable and unobtrusive way. Sertis deploys a large internal team of UIUX designers that not only deploy the latest objective-oriented design tools that can optimize usability and design but are specifically tasked to temper that with an overt design approach to integrate enterprise tools within existing processes in the most comfortable way possible for human end-users.
5) Rightsizing AI Models
For enterprises that have already taken a plunge, 2025 will see them internally rightsizing the AI approach to enterprise capabilities, leaning into a Unified AI Framework, and taking a highly structured approach to implementation. By focusing on modernizing legacy systems, cleaning up data once and for all, and empowering employees with the tools they need to perform, enterprises can bridge the gap between expectations and the reality of what AI will deliver. Success in rightsizing will embed intelligence directly into workflows where it truly matters rather than overkilling or taking a one-size-fits-all approach to single-threaded business objectives. AI deployments have a tendency to miss the mark once deployed or fail to pass through the mid-funnel when this is the case. 2025 will be a year in which Sertis can work with its existing customers and new enterprise customers to ensure overkill and underkill are uniquely avoided.
Bonus: Investment Sectors
We see these 5 trends spurring fresh project investment from (1) the Retail sector which has been slow to adopt in part because of poor margins and legacy system traps, (2) the Banking/Financial Services sector including insurance where everything from Computer Vision AI to Predictive AI to Generative AI can have a clear, measurable business impact and the organization is accustomed to teeing up, analyzing, and acting on cost-saving operational measures, (3) the Manufacturing sector where supply chain rationalization interfaces directly with internal organizational processes that can be impacted significantly by AI deployments, (4) the Healthcare sector and (5) universities.
Regardless of your industry vertical, 2025 will be a year in which all enterprises will need to make AI investments to “keep up” and “make comp” or else fall behind competitors. There is a sustainable cost structure and revenue advantage in being the first in a sector or competitive grouping to successfully launch AI initiatives. The boom-bust cycle of business has become consistently and unmistakably more rapid ever-consistently over the past 500 years as technology makes its impact felt, and now literally a single year can make a world of difference in competitiveness and viability to the market. Buckle up, the car is near the top of the first roller coaster hill.
Sertis is your partner in navigating these waters, a consulting design and delivery firm reaching over 75 public sphere enterprises and 400+ AI and Data deployments over the last 10 years. Contact us to see how we can help you: https://www.sertiscorp.com/th/contact-us
Yorumlar