How To Build An AI Action Plan

AI Action Plans create the first steps for a business to accelerate its AI strategy and product roadmap. Most people’s biggest gripe about strategy is it never leads anywhere, so the AI Action Plan is a litmus test. A data and AI strategy is only valuable when it is actionable and informs decision-making across the business.

However, CEOs have realized that their companies aren’t transforming, innovating, and delivering fast enough to survive. They need strategy planning and implementation to run parallel with execution. That’s where an AI Action Plan adds value. We added the AI Action Plan to our frameworks in 2021, and it has been a success for V Squared’s clients.

For mature businesses with a data and AI strategy, it’s a way to quickly pivot from planning to action. For early-maturity businesses and startups, it’s a way to compress strategy planning, implementation, and execution into a much shorter cycle.

Decisions connect strategy with implementation, and carrying out those decisions connects strategy with execution. If developing an AI Action Plan isn’t a simple extension of the data and AI strategy, it’s time to go back and fix the disconnect. A single framework for decisions about AI is crucial for enterprise-wide alignment.

We must enable every part of the business to own the decisions they’re accountable for, or data and AI will get stuck in the technical team’s silo. Still, decisions made in different parts of the business must align, or the result is business units moving in different directions. Alignment requires common goals, decision frameworks, context, and transparency.

AI Action Plans also help businesses avoid chasing shiny objects or getting caught up in the hype cycle. They support building value and ensuring that AI initiatives are not just experiments but drivers of tangible business outcomes. I’ll start this article with advice about developing an AI Action Plan and provide a template you can follow to create one. The article is based on my experience implementing AI Action Plans and leverages the frameworks I teach in my certification courses.

Begin With Pragmatic AI Opportunities & Business Realities

Data and AI strategies exist in messy, complicated, real-world challenges. The AI Action Plan must work under imperfect conditions and meet the business where it is. What are the pain points? Where are the gaps? What's your current data maturity? Don't skip the assessment step, or you’ll end up with an AI solution looking for a problem or something that isn’t feasible and pragmatic.

In the assessment phase, you start cataloging your data, evaluating your current workflows, and narrowing down where AI can make a difference. Once you have that foundation, it's time to start identifying opportunities. Avoiding the "let’s use AI for everything" approach requires finding those specific use cases where AI can deliver value that other technologies can’t.

Building an opportunity pipeline with 3 or more strategic opportunities is essential. Too many opportunities and the business spreads its limited resources too thin. Too few, and the business will continue to work on opportunities that don’t pan out because it doesn’t see any other choice.

Businesses must focus on problems where people and digital solutions perform poorly, where you can increase efficiency, or where you can create entirely new value propositions. What can AI do that you couldn’t do before? Which opportunities will customers care about? I’ve seen so many companies get bogged down in long-term AI projects that never see the light of day because the opportunities they chose were incremental improvements with little real value.

The AI Action Plan should have a 1:1 ratio of productivity and efficiency to revenue generating and growth initiatives. Another critical mistake I see repeated is focusing on AI as a productivity driver. Data and AI are new inputs for growth. Most of the value businesses realize will come from new features, products, and platforms.

The AI Action Plan must include guidance on how to make buy vs. build decisions, especially for internal operations use cases. The data and AI teams are finite resources and must be leveraged for the highest-value initiatives.

The Big Picture & Decision

Business leaders don’t feel like they can ask a nagging question. Why use AI in the first place? No one says this out loud, but everyone outside the technology organization thinks it. The AI Action Plan should connect opportunities with investment and transformation.

Business leaders must own the decision to invest in data and AI because they own the consequences. I use a framework called “The Big Picture” to showcase the layers involved in strategy implementation and execution. The image below shows one part of the framework.

The top two levels are some of the pieces required for holistic, enterprise-wide data and AI strategy planning and implementation. Both layers must align with technology (layer 3 is part of the data capabilities maturity model) and business transformation (layers 4 and 5). Layer 6 at the bottom shows the connection to the big decision.

Competitive maintenance is where most businesses get stuck. Technology keeps the business at the same level as its peers. The business doesn’t know how to use technology as a differentiator, so it isn’t more efficient than competitors. It also doesn’t leverage technology to deliver features or products that its competitors can’t. Most technology organizations are cost centers until the business learns to leverage technology for competitive advantage, innovation, etc.

Business leaders must make this decision, and the AI Action Plan must support them. Once they have, the AI Action Plan explains the opportunities and finally answers the question, “What’s the point of all this technology?”

Incremental Delivery & Estimating ROI

Incremental delivery, delivering value quarter to quarter, is critical. The Big Picture’s big initiatives must be broken down into smaller, more manageable chunks. Early maturity businesses must start with the basics, like a simple data product, and incrementally add more advanced AI capabilities as the business matures. Ensure that each phase of the product roadmap builds upon the last, creating a solid foundation for the next phase of AI deployment.

The AI Action Plan must detail a framework for quantifying the value of each feature. Here’s where the rubber meets the road: you must measure the impact upfront and validate it after each delivery. Business leaders aren't interested in AI for AI’s sake anymore. They are accountable for top and bottom-line impacts, so we must quantify AI's impact on business goals.

The product roadmap must deliver value every 8-12 weeks, or the business will lose interest. When that happens, initiatives lose momentum, and investment goes elsewhere. The AI Action Plan must define how the product roadmap serves the CEO’s need for immediate returns with their need for long-term, large growth drivers.

