Making a big technology change can be overwhelming, so leaders often turn to major technology consultants and companies for guidance on how to start, avoid pitfalls, and approach the change strategically. While I've had to clean up messes left by some of these big technology consultants, I recognize their frameworks can be well-researched and useful for selling your transformation plan internally.
Ultimately, the right answer for your organization will depend on your culture, customers, strategy, and resources. The framework also isn't enough alignment to your organization's strategy, collaboration, resources and Consider these frameworks as inspiration and guidance rather than rigid dogma.
So I'll break them down and give you my take as someone with a lot of experience in transformation and AI jobs.
IBM's AI Ladder
Overview
The AI Ladder by IBM provides a structured, step-by-step approach to AI implementation, consisting of four main stages:
Collect: Making data simple and accessible
Organize: Creating a business-ready analytics foundation
Analyze: Building and scaling AI with trust and transparency
Infuse: Operationalizing AI throughout the business
My Take
This approach is very expensive and requires a sustained long-term effort, particularly for the Collect and Organize phases. It seems to target enterprise customers with vast scale and budgets. The cynical part of me thinks IBM is aiming for lucrative contracts to clean up your data mess. Does all your data need to be simple and accessible? Probably not, unless there is a clear path between that data and a use case. You can always go back and fix data issues later if a use case arises.
While I appreciate the focus on trust and transparency, it feels like a miss not to address talent and organizational concerns. Just following the AI Ladder wouldn’t be enough to truly transform a company.
Andrew Ng's AI Transformation Playbook
Overview
Andrew Ng, a professor and leader at Landing AI and Deep Learning, created a playbook offering five key steps:
Execute pilot projects to gain momentum
Build an in-house AI team
Provide broad AI training
Develop an AI strategy
Develop internal and external communications
My Take
This playbook is comprehensive and practical, drawing from Ng’s experience with the Google Brain team and Baidu AI Group. Ng suggests doing pilot projects, building a team, and training across the organization before developing a strategy so the strategy is based on the organization’s learnings and capabilities. He encourages companies to focus their AI strategies on their competitive advantage, develop a virtuous circle, and elevate data collection and procurement to the strategic level.
In my experience, leaders often underestimate the effort and time required to get the data in place for ML or AI applications. The virtuous cycle (similar to Jim Collins' flywheel concept from Good to Great) is powerful. I’ve seen how releasing a new feature can quickly lead to companies contributing new data, elevating the experience for everyone.
Ng addresses the skepticism of AI and the need to recruit internal support. The playbook identifies who needs training (executives, division leaders, engineers) and what the sessions should cover. He also emphasizes the importance of internal and external communications. Many corporations previously invested heavily in innovation labs, which often failed because they couldn’t transition from the lab to a business unit.
Ng argues for focusing on reskilling existing talent rather than seeking new AI talent. Given the high demand and price tag for AI talent, this approach makes sense, especially if you hire a few high-level people who are good at coaching and can bring others along. They will also have a better understanding of your business, data, and customers.
Microsoft's AI Foundations
Overview
Microsoft's AI framework emphasizes a holistic and systematic approach, focusing on several dimensions:
Business Strategy: Clearly defined objectives and use cases, with measurable AI value.
Technology Strategy: AI-ready application and data platform architecture.
AI Strategy and Experience: Customer-centric approach with the right models for the right use cases.
Organization and Culture: Strong operating model, leadership support, change management, and continuous learning.
AI Governance: Processes, controls, and accountability for data privacy and security.
My Take
This framework is less actionable than the others, despite outlining specific actions towards the end. It highlights the most important elements that need to be part of your plan or roadmap. It feels odd to have three types of strategies when there should be a single business strategy, with technology and AI serving that strategy. However, the substance of the document focuses on making smart technology and AI decisions strategically.
A significant part of Satya Nadella’s transformation of Microsoft was moving from a know-it-all culture to a learn-it-all culture. This spirit carries through to Microsoft’s approach to Responsible AI. Natasha Crampton, Chief Responsible AI Officer, writes, “A theme that is core to our responsible AI program and its evolution over time is the need to remain humble and learn constantly. Responsible AI is a journey, and it’s one that the entire company is on.”
BCG's Approach
Overview
BCG outlines three strategic initiatives to drive value and achieve end-to-end enterprise transformation:
DEPLOY: Utilizes off-the-shelf tools to boost workforce productivity by 10-15%, improve employee satisfaction, and generate excitement for broader AI change.
RESHAPE: Re-imagines functions through workflow re-engineering, driving 30-50% improvements in efficiency and effectiveness.
INVENT: Leverages AI to expand revenue streams and "invent before getting disrupted" by introducing new offers, services, and experiences to the market.
My Take
BCG is the only company focusing on workforce productivity. This approach creates the capacity and excitement for moving into AI. Often, leaders add another strategic priority without removing anything else, and they are surprised when it doesn’t get done.
BCG suggests a 10-20-70 approach focusing on algorithms (10%), tech & data (20%), and people & processes (70%). This strong focus on data, talent, and culture is commendable. However, this framework targets CEOs and doesn’t address technology, data, or infrastructure concerns in depth. BCG misses an opportunity to educate CEOs and other C-suite leaders about the level of investment and risk involved in these projects.
Towards the end, BCG poses some crucial questions:
How do I empower my C-suite to stay up to date with the rapidly evolving AI landscape?
How do I think about whether to build, buy, or partner for the models needed?
What kind of (new) data capabilities do I need to leapfrog my competition?
How do I simplify my legacy systems to adopt new AI tech stacks and platforms?
How do I control costs and ensure a return on investment going forward?
How do I mobilize senior leaders to embrace and actively champion our (Gen)AI ambition?
How must our roles, departments, and operating model adapt to capture value from AI?
What are the best ways to fill the AI talent gap within my organization (e.g., upskill, reskill)?
How do I effectively manage change across my enterprise?
How do I communicate our ambition and set the right expectations, while promoting trust and preventing misconceptions?
Conclusion
AI is evolving rapidly, presenting numerous opportunities, complexities, and risks. These frameworks offer valuable guidance on strategy, technology, data, responsibility, talent, communication, and culture. Choose achievable steps and keep learning to navigate your AI transformation successfully.
Image: An NPS ranger let me look into the very top of the Bodie Island Lighthouse in North Carolina. Crazy to think it uses a small 1000 watt light blub. It was great getting to learn about the martime history and ecology of the Outer Banks while the girls worked on their Junior Ranger badges.
Opmerkingen