There is a lot of hype around AI right now, which is why it was unsurprising when we reached out for predictions, that this topic received the most responses. We are excited to share the predictions experts and colleagues in the field shared with us and hope they help inform your AI decisions for 2025.
Retailers Look to Co-Create Experiences with Customers Using AI
We’ll continue to see the adoption of AI rise and become increasingly capable, enabling what retailers can do to enhance the customer experience. AI is already good at personalizing product recommendations, tailoring search results, and generating segment-specific content at scale. I expect to see these use cases continue to be refined and expanded so that complete experiences, not just parts of it, are co-created between a brand and its customers in real time. In order to stand out, retailers should leverage the enthusiasm younger shoppers have for AI by proactively co-creating personalized shopping experiences that feel like an authentic, fluid, familiar conversation, rather than a sales pitch. – Bobby Meixner, Senior Director of Industry Solutions at UserTesting
AI Governance Goes Mainstream
A good analogy for successful AI adoption can be drawn from sports. Winning teams have a solid balance of effective offense and strong defense. In 2024, many companies faced challenges in advancing with AI due to concerns about security and governance. Given the potential benefits of AI adoption for businesses, enterprise-wide AI governance programs are expected to become a key focus in 2025.
Risk management involves understanding both the potential harm and the likelihood of it occurring. For AI this involves handling traditional data risks (sensitive information exposure, privacy, security, and quality) and new generative AI risks (hallucinations, harmful content, bias amplification, and explainability). Ultimately for AI, it’s about implementing trustworthy solutions through a combination of policies, procedures and best practices.
Another key aspect of AI governance to watch in 2025 is regulation. Local, state, federal, and industry-specific regulation will be a key factor in AI governance and overall adoption for 2025 and beyond.
Enterprise tools are rapidly emerging to help organizations manage the entire scope of an AI governance program. There has also been substantial progress on foundational frameworks for AI governance. The NIST AI Risk Management Framework and Playbook is a great example. In 2025, companies that aspire to be AI leaders will move aggressively to adopt these tools and frameworks. – Steve Tranchida, Vice President, Digital Architecture & Strategy at Verinext
Integrating AI in Human Resources
As we wind down 2024 and ramp up into 2025, the integration of AI in Human Resources will continue to accelerate and evolve. The true potential of AI lies in its ability to enhance, rather than replace, human interaction. By leveraging AI tools, HR professionals can provide employees with more meaningful engagement and support by freeing up their time to build deeper, more impactful relationships. For example, AI can analyze employee feedback and sentiment in real-time, enabling HR to address concerns promptly and effectively. These AI-driven insights can help tailor personalized development plans, ensuring that each employee feels valued and supported in their career journey.
In 2025, I believe the key to successful AI integration in HR will be to find the right balance. While AI handles the administrative workload, HR teams can concentrate on fostering a positive workplace culture and building strong, personal connections with employees. This dual approach will not only improve operational efficiency but will also enhance employee satisfaction and retention. – Louise Willoughby-Petit, Chief People Officer at Beekeeper
AI-Driven Autonomous Debugging Advances as the Next Level of AIOps and AI Code Assistants
AI-powered autonomous runtime debugging will likely emerge by 2025 as a complement to traditional live debugging tools, marking a new maturity level in observability. This advancement is expected to converge with AIOps (Artificial Intelligence for IT Operations), driving automation in root-cause analysis to pinpoint and diagnose issues at the code level with minimal human intervention. Simultaneously, it will align with AI Code Assistants, supporting developers in creating more robust applications and accelerating issue resolution.
Autonomous debugging tools will enable developers to detect issues early and trace errors back to specific lines of code, transforming critical issue tickets into actionable insights. By analyzing both historical and real-time data, these AI-driven tools can also provide proactive recommendations and even enable auto-remediation for recurring issues. Additionally, when paired with AI code assistants like AWS CodeWhisperer and GitHub Copilot, autonomous debugging positions developers to build more reliable, high-quality applications with greater efficiency and precision. – Ilan Peleg, CEO at Lightrun
Structured Data Will Become the Backbone of AI-Driven Insights
In 2025, structured and semi-structured data sources will form the foundation for robust AI-driven insights. By leveraging data lakes, data warehouses, and data streams, organizations will enhance AI model training, improve data quality, and generate more accurate, actionable intelligence. This strategic approach will empower businesses to unlock the full potential of AI, driving continuous innovation and maintaining a competitive edge. – Rohit Choudhary, Co-founder and CEO at Acceldata
AI’s Next Leap: Specialization Over Smarter Models
As foundational AI models continue to evolve, there’s a limit to how much smarter they can get, given their dependence on vast amounts of data. The next frontier lies in enhancing these models with specialized add-ons that boost their adaptability and domain-specific capabilities. Expect genAI to become more context-aware, customized, and deeply integrated into industry workflows, transforming advanced machine learning from a technological marvel into a practical, indispensable tool that enhances human potential across sectors. – Eric Sydell, Ph.D., CEO at Vero AI
Open Source LLM vs. Subscription-Based: Who Will Win in 2025?