Too often, I see businesses jumping headfirst into AI initiatives without a clear plan that aligns features with products or explains the larger AI platform vision. That's a recipe for disaster, and a robust AI action plan is the antidote to this chaos. The feature, product, and platform approach ensures that every step with AI is intentional, value-driven, and contributes to achieving the business’s goals.

Transformation Strategy

People are an essential part of any AI strategy. That means investing in training and upskilling employees across the enterprise, not just the technical organization. External business units will need budget and time in their schedules specifically allocated to the activities that support transformation.

I have seen businesses make the mistake of committing to transformation but assuming its scope is limited to the technical organization. When external teams don’t have time and incentives to change, transformation won’t happen. Leaders across the business need authority and resources to drive initiatives forward.

The AI Action Plan should give people the autonomy to act. Embracing an innovation culture requires experimenting, iterating, and learning. Few business units are set up this way, and learning from failures is typically punished for the failure, not rewarded for the learning. It won’t happen unless resources are earmarked for transferring ownership of innovation cycles to front-line teams.

AI should be integrated into all aspects of the business, not just siloed in the data and AI teams. AI strategy informs decision-making across all levels of the company, so the AI Action Plan must be relevant and actionable across the enterprise as well. Strategy flattens decision-making, and the AI Action Plan should prevent bureaucracy and unnecessary committees.

The AI Action Plan Template

The template isn’t a rigid set of steps or actions. Strategy reveals multiple paths to success, and the AI Action Plan supports leaders across the business as they work to choose the best one. Frameworks are dynamic and adaptive. The AI Action Plan must be pragmatic, not aspirational. If it only looks good on paper, it will fail to deliver. Adapt the template to fit the business, not the other way around.

Assess the Business's Current State

Begin by understanding the business's history, context, needs, strengths, weaknesses, and gaps. This involves activities like cataloging data and listing current and potential use cases. This section should detail the findings of your initial assessment.

Identify Opportunities

Look for opportunities where AI can support strategic goals. Write up high-value opportunities, select appropriate use cases, and put down as much detail about the corresponding workflows as possible. Define a value-based prioritization framework. Focus on use cases that AI makes possible for the first time and that customers will perceive as valuable.

Define Use Cases and Problem Spaces

Decide which technologies fit best based on the problem space. Map the problem space to the workflow and define success with outcome metrics. Evaluate multiple technical solutions based on costs and impact on outcome metrics. Detail the tradeoffs and write up a value-based justification for the use cases where data and AI deliver value more efficiently than the alternatives.

Develop a Product Roadmap

Create a product roadmap that delivers value every quarter. The roadmap should break initiatives down and define the path from data to AI products. AI products should be built incrementally, with each phase delivering value and laying the foundation for the next. This could include starting with simple data products, adding descriptive models, and eventually incorporating more advanced AI capabilities as the business matures.

Quantify AI's Impact

Measure AI's impact on business goals with quantifiable metrics. Avoid metrics like customer satisfaction and productivity. Demonstrating tangible business impacts using the same KPIs that executive leaders do rather than generic metrics. Make the connection to financial metrics to close the loop.

Incorporate AI Value Governance

Use an AI Value Governance framework to define AI projects from a value creation perspective. I use a three-phased framework. Productization defines the need and potential value, commercialization defines the solution and costs, and monetization assesses the feasibility of realizing the project's expected returns. Detail those in the AI Action Plan.

Align with Business and Operating Models

Detail AI initiatives’ alignment with the business’s strategic goals and operating model. Ensure the AI strategy informs decision-making across the business, not just in the data team. Make this section of the AI Action Plan relevant to the teams and products that AI will benefit most. If the AI strategy isn’t actionable, go back and fill the gap. If you’re starting from the AI Action Plan, call out the steps that will be undertaken to deliver an AI strategy.

Monetization Strategy

The AI Action Plan should detail how AI products will be monetized, such as charging per use, offering subscriptions, or bundling AI features. Discuss current customer segments and how AI will open up new segments. Ensure that the monetization strategy aligns with the business model or, in more mature businesses, the business model transformation strategy.

Data Strategy

Explain why data is still a competitive advantage. In an early-maturity business, explain the steps required to develop a strong data strategy. It should include data gathering, curation, transformation, and information engineering plans. Explain how the business ensures data gathering aligns with customer needs, privacy, and ethics.

Internal Talent and Leadership

Justify investing in training and upskilling programs for employees. Empower leaders with the authority and resources they need to implement decisions, execute strategy, and drive initiatives.

Innovation Culture

Encourage experimentation, iteration, and learning. Develop a managed innovation process that balances exploration with exploitation. Use this part of the AI Action Plan to explain how you’ll implement both and what that means for the rest of the business. Justify the need for innovation to C-level leaders, and don’t forget to justify the investment to the CFO. Innovation can’t succeed without a budget, and you can’t get a large enough budget unless the CFO is bought in.

Holistic Approach

Ensure the AI strategy is comprehensive and addresses all aspects of the business that AI will impact. Call out those connections and the implications in every section of the AI Action Plan. Explain the alignment between the business and the data team.

Go-to-Market Strategy

Provide high-level details for the go-to-market strategy for AI products, including pricing, scaling, feature pruning, customer adoption, design, cost optimization, and ongoing product improvement.