From 2024 to 2025: How Enterprise AI Moved from Experimentation to Execution

Comparing a year of transformation — what changed, what scaled, and what’s next

Oct 07, 2025

From Proof to Production: AI’s Real-World Momentum

A year ago, only a few organizations were running AI in production. Today, many have joined them, reshaping workflows, decisions, customer engagement, and even commerce. Between October 2024 and October 2025, AI moved from speculative pilots to functioning systems embedded in business operations. Organizations advanced from experimenting in labs to deploying value-driving models in production, especially in front office, operations, and customer domains.

Enterprise AI spending grew more than sixfold to $13.8 billion, and three-quarters of knowledge workers now use AI tools daily (Microsoft Work Trend Index 2024). Meanwhile, production deployments have scaled sharply. Bain & Company found that enterprise AI use cases in production doubled year over year (Bain & Company, 2025). Among OpenAI customers, 67 percent report running non-frontier models in production (a16z, 2025). Yet scaling remains uneven, with 42 percent of firms abandoning most AI projects before full deployment, up from 17 percent a year earlier (CIO Dive, 2025).

The difference between 2024 and 2025 is focus. Success now depends on embedding AI in measurable workflows rather than running endless pilots.

Financial Services — AI Becomes a Trusted Partner in Client and Investment Operations

In 2024, financial institutions used AI for narrow automation such as drafting reports, summarizing research, and testing proof-of-concept copilots. By 2025, those tools have matured into reliable systems supporting analysts, advisors, and client teams in daily decision making.

Morgan Stanley expanded its GPT-powered assistant to nearly every advisor, automating meeting notes, research lookups, and follow-up tasks. Advisors report saving 10–15 hours weekly, freeing time for strategic client work (OpenAI Case Study).
UBS integrated private-cloud copilots to condense market research and detect portfolio signals, reducing report preparation time by up to 80 percent.

Financial institutions have shifted from isolated experiments to AI systems that act as trusted operational partners, enhancing productivity while preserving the compliance, accuracy, and accountability critical to financial services.

Healthcare — From Data Analysis to Clinical Augmentation

Healthcare’s AI transformation is visible in how clinical teams interact with data. In 2024, most efforts focused on image analysis and risk prediction. In 2025, they have expanded to center on workflow automation and EHR integration, marking a pivotal year for physician-facing AI.

Stanford Health Care ChatEHR

Stanford Health Care’s award-winning ChatEHR lets clinicians query patient records in natural language, such as “show me the last three lab results for renal function,” and instantly retrieve summaries within Epic (Stanford Medicine News, 2025). Nigam Shah, Chief Data Science Officer, said, “ChatEHR reduces the administrative burden that contributes to clinician burnout.” Early results show 30–40 percent faster chart reviews and improved physician satisfaction (Health Journalism, 2025). Stanford will present this case study at AI Realized Summit 2025.

Mayo Clinic — Generative Documentation

Mayo Clinic piloted AI-generated clinical notes and discharge summaries that cut documentation time by roughly 30 percent, allowing clinicians to focus more on direct patient interaction.

“AI’s role now includes clinical augmentation as well as image analysis and risk prediction—accelerating workflows and elevating patient care.”
— Nigam Shah, Chief Data Science Officer, Stanford Health Care

AI’s role now includes clinical augmentation as well as image analysis and risk prediction, accelerating workflows and elevating the quality of patient care.

Customer Experience and Commerce — The Rise of Conversational Engagement

Conversational AI was one of the earliest production domains. In 2025, it evolved into agentic commerce.

OpenAI’s Instant Checkout feature now allows users to purchase directly through ChatGPT. Initially launched for U.S. users with Etsy integrations and planned expansion to Shopify, it merges discovery, suggestion, and purchase into one seamless flow. ChatGPT surfaces relevant products, and when a product supports Instant Checkout, users can tap “Buy,” confirm payment and shipping information, and complete the purchase without leaving the chat (Reuters, 2025).

This shift collapses the funnel: discovery, conversation, and transaction now happen in a single interface. For brands and product leaders, that demands rethinking search, personalization, and conversational design.

Legacy deployments like Compass’s AI-based support (65% one-touch resolution, 98% CSAT) and Erica’s billion-plus customer conversations laid the groundwork. Now, Fandom (CTO Adil Ajmal) and CarGurus (SVP Sarah Rich) are pushing the frontier further, using generative and predictive analytics in real time to power personalized fan engagement and product discovery.

