Industry executives and experts share their predictions for 2025. Read them in this 17th annual VMblog.com series exclusive.
By Vall Herard, CEO of Saifr.ai
According to IDC’s latest forecast, global
spending on AI-supporting technologies is projected to exceed $749 billion by 2028. That means investments
in AI and supporting tech will soon outpace spending in key sectors like cloud
and digital services.
Additionally, the AI market is evolving from
an experimental phase to a more strategic one as business leaders demand
concrete proof of AI’s financial value. Companies are no longer content with
merely exploring AI’s potential; they now seek measurable returns on their
investments.
As we enter 2025, these new expectations are
transforming the AI landscape. This shift will likely be characterized by a
more pragmatic and results-oriented perspective. Agentic AI is emerging as the
most promising trend that decision-makers will be closely examining in the
coming year, alongside the rise of specialized AI models that challenge
traditional large language models, and the growing adoption of RegTech
solutions to help navigate increasingly complex regulatory requirements.
Breakthroughs
expected in agentic AI capabilities
Gartner has pinpointed agentic AI as the most critical technology trend to watch in 2025,
sparking widespread interest in its transformative potential. But what exactly
makes these intelligent systems so groundbreaking?
Agentic AI represents a quantum leap beyond
traditional AI approaches. Conventional AI systems follow linear command
execution. Agentic AI, on the other hand, can analyze the entire task,
determine the most logical approach, and execute steps in the most logical
order.
Consider the mathematical principle of order
of operations (PEMDAS) as a metaphor for agentic AI’s core strength. Just as
solving equations requires following a specific sequence to arrive at the
correct result, these advanced AI systems understand that the order of
problem-solving is often as crucial as the individual steps themselves. This
capability suggests profound implications for business intelligence, promising
more accurate and dependable AI-driven solutions.
Furthermore, agentic systems are capable of
understanding context-rich facts-such as the current time, weather, and
date-thanks to advancements in interoperability. These systems can access
third-party tools, agents, and applications to better understand the
environment in which they receive prompts.
In the financial services sector, agentic AI
could help revolutionize complex processes like loan underwriting. Imagine an
AI system that can autonomously break down intricate tasks into specialized
subtasks, seamlessly collaborate across multiple systems, and iteratively
refine its outputs based on continuous feedback. Such capabilities could
dramatically compress review cycles and enhance decision-making efficiency.
Specialized
AI has the potential to disrupt broad-based LLM market
Large language models have long been the
centerpiece of AI conversations, but 2025 is set to herald a significant shift
toward more specialized solutions. Smaller, domain-specific models are emerging
as powerful alternatives, offering unprecedented precision and efficiency
across critical industries like finance, healthcare, and robotics.
These targeted models represent a strategic
evolution in AI deployment. By focusing on narrowly defined datasets, they can
deliver superior performance with dramatically reduced computational
requirements. This efficiency translates into tangible benefits, such as lower
operational costs, enhanced accuracy, and the ability to operate in
resource-constrained environments.
These niche models shine by addressing complex
compliance challenges, particularly in highly regulated sectors. Their
industry-specific training allows for more controlled and predictable outputs,
reducing the risks associated with generalized AI systems. As business leaders
scrutinize the return on investment for AI technologies, the economic
advantages of these streamlined models become increasingly compelling.
RegTech
gains momentum
Naturally, small language models rely on
smaller datasets and therefore face significant limitations in addressing the
intricate compliance landscape of modern AI deployment. As AI becomes
increasingly sophisticated, organizations require more robust governance
mechanisms to help navigate complex regulatory environments. Regulatory
technology (RegTech) is becoming an increasingly compelling solution to this
issue.
We can understand RegTech’s potential by
considering generative AI’s current limitations. GenAI models typically provide
responses that satisfy a prompt’s basic requirements. However, these responses
rarely take nuanced regulations into account, which can pose risks for
communications in highly regulated industries. RegTech can provide critical
oversight in these instances, enabling businesses to monitor AI systems,
identify potential risks, and help ensure ethical operations. In other words,
companies can confidently integrate efficiency-boosting tools like AI into
their workflows while maintaining rigorous compliance standards.
Financial services stand at the forefront,
where RegTech applications are transforming data management, fraud prevention,
and regulatory adherence. This strategic approach not only helps mitigate
potential risks but also enables more intelligent and responsible technological
innovation.
##
ABOUT
THE AUTHOR
Vall
is the founder and CEO of Saifr, a RegTech incubated within Fidelity Labs. He
specializes in the intersection of financial markets and technology and has a
mastery of emerging methods like AI, machine learning, blockchain, and
micro-services. He has extensive experience in the financial services industry,
including FinTech, RegTech, InsurTech, capital markets, and hedge funds.
Vall
is a technical CEO with a proven track record of taking companies from ideation
to scale on a global basis, including multiple FinTech exits. His insightful
go-to-market and executable strategies have helped build profitable businesses
in banking, investment management, analytics, consulting, and sales. He has
previously worked at UBS Investment Bank, BNP Paribas, BNY Mellon, Numerix, and
more.
Vall
holds an MS in Quantitative Finance from New York University and a BS in
Mathematical Economics from Syracuse and Pace Universities, as well as a
certificate in big data & AI from MIT.