Meta changed the rules of the Large Language Model (LLM) game by open sourcing their model, Llama. Now, Meta is on track to have the most widely deployed chatbot in the world by the end of the calendar year 2024, despite OpenAI’s initial leadership with ChatGPT.
As the GenAI race heats up and more native artificial intelligence Independent Software Vendors (ISVs) emerge, open source models will continue experiencing exponential growth. ISVs will adopt an open source model like Llama instead of building on top of a model with a licensing fee involved. Ecosystems will form around open source LLMs, and they will gain critical mass. – Ratan Tipirneni, CEO at Tigera
AI-Generated Document Experience Evolution: Human-Computer Interfacing in Real-time
As we embark on the journey of understanding AI’s impact on productivity, particularly in how we create, interact with, and experience documents, we find ourselves at an exciting crossroads. Currently, generative AI is revolutionizing content creation, enabling us to produce new material with unprecedented ease. Additionally, AI’s capability to access and summarize text from images has transformed our interactions with documents, making them more intuitive than ever.
Looking ahead, I anticipate a significant evolution in how we experience these documents. One of AI’s groundbreaking advancements is its ability to establish a direct interface between humans and computers. While the popularity of natural language chatbots is currently capturing attention, they serve as a preliminary step in demonstrating the potential of transformer models.
In the coming year, we can expect this new human-computer interface to facilitate real-time personalization of document content. This means that interactions will become dynamic, tailored to individual preferences and past experiences, all while leveraging the most current information available. Over time, the conveniences brought by AI will become so integrated into our daily lives that they will be taken for granted, much like our constant connectivity to the internet today. – Jonathan Rhyne, CEO & Cofounder at Nutrient
Low Code Tools Will Streamline Compliance
Industries such as healthcare and finance, where compliance with strict regulatory standards is critical, often face extended development timelines due to the rigorous testing required. However, the growing adoption of low code tools is poised to revolutionize the time needed to adhere to these standards. Low code platforms not only accelerate app development but also ensure that applications are built in alignment with legal and regulatory requirements. By integrating industry best practices into the development process, low code solutions will streamline compliance, enabling faster delivery of secure, compliant apps without sacrificing quality or oversight. –JJ McGuigan, App Builder Product Manager at Infragistics
The AI Trojan Horse
Everyone’s worried about AI generating convincing phishing emails, but they’re missing the real threat. Your shiny new AI features are actually massive data sieves, quietly hemorrhaging corporate secrets faster than a WikiLeaks dump. The only companies left standing will be those who encrypted their data before letting AI anywhere near it. – Ameesh Divatia, CEO of Baffle
AI’s Role in Precision Psychiatry
Advanced AI tools will enable hyper-personalized treatment plans for mental health, improving outcomes for TRD and MDD. – Dr. Hans Eriksson, Chief Medical Officer at HMNC Brain Health
Increased Focus on Copyright and Data Provenance
As AI-generated content proliferates, questions about copyright infringement and data lineage will intensify. The industry will see a push toward “provenance-aware” AI models that provide transparency about the sources of their training data. Legal and regulatory developments in this area will be closely monitored and could redefine how AI models are trained and deployed. – Greg Benson, Professor of Computer Science at the University of San Francisco and Chief Scientist at SnapLogic
The Race to “Inference on Laptop” Will Continue
Running models locally on devices like laptops or phones offers major benefits for cost savings and data privacy, spurring a race to create models lightweight enough for such applications. This goal is actively pursued in academia and industry, including by companies like Apple. Although a clear path may not emerge soon, achieving efficient on-device AI remains a top priority for the next generation of AI deployment. – Philip Derbeko, Head of AI at ControlUp
AI Adoption Will Lead to More Non-Human Identity Risk
AI adoption is creating new challenges when it comes to non-human identity management and security. A growing trend, termed “LLMJacking,” involves threat actors targeting machine identities with access to Large Language Models (LLMs), and either abusing this access themselves, or selling it to third parties. This threat will escalate in the year ahead, amplifying the need for robust non-human identity security measures. – Danny Brickman, CEO and Co-Founder of Oasis Security
Artificial Intelligence (AI) Will Enhance Business Operations, But Security Will Remain Crucial
In 2025, organizations will embrace AI-powered solutions across different business functions to increase productivity and speed decision-making. This new technology stack creates new attack surfaces and exposes organizations to previously unknown threats. To mitigate these new risks, security teams must adapt existing processes and controls, such as data access governance, privileged access management, and activity monitoring. – Ilia Sotnikov, Security Strategist at Netwrix
GenAI – Friend and Foe
The fiery AI hype cycle of 2024 will start to ebb as practitioners get their hands on AI-enhanced solutions with 2025 separating the marketing promise from the “boots on the ground” reality. In 2025, GenAI will begin to show real value in select areas of cybersecurity, but not in the pervasive ways some expect.