Fandom’s predictive and generative ad system blends real-time behavioral analytics and generative targeting logic to align brands with fan conversations, achieving double-digit engagement gains. Fandom will present this system at AI Realized Summit.
CarGurus introduced an AI-powered conversational search experience that lets users describe their ideal car, for example, “SUV under $30 000 with high safety ratings,” instead of clicking through filters. Users spend 20–30 percent longer on site and convert at higher rates. CarGurus will also share its results at the Summit.

“AI is transforming how customers experience brands—making interactions more relevant, contextual, and human.”
— Zendesk 2025 Customer Experience Trends Report

The focus has shifted from efficiency to personalized discovery and relationship-driven engagement.

A dedicated panel at AI Realized Summit 2025 will explore how companies are managing customer relationships, using automation and personalization to create deeper engagement.

Life Sciences and Pharma — From Automation to Acceleration

In 2024, AI in life sciences was limited to automating documentation and data preparation. In 2025, it is accelerating the discovery pipeline itself.

Pfizer’s PACT initiative with AWS now spans 14 AI/ML projects optimizing drug development. The company has saved 16,000 hours of search time annually and cut infrastructure costs by 55 percent. Anastasia Christianson, Pfizer’s Head of AI/ML, explains, “AI lets us find patterns humans may miss, helping us make faster, smarter decisions” (AWS Case Study, 2025).
Johnson & Johnson uses AI to analyze surgical video data and monitor device performance through digital twins, improving predictive maintenance and clinical outcomes.

“AI has moved from automating workflows to accelerating science itself, compressing the time between hypothesis, experiment, and validated result.”
— Anastasia Christianson, Head of AI/ML, Pfizer

Operational Lighthouses That Scale Across Functions

According to BCG’s AI Value 2025 survey, only 4 percent of companies have achieved significant value from AI at scale. Those that did realized 1.5 times revenue growth and 1.6 times shareholder returns (BCG, 2025).

These early “lighthouse” projects, such as clinical documentation or claims processing, are being replicated across new domains, proving that disciplined scale, not experimentation, drives sustained advantage.

Leadership and Organizational Change — From Innovation Management to Enterprise Redesign

In 2024, AI leadership lived within the CIO or R&D function. By 2025, executive teams treat it as an organizational design challenge that determines how the business competes, operates, and learns.

Key shifts since 2024:

AI in the boardroom: AI strategy is now a standing topic in CEO and COO sessions, tied directly to long-term competitiveness.
Rise of the Chief AI Officer: CAIOs are being appointed not just to manage technology but to integrate AI strategy across product, operations, and culture, helping redesign how the organization creates and delivers value.
Cross-functional fusion teams: Business leaders and data scientists now work side-by-side to test and scale initiatives rapidly while embedding governance from the start.
Workforce evolution: More than 78 percent of employees now use personal AI tools at work. Leading firms invest in training and transparent adoption programs to sustain trust and productivity.
Cultural maturity: The language has shifted from pilots and prototypes to platforms and productivity. As Walmart CEO Doug McMillon noted, “AI is going to change literally every job.”

Leaders are no longer managing AI as innovation; they are using it to redesign their operating models and workforce strategy for the decade ahead.

Governance and Risk — The New Discipline of AI Management

AI regulation and governance matured substantially between 2024 and 2025.

Policy landscape: The EU AI Act (effective August 2025) introduces transparency and documentation requirements for general-purpose models. In the U.S., the Colorado AI Act (effective February 2026) establishes standards for “high-risk” systems in employment and finance (European Commission, 2025).
Framework adoption: The NIST AI Risk Management Framework has emerged as the leading blueprint for responsible AI deployment, guiding model evaluation, monitoring, and bias control across industries (NIST, 2025).
Certification leaders: Zendesk achieved ISO 42001 certification, demonstrating formal, auditable governance of AI systems. Microsoft and SAP are now pursuing similar recognition.
Multi-model strategies: Roughly 37 percent of enterprises now run five or more AI models in production, balancing innovation with resilience and regulatory compliance.

AI governance in 2025 is an operational discipline grounded in standards, frameworks, and cross-functional accountability.

Conclusion — From Promise to Performance

The difference between 2024 and 2025 is clear: AI has moved from experimentation to execution. Companies like Stanford Health Care, Fandom, and CarGurus are demonstrating measurable impact and sharing what they’ve learned at AI Realized Summit 2025, not about potential but about performance.

The lesson from this year is straightforward: AI is no longer an experiment; it is a capability embedded in how leading organizations operate, innovate, and compete.

AI Realized Now | Shaping Enterprise AI Adoption

Discussion about this post

Ready for more?