Skip to content

Scaling AI Within Global Corporate Enterprise For Better Decision Making

QUICK SUMMARY

The session explores the challenges of scaling AI beyond proof of concepts in corporate environments, highlighting the need for both technological and organizational readiness. The speaker from Regeneron shares practical examples of AI implementation across various business functions, from omnichannel marketing to causal analytics and synthetic data generation. The presentation emphasizes that AI should augment human decision-making rather than replace it, requiring companies to develop a clear AI strategy that balances everyday efficiency gains with transformational opportunities.

KEY QUOTES

  • “People confuse scaling with the in-between moment, and the challenge we face is you don’t want to be at the end of the in-between moment because then you’re gonna be behind or out of business because there are fundamentally transformational things going on in this period.”
  • “You have to define your AI position, not on the basis of a project, but a vision. Because if you don’t do that, you are gonna constantly be pegged to funding within a project that will never scale.”
  • “For many years, the person who did the number crunching and the insight management was the one who was making the decision. AI is actually gonna allow us to separate those two.”

FULL SESSION SUMMARY

The AI Scaling Challenge

The session begins with the speaker acknowledging the significant presence of analytics and data science professionals in the audience. Working at Regeneron, a biotechnology company that recently acquired 23andMe, the speaker addresses a critical challenge in the AI space: the difficulty of scaling beyond proof of concepts. Despite clear innovation paths for most technologies, AI faces unique obstacles, with research showing that up to 80% of AI proof of concepts fail to scale, and 49% of companies question the value of their AI implementations.

The speaker introduces the concept of the “in-between moment” – the gap between when a transformative technology is demonstrated and when it’s widely adopted. This phenomenon isn’t unique to AI; it happened with the light bulb and other innovations. However, the speaker emphasizes that companies often mistake this in-between period as mere hype rather than recognizing it as a crucial transition phase. Those who wait until the end of this period risk falling behind competitors or becoming obsolete.

From Proof of Concept to Organizational Readiness

A key message is that organizations get caught up in proof of concepts without considering the broader implementation strategy. Successful AI adoption requires both technological readiness (infrastructure, tools) and organizational readiness (training, stakeholder understanding, consumption methods). The speaker notes that leadership often lacks technical understanding of AI concepts, leading to vague discussions about “AI” that create misaligned expectations and overemphasis on risks rather than benefits.

The speaker recommends using Gartner’s AI Ambition framework, which categorizes AI initiatives into three types:

  1. Everyday AI (Enable) – Providing bottom-line efficiencies
  2. Middle-ground AI (Extend) – Offering both top and bottom-line benefits through improved analytics and decision-making
  3. Game-changing AI (Upend) – Transformational applications that fundamentally change business models

Regeneron’s AI Implementation Examples

The presentation showcases several practical AI applications at Regeneron:

  1. Omnichannel Marketing and Customer Orchestration: AI helps manage real-time information flow to physicians, patients, and payers, creating more personalized customer journeys. This application is scaling across all brands and becoming a standard way of working.
  2. Conversational Analytics: Natural language interfaces allow non-technical stakeholders to query data without coding knowledge. The system not only answers initial questions but suggests follow-up queries based on context, creating an “immersive experience” that’s changing how people interact with analytics.
  3. NLP for Text Summarization: AI summarizes textual feedback efficiently, saving days of processing time while preserving the need for human interpretation of results.
  4. Synthetic Data Generation: In healthcare, privacy restrictions limit access to patient data like eye scans. Synthetic data technology creates new datasets that preserve the relationships and context of original data without privacy concerns, opening new analytical possibilities.
  5. Causal AI: Moving beyond correlation to understand system dynamics, causal AI helps Regeneron understand the complex journey of biologics from prescription to patient. This enables more targeted interventions and efficient resource allocation.
  6. Multimodal Analytics: The company is exploring capabilities to analyze images, video, and other rich media to better understand consumer sentiment and behavior beyond text-based insights.

Demonstrating Value to Leadership

The speaker emphasizes the importance of translating AI initiatives into business value. They share that they regularly review a chart with their CFO showing how each AI initiative impacts business metrics. This approach helps leadership see the return on AI investments and builds support for scaling efforts.

Human Judgment Remains Essential

The session concludes with an important distinction: while AI can help mine insights and understand customer needs, human judgment remains essential for decision-making. The speaker demonstrates this with a simple test showing how humans understand context in ways that large language models cannot, reinforcing that AI should augment rather than replace human decision-makers.

KEY TAKEAWAYS

  1. Successful AI scaling requires both technological infrastructure and organizational readiness, including stakeholder education and clear communication about capabilities and limitations.
  2. Companies should develop a comprehensive AI strategy that balances “everyday AI” for efficiency gains with more transformational applications, rather than pursuing disconnected proof of concepts.
  3. AI should be positioned as augmenting human decision-making, not replacing it, with a focus on how it can help separate insight generation from judgment-based decisions.

Delivery on Event Focus:
Aligning Innovation with Business Strategy

This session directly addresses the focus of aligning innovation with business strategy by demonstrating how AI initiatives must be tied to specific business outcomes. The speaker’s emphasis on creating a vision-based AI strategy rather than project-based implementations shows how technology can be aligned with broader business goals. The presentation of a business impact chart used with the CFO illustrates how innovation leaders can translate technical capabilities into language that resonates with business leadership.

Delivery on Event Theme:
Harvesting Innovation and Sowing the Seeds of Future Growth

The session supports the theme of “harvesting innovation and sowing seeds of future growth” by showing how companies can balance immediate AI applications (harvesting) with investments in more transformational capabilities like synthetic data and multimodal analytics (sowing seeds). The speaker’s discussion of the “in-between moment” highlights the importance of investing in emerging technologies before they become mainstream to ensure future competitiveness and growth.

Action Steps for Innovation Experts and Corporate Changemakers

  1. Develop a comprehensive AI strategy using frameworks like Gartner’s AI Ambition to balance everyday efficiency gains with transformational opportunities.
  2. Build organizational readiness by educating stakeholders about AI capabilities and limitations in accessible language, avoiding technical jargon.
  3. Create clear business value metrics for AI initiatives that can be shared with leadership to demonstrate ROI and secure ongoing support.
  4. Start exploring transformational AI applications like synthetic data and causal AI now, even if full implementation is years away, to avoid falling behind during the “in-between moment.”
  5. Position AI as augmenting human capabilities rather than replacing them, focusing on how it can help separate data processing from judgment-based decisions.