This goes without saying but an AI strategy is nothing but a way to use AI so that you can improve the execution / impact of your stated business strategy.
That’s because AI, including generative AI and predictive analytics (both supervised and unsupervised), has the potential to really accelerate business results.
In this blog I’ll outline a simple methodology to develop a solid AI strategy that is aligned to your stated business goals.
That way you can maximize the return on your AI investments.
PS: If you already have AI use cases, here is a AI Idea Scorecard tool to score them to see if they should get the green light.
Alignment with Business Strategy
An AI strategy must be aligned with your business strategy to be effective.
What do I mean by it?
Let’s look at two examples when this alignment is obvious:
- Banking: Use predictive analytics to identify high-value customers, and use generative AI to tailor personalized offers, boosting customer retention and increasing cross-sell opportunities.
- Supply Chain: Employ AI to optimize inventory management and use generative AI to create attractive and customized product catalogues – there by reducing costs and improving sales.
And here are 2 examples when the alignment is wanting.
- Banking: An AI model is developed to predict stock market trends, but the bank’s strategy for this year focuses on expanding mortgage services, payments infrastructure, and reducing fraud (among other things).
- Supply Chain: An AI model is built to forecast fashion trends for new product design, but the company’s strategy for this year is to enhance inventory management, improve eCommerce capabilities, and improve return logistics (among other things).
In the examples above, we have excellent innovation, but that innovation is not aligned. As a result, we’ll end up spending a lot of cycles, and ultimately the ideas will not be operationalized.
This misalignment diverts resources from strategic goals.
Evolution of Business Strategy?
I said above that AI strategy should be aligned with business strategy.
But sometimes, new AI models and their feasibility may open up new possibilities to evolve or adapt our business strategy.
These new business capabilities are not obvious or possible without considering AI.
For instance,
- Tesla’s Approach to Autonomous Vehicles illustrates how AI in self-driving technology shifted their strategy from simply producing electric cars to becoming a leader in autonomous driving. This has potential now to disrupt currently known ride-hailing and mobility services.
- Spotify’s Music Discovery Model shows how AI-driven personalized recommendations enabled Spotify to become a key player in music discovery and artist promotion.
- Google’s DeepMind and Healthcare, shows us how AI advancements in early diagnosis and treatment recommendations opens new opportunities for product developments in health diagnostics.
These examples are all about how AI can catalyze fundamental changes in business strategy, enabling innovations that were once out of reach.
So as you think of your current business strategy, do bring up new things that can be done in the business.
AI Strategy Methodology Overview
Now let’s look at the methodology to adopt AI effectively to maximize business impact.
It has 3 parts:
- Empowered innovation & feasibility at the level it matters
- One key business leadership mandate
- One key technology leadership mandate
Let’s look at these in order.
1. Empowered innovation & feasibility
Empowering innovation at all levels means that all business units are tasked with adopting AI to further the stated business strategy as it applies to their business unit.
Each idea should be measured against quantifiable business benefits. These can include any of the following:
- Reduced costs
- Improved efficiency
- Enhanced customer retention
- Increased cross-sell
- More new customer acquisition
- Reduced time to market for new products
- Improved innovation through new or customized products and services
- and so on
Each idea should also be checked for feasibility. This may include:
- Availability of the right data
- Complexity
- Ability to operationalize the insights
- Compliance or privacy risks
- and so on.
Think of it as a 2×2 matrix of benefits vs feasibility.
Design thinking workshops are an excellent tool for ensuring discussions are customer- and user-centric rather than fragmented.
This approach helps focus on real needs and creates viable solutions.
2. Business Leadership Mandate
Creating a business-aligned AI strategy starts with understanding and clearly communicating the business strategy.
Include the AI mandate in the strategy communications along with projected financial information as appropriate.
In addition, for AI adoption, it is important to promote a culture of acceptance for AI and address fears related to automation. It’s an empirical fact that for every automation and transformation program, the biggest impediment comes from mis-aligned people.
Each leader will handle this differently, but a clear reskilling and upskilling charter can help. This ensures that talent and knowledge are retained within the enterprise. For instance, if you rely heavily on outsourcing, explore ways to mitigate internal impacts through more insourcing.
This approach can still reduce costs and amplify business benefits by empowering people rather than simply trying to power through.
Nothing damages momentum more than when employees fear for their future.
And empirically, this is a sign of missing business strategy mandates and communications, aka, orphaned programs.
3. Technology Leadership Mandate
Ensuring that innovation can be operationalized means raising both technology and data maturity of an organization.
The best AI ideas will struggle if it takes too long for creating the right infrastructure and platform.
The two key components are:
- Data strategy: we need the right data with low latency and high quality. Security and privacy risks should also be mitigated.
- App modernization: every application should be capable of using the insights. To a large extent it means having APIs and modular architectures. Different types of cloud adoption may also be important to improve agility. See my post on app modernization and use of AI.
Successful data strategy will also include MLOps (machine learning operations). This will ensure that the AI models are used effectively and updated at the right frequency.
A great way to look at MLOps is to visualize the entire 5 stage data lifecycle
- Data engineering (get the data efficiently, aka ETLs)
- Data lakes (storage)
- Dashboards & reporting
- AI Modeling (and refreshing the models periodically)
- Consumption of insights (e.g. with APIs)
At the tech leadership level, the importance of managing these elements well is crucial to avoid creating enterprise silos.
A systematic implementation of enterprise technology and data architecture is crucial for sustained success.
Take the Next Step
I hope this methodology was helpful. Coming up with the right AI use cases, and ensuring that tech/data strategy is up for it are two things under every tech practitioner’s control.
To help you prioritize and score your AI ideas, try the AI Project Scorecard: Do You Have the Green Light? quiz. This tool will guide you in evaluating your initiatives and determining their potential impact and feasibility.
Good luck in harness the full potential of AI to drive significant business impact.
Connect with me if you’d like to brainstorm.