Machine learning driven anomaly detection has been integral in cyber for years, helping to discover unknown threats. In 2024, the focus was on GenAI and its potential for cyber security breakthroughs. Except, its focus on regurgitating and organizing all the knowledge collected in LLM’s cannot keep up with the real time, dynamic nature of cyber threats.
Next year AI breakthroughs will focus on security team productivity through the automation of lower value but highly time-consuming tasks. These less sexy successes will add real value and pave the way for further advancements that will include automation of decision making. Only the superior uses of the technology will see continued development into 2026. – Andy Grolnick, CEO of Graylog
AI Agents Will Change How We Interact With the Web
Over two years ago, ChatGPT burst onto the scene and changed the technology landscape. It also set us on a path towards a reimagining of the World Wide Web as we know it. In the next decade, the web will become the realm of AI agents that work on our behalf to handle tasks like scheduling appointments, making online purchases, paying bills, and more, freeing us up to step away from our screens and spend more time interacting face-to-face. 2025 will be the year we see the first steps towards this new normal. The chatbots of the last two years will evolve into basic AI agents that don’t just tell you how to navigate screens and menus on a website, but can also complete low-stakes tasks. For example, your healthcare provider’s chatbot, which today can tell you how to navigate the webpage to make an appointment, will soon be able to schedule that appointment. By embracing this shift, we stand to gain a future where technology fades into the background, letting us lead more fulfilling, present, and connected lives. – Robert Blumofe, CTO of Akamai
AI-Driven R&D as a Strategic Imperative
In 2025, enterprises will increasingly view AI-driven R&D as a strategic priority, with significant investments in tools that enhance developer productivity, such as AI-powered code generation and automated testing. The adoption of these solutions will amplify engineering capacity, forcing leaders to align product roadmaps with the newfound velocity. Companies that fail to adapt to this pace of innovation risk falling behind as competitors leverage AI to dominate market dynamics. – Justin Holtzinger, Chief Revenue Officer, DevOps at Cirata
Deepfakes Will Become Unrecognizable
In 2025, AI will become sophisticated enough that even AI experts may not be able to tell what’s real and what is manufactured. People will have to start asking themselves every time they see an image or watch a video: did what I’m seeing really happen? Unfortunately, this will be used by people will bad intentions, whether that’s a scorned ex-partner spreading rumors via fake photos on social media or governments manipulating entire populations and swaying votes by releasing videos that spread political misinformation. – Siggi Stefnisson, CTO at GenDigital
The Answers to Innovation Will Also Become the Biggest Financial Threats
Generative AI and large language models are revolutionizing business processes, empowering organizations to enhance communication, automate workflows, and extract insights like never before. However, this comes with a hefty price tag.
- Rising Cloud Costs: GenAI has contributed to a 30% surge in cloud expenses, with 92% of IT and finance leaders identifying AI-driven cloud costs as “unmanageable.” Without rigorous governance, GenAI could make innovation financially unsustainable.
- Operational Risks: As GenAI adoption grows, ensuring ethical use, avoiding bias, and implementing guardrails will become paramount to prevent reputational and financial damage.
Take Action: Implement strong FinOps practices to govern AI-related cloud costs, ensuring investments in GenAI yield tangible returns. Develop frameworks for ethical AI use, supported by continuous monitoring of cloud utilization. - Chris Ortbals, Chief Product Officer at Tangoe
Responsible AI Will Gain Importance
In 2025, I think we’ll see the topic of responsible AI gain importance, specifically due to an increase of public scrutiny around risks and remediation practices. Organizations will need to strike a careful balance over taking risks with AI and having rapid remediation strategies available. At the same time, we’ll see fraud techniques become more effective, and more precise at targeting aging boomers who control significant wealth. Businesses and brands will see dramatically increased risks, as bad actors using AI will launch convincing impersonation attacks, making it easier and more accurate than ever to fool customers. And unfortunately, disinformation will continue to proliferate, as social media platforms reduce their efforts to regulate its spread. Due to the documented increase in scams, phishing will be considered an expected cost of doing business and this will intensify the push from brands and regulators for improved protections. – Rod Schultz, CEO of Bolster
More AI News
