Unfiltered Stories | Moving Beyond the Hype to Create Value from Generative AI

Summary

The conversation explores the rise of generative AI in enterprises, highlighting challenges and opportunities for scaling its adoption. Matt Candy from IBM shares five key foundations for successful AI implementation: value identification, AI control centers, data foundations, transformation operating models, and reskilling the workforce. The discussion emphasizes the need to simplify tech stacks, tackle organizational debt, and address cultural barriers. Candy also explains IBM’s internal AI transformation, from HR to marketing, and stresses that the biggest challenge is not the technology itself, but the need for people to embrace and adopt it at scale.

Takeaways

• 💡 Generative AI has rapidly transformed enterprise discussions, shifting from 'what is it?' to 'how do we scale it?'
• 📈 Companies that lay a solid groundwork for scaling AI—through strategy, value identification, and ROI frameworks—tend to succeed faster.
• 🔑 A key to scaling AI is building an 'AI control center,' ensuring security, governance, and seamless integration of proprietary and open-source models.
• 📊 Data foundations are crucial; organizations must integrate proprietary enterprise data with AI models to unlock true value.
• 🚀 The transformation model for organizations needs to be agile, experience-led, and iterative to accelerate AI adoption.
• 👩‍🏫 People and culture are central to scaling AI. Organizations must focus on reskilling and change management to overcome fears or misunderstandings about AI.
• 📊 IBM itself has transformed through AI, especially in HR, where 94% of employee interactions now happen via digital AI solutions.
• 🤖 AI-powered tools are used across IBM’s departments, including marketing (for content generation) and consulting (for task assistance and collaboration).
• 🛠 Enterprises face various types of 'debt'—tech, data, process, skills, and cultural—that hinder AI adoption. AI can help mask some of this debt while organizations work on remediation.
• 💭 The biggest wish for AI adoption? Getting people more comfortable and hands-on with AI to accelerate its widespread use and impact in the workplace.

Q & A

What are the main challenges enterprises face with generative AI, according to the speaker?

The main challenges enterprises face include understanding how to make generative AI real and scalable, moving beyond proof of concepts (POCs), and embedding AI into workflows with clear ROI tiebacks.

What are the five key factors for scaling generative AI in enterprises?

The five key factors for scaling generative AI are: 1) Strategy and value identification, 2) AI control center and governance, 3) Data foundation, 4) Transformation operating model, and 5) People and reskilling.

How is IBM applying generative AI within its HR department?

IBM has implemented a digital-first AI solution in HR, allowing 94% of employee-manager interactions to be handled through AI. This has improved efficiency, saved time, and increased employee NPS.

What role does IBM's partnership with Adobe's Firefly platform play in their marketing function?

IBM uses Adobe’s Firefly platform for image generation and has created a brand-specific large language model. This enhances content generation with AI, driving faster content delivery, personalization, and lower costs.

What is IBM's Consulting Advantage Platform and how does it aid consultants?

The Consulting Advantage Platform serves as an AI control layer that wraps around various AI models. It helps consultants by providing AI assistants for tasks such as user story generation, business requirement definitions, and code conversions.

What types of organizational debt are discussed in the script, and how can generative AI help address them?

The script mentions technological, data, process, skills, and cultural debt. Generative AI, particularly through natural language interfaces, can help mask and gradually address these debts while simplifying back-office operations.

How can enterprises reduce skills debt with generative AI?

Enterprises need to invest heavily in reskilling their workforce to use generative AI. This involves hands-on learning and ensuring that AI tools are democratized across the organization, not just limited to the IT or AI teams.

What is Matt Candy’s biggest wish regarding generative AI adoption?

Matt Candy’s biggest wish is to accelerate the adoption of generative AI across enterprises, focusing on getting people comfortable with the technology, removing fear, and fostering a culture of innovation.

Why is generative AI considered more of a 'people issue' than a 'technology issue'?

Candy believes that the success of generative AI hinges on getting employees to understand and adopt the technology. The challenge is not with the algorithms, but in overcoming cultural and behavioral barriers within organizations.

What does 'death by a thousand POCs' refer to, and why is it a problem?

'Death by a thousand POCs' refers to the proliferation of proof of concept projects that do not scale or deliver real business impact. It’s a problem because it prevents enterprises from fully leveraging AI’s potential by keeping initiatives in experimentation mode.