Before getting into the details of an Artificial Intelligence (AI) Strategy, let’s first set the foundation to understand what a strategy is. A strategy is the roadmap of the enterprise goals or tasks that a business must achieve.
But what are these goals, and how do enterprises prioritize among multiple initiatives to meet the end goal?
This post exemplifies what an AI strategy is and why it is needed. You will also learn how to take an approach from design to implementation. Further, we’ll discuss key factors an enterprise must consider before starting an AI initiative and how to focus on people, processes, and technology to see your AI strategy come to life.
Why Enterprises Need an AI Strategy
Enterprises can not depend on wait-and-watch or hit-and-trial to understand the genuine rewards of AI solutions. The pandemic has brought rapid digitalization, generating valuable data to tap and convert into business-ready and actionable insights. The user preferences are changing rapidly, creating a need for scalable AI solutions that can develop near-to real-time insights.
Akin to the challenges and woes accompanying the adoption of any new technology, many enterprises are struggling with failing to conceive the right AI strategy for their business.
A lot of things can go wrong without proper AI planning:
- Choosing the wrong projects
- Difficulty in new technology adoption
- Lots of trials but no reward
- Not knowing where to start
- Lack of evangelizers supporting new technology
Do You Require an AI Strategy?
AI can be a core enabler of transforming and pivoting business into a foray of success. No wonder AI has table stakes in many board discussions and plays a significant role in defining business strategy and setting the direction.
If you are already an AI-first enterprise in the industry, it strengthens your position as a leader and industry pioneer. But, if you have not yet thought of leveraging AI algorithms' prowess, you will conceive one sooner to combat the competition.
Designing a successful strategy is a challenging task that becomes tougher when done specifically for AI projects.
AI projects are iterative by nature and are dependent on extensive cross-collaboration. Extending from the experimentative nature of AI, even the strategic roadmap for AI initiatives goes through multiple iterations and requires support from AI endorsers in the organization. This gives rise to the role of an AI strategist who lays down the roadmap of AI products and services for the organization.
An AI Strategist possesses core project management skills and has a working knowledge of AI workflows, algorithms, and processes. Additionally, AI strategists also consider ethical concerns for each initiative that is baked into the strategic business goals.
Choosing the Right AI Project
The starting point is always tricky - it generally springs from a high-level overview of the critical business pain points and how they can be solved.
Listing down the potential pathways to achieve the business requirements is the fundamental step in charting the AI journey.
Once you have identified the viable ways to meet the business prerogative, you should
answer the following questions for each solution:
- What is the cost of developing this solution?
- Is the solution scalable?
- Does it require scratch-up groundwork, or can it be built upon existing solutions?
- How to decide whether to build in-house or buy from vendors?
- Where does the solution add value? Does it bring operational efficiencies, or has customer focus the prime objective?
- What does success look like? What are the KPIs to measure the results of my AI initiative? For example, does the business plan to improve the user footfall on their web property i.e. clicks, or does its aim to increase conversions?
Source: Image by jcomp on Freepik
To qualify for an AI-based solution, you would need to find the answers to the supplementary questions stated below:
- Does it have pattern and quality data to learn those patterns from?
- Does business agree with AI's probabilistic solution as compared to a deterministic rule-based solution?
- What are the ethical concerns?
- Does business feel comfortable with the black-box nature of algorithms?
- What is a good evaluation metric? A machine learning metric should not just be a scientific metric reflecting the goodness of a model but must tie back to the business objective. For example, how do clicks or conversions increase with a model having 80% accuracy?
These questions form a good starting point to weed out the ineligible options of different AI solutions at first. Secondly, the leaders from different teams come together to collectively call on the trade-offs between the selected options and hash out the priority order of the other objectives zeroed in for the year.
Considerations When Implementing AI Initiatives
The strategy extends beyond just setting goals and requires a clear roadmap to achieving those goals. It must start with a clear plan, objective, and action items.
Once you have decided on an AI initiative, your next goal should be to find answers to the following questions to accelerate the implementation journey:
- What are potential risks and how can we mitigate them?
- Not knowing the unknowns - can they act as roadblocks that can doom the AI project?
- What kind of cross-team collaboration is required - what resources are needed to take the project to successful execution?
- Who is the orchestrator and will own the accountability if something goes south?
- Is the strategist AI-aware to design it in the first place?
- Who are the consumers of AI applications?
Solid strategy execution requires a clear focus on people, processes, and technology. Is it forward-looking, or is it coming out of a reactive stimulus driven by what other players in the market are doing?
Bringing Together People, Process, and Technology
Successful AI adoption and digital transformation sit on three essential pillars: people, process, and technology:
It is imperative to know what skill sets are required to execute the proposed strategy. Enterprises must take stock of what skills are already present in-house or need to hire from outside. The analysis brings clarity on whether the initiative requires specific expertise or a generalist’s skills. In essence, rightly skilled people with vast backgrounds and experience are fundamental in taking the next step towards strategy execution.
Hiring the right talent helps you in your next steps, i.e., process and technology. Processes define the ‘how’ part of execution and are iterative by nature. You need to set up the processes that produce the same results and are agnostic of who executes them. One standard practice to embed processes in organizational culture is to document and communicate key steps to ensure that everyone in the team is using the same language.
Organizations must keep themselves technologically updated. Obsolete technology or failure to adapt to changes in technological advancements can adversely impact the productivity of the people. Regardless of how skilled the people are or how effective the policies are, they prove to be counterproductive if the organization is not able to keep up with the advances in technology.
In short, leadership support is the common thread that connects people, process, and technology as tersely highlighted below:
"You need people who endorse the process and adopt the latest technology with the help of leadership support to develop a successful AI solution."
This post highlighted the importance of designing an effective AI strategy similar to a business strategy. It covered the main concerns that need to be considered in an AI strategy such as the availability of data, the presence of repeatable patterns, and ethical concerns.
Further, the role of AI strategists was discussed and what factors they contemplate when stack-ordering the key AI initiatives among many other promising projects. The article then concludes by sharing strategic questions that can help an enterprise prepare for the potential risks of building AI solutions.
Learn more about AI best practices and challenges in the enterprise at Data Science Salon Austin on February 21-22, 2023! The intimate event curates data science sessions to bring industry leaders and specialists face-to-face to educate each other on innovative solutions in AI, machine learning, predictive analytics and acceptance around best practices. We are almost sold out, but you can still apply to attend HERE! See you in Austin!