AI Strategy: McKinsey

McKinsey seems extremely bullish on AI and how it will transform industries, as well as the economic value creation as a result of AI adoption. Their AI playbook talks about $9.5T to $15.4T economic impact!

Their AI strategy has 3 elements:

Value and Asses: Similar to INSEAD and Microsoft AI strategy, McKinsey recommends first looking at the potential opportunity and impact in your industry. They recommend sizing the business value and zeroing on specific areas of focus for AI adoption in your organization, leveraging their top level sizing data.

Execute:

This stage is further broken down into:

Aligning on strategy

Recommendations can be summarized as:

<A> Gain strong executive alignment

<B> Secure and allocate appropriate budget (up to 25% of IT spend)

<C>Allocate large portion of the spend to embed AI / Analytics into rest of the org (versus just building it in a silo)

<D> And get ready to execute 3 or more use cases in parallel (scale AI adoption across org).

Details can be found here.

Building tech and data capabilities

Recommendations can be summarized as:

<A> Define and align on a Data strategy (as a core pre-requirement to AI strategy)

<B> Define and execute on data governance across org

<C> Define analytics methodology and have constant ‘test’ / challenge methodology (A/B testing).

<D>Have deep Data science / analytics expertise (>25 per 1000 FTE), with clearly defined talent strategy that includes training, on-boarding, career paths, and integration with rest of the org.

You can read the details here.

Completing the last mile

Here the focus is on changing organizational decision making process to be ‘data and AI driven’. Recommendations here are appropriate budget and accountability process towards such a decision making.

You can read the details here.

Beware: Warning signs of AI program failure

McKinsey lists a few warning signs:

1. Executive vision and buy in. recommendation is to correct this with appropriate workshops and training and aligning across the executive team. There is a need for a champion role (such as CDO / CTO) to ensure this buy in.
2. Not defining clear ‘low hanging fruit’ use cases with clear business KPIs for success. While this is self explanatory, the right leadership is key in achieving this as it does need both external (AI / Tech) and internal (business) knowledge.
3. No AI / analytics strategy beyond a few use cases. This risk centers around scaling AI across the org to deliver the true gaming changing benefits that the investment justifies. After all the initial ‘low hanging fruit’ use cases would typically be incremental and not game changing.
4. Not having a talent strategy for AI / Analytics talent: Acquisition, Training, Growth as well as Integration across rest of the org.
5. Too much focus and effort on solving all data problems at once. A multi-year platform effort that has to be fully complete and touch across all organizations, should not be a prerequisite for delivering value with AI.
6. Opposite of the above situation is not having focus and investment for a small but proposed-built data platform to support new AI use cases that can be developed in tandem with legacy data platforms.
7. The ‘business KPIs’ that capture the impact of AI projects are not part of the CFO’s or CEO’s scorecard.
8. Last but not the least is not having focused efforts on social, ethical and regulatory impact of AI and Data priorities.

Overall has a lot of good insights on their AI approach that seems to be more focused on transforming existing large organizations through Data, Analytics and then AI in those steps.

Their overall executive guide to AI is a good place to ground business leaders on AI.