2025: The Year of Breakthrough AI? Tech Leaders Predict the Next Stages of Enterprise AI

AI’s influence is being felt everywhere, from chatbots that help people shop online to talk of autonomous drones or other Internet of Things devices that can act independently. 2025 shows no signs of slowing what organizations may want to tap into or create AI solutions for.

According to Ville Somppi, senior vice president, industry solutions, M-Files, in 2025, many companies will move beyond the exploratory phase of generative AI adoption, gaining a clearer understanding of where it can drive value and where it falls short.

“As large language models continue to evolve, becoming smarter and more affordable, organizations will experience improved consistency and quality in AI outputs. This affordability will open new use cases previously unconsidered. Building on this, enterprises will increasingly focus on targeted, strategic AI use cases, refining their automation strategies. Successful organizations will recognize when AI can fully automate tasks and when it should support human work, shifting toward smarter, more practical applications that deliver measurable impact,” Somppi explained.

Ethics in AI will take a step forward in 2025, predicts Stephen Manley, CTO of Druva, as businesses will choose to lean into transparency and ethical AI to win over customers in an age of information chaos.

“In 2025, geopolitical turbulence will continue, and misinformation is likely to abound. It’s unlikely that new data privacy and AI policies will be passed and enforced in 2025, so customers will expect businesses to take responsibility for ethics in AI,” Manley said. “As companies incorporate AI into their products, they have a responsibility to protect what and how the AI uses customer data, especially as it relates to sensitive data. Businesses must invest in ethical AI development, with an emphasis on transparency because AI adoption will directly correlate to the amount of trust the customers have in it.”

Baris Gultekin, head of AI, Snowflake, believes AI “backlash” or negativity will be mitigated one successful use case at a time.
AI hallucinations are the biggest barrier to getting generative AI tools in front of end users. Right now, a lot of generative AI is being deployed for internal use cases only because it’s still challenging for organizations to control exactly what the model is going to say, and to ensure that the results are accurate, Gultekin said.

“However, there will be improvements, especially in terms of keeping AI outputs within acceptable boundaries. For example, organizations can now run guardrails on the output of these models to constrain what generative AI can or can’t say, what tone is or isn’t allowed, etc. Models increasingly understand these guardrails, and they can be tuned to protect against things like bias. In addition to establishing guardrails, access to more data, to diverse data, and to more relevant sources will improve AI accuracy,” Gultekin noted.

Ryan Janssen, CEO, Zenlytic, offers similar sentiments with AI renewing the focus on data quality, for two reasons, “First, high quality data is required for training and fine-tuning models. Second, AI-powered analytics tools will offer a higher-resolution view of data, revealing previously undetected quality issues.”

Susan Haller, senior director of advanced analytics, SAS, thinks 2025 will be the year synthetic data begins to shine. As more organizations discover the incredible potential of synthetic data—data that is statistically congruent with real-world data without resorting to manual collection or purchased third-party data—the perception of this technology will inevitably shift. Making the generation of synthetic data more accessible across a range of industries, from healthcare to manufacturing, will prove to be a significant strategic advantage.

“The future possibilities for leveraging this type of data are endless,” Haller said.

Nick Durkin, field CTO at Harness sees the hype around agentic AI reaching a boiling point in 2025, as it overtakes GenAI at the top of the agenda for software development leaders. Organizations will increasingly use agentic AI frameworks to streamline the software delivery lifecycle, giving rise to the next iteration of DevOps automation.

“Instead of having one, basic GenAI assistant, organizations will bundle together specialized agents for the different stages of software delivery—such as code generation, testing, and quality assurance. One agent will act as the orchestrator, directing the others and producing more accurate insights into the end-to-end process. This will drive a more intelligent approach to DevOps automation, with AI agents able to make more effective decisions, based on a holistic understanding of the processes they support,” said Durkin.

However, the trough of disillusionment looms for GenAI, and the request for ROI will quicken the industry’s descent into said trough.
“Every business is striving to understand the impact of GenAI, and savvy business leaders are already asking questions around accuracy, efficiency, and outcome to validate the IT spend allocated to it. Unless it’s incorporated into a purpose-built tool from the ground up, GenAI won’t drive significant measurable efficiency and many will feel let down by its initial promises,” said Manley.

Prince Kohli, chief technology officer at Automation Anywhere, predicts that in 2025, the number of companies claiming to offer AI agents will surge 100-fold, yet far fewer will have customers that capture real and measurable value from these solutions.

“As foundation models commoditize, enterprise interest in yet-another-LLM will decrease significantly. The focus will instead shift to deploying generative AI effectively through well-planned and projected high-value implementations. Only those investing in cohesive, process-driven AI deployments will unlock the full potential of AI agents, setting a new standard that distinguishes genuine innovation from mere claims,” Kohli said.

By now, it’s common knowledge that AI has significant energy demands that are putting a strain on the grid and data centers. But not all AI’s power demands are the same. Heiko Claussen, chief technology officer, Aspen Technology said industrial AI will gain attention for its energy efficiency benefits.

“For example, industrial AI, which combines domain expertise and engineering fundamentals with AI, is more narrowly focused with smaller models, which means it is more data efficient and requires far less power. In addition to its energy efficiency benefits, industrial companies are turning to Industrial AI for agility, guidance, and automation, such as helping companies avoid unplanned plan shutdowns or enhancing operational decision support for users,” Claussen said.