Technology as a Strategy: Key Skills for Future Business Leaders

Tech literacy is integral to strategic leadership. Regardless of whether you’re launching a new product, entering a new market or optimising internal processes, there is a significant chance you will find that cloud, data and machine learning will be an integral part of the solution.

The leaders that will succeed in such an environment will not be communicators or financial sharp shooters, but people who are technology savvy and can bridge strategy with engineering, business objectives with technical limitations.

You do not have to be a coder. However, you need to be conversant.

The quicker you can develop that fluency, the better prepared you will be to make the future of AI.

—

We are at a time when businesses operate on the cloud, decisions are made by algorithms and not understanding the mechanics of cloud computing systems is nothing short of a strategic flaw. Cloud, artificial intelligence and machine learning are no longer niche IT topics – they have permeated almost every aspect of decision-making and influence operations, marketing, finance, customer experience – you name it.

This also means aspiring professionals and future MBAs of the world must develop a working fluency with the machines that drive business. Coding is not a prerequisite, as you do not have to write Python or tune neural networks, but they must know how systems, pipelines, and architectures work. The potential to facilitate the connection between business plan and technical implementation is rapidly becoming the signature of successful managers.

This is what leadership skills in the AI era looks like.

Cloud Certification and the Fluency

The value of cloud fluency today extends far beyond the scope of scoring well in exams. Fundamentally, corporate executives need to learn the nitty-gritty of cloud computing, not only in the technical sense, but also economically and strategically. Provider certifications, such as AWS, Microsoft Azure or Google cloud can provide a good foundation. These prove executives are aware of cloud service models (IaaS, PaaS, SaaS), storage possibilities, security sets and deployment workflows.

Leaders also need to analyze cost models like the operational and capital equipment expenditure versus trade-offs with cloud adoption. They ought to learn integration complexity – the way new cloud services are hooked up to old systems, as well as the dangers of vendor lock-in between going with one cloud ecosystem as opposed to another. These are issues that are located squarely within the intersection of business and technology.

Understanding the Machine Learning Lifecycle

This is not limited to learning just the outputs of models. The need, for any good business leader, is to have a holistic understanding of the entire ML lifecycle. This requires leaders to familiarize themselves with aspects like data collecting and cleaning, training, validation, deployment and continual monitoring.

It is this understanding that will underpin how well managers know what is expected of them, how to distribute the right resources and pose the appropriate questions. Such as, for a model that is not performing well, is it bad training data or an inappropriate algorithm? Should the team focus more on feature engineering or think of collecting new data?

This situational awareness enhances a more productive synergy between business partners and data science groups – and results in superior overall decision-making.

MLOps Awareness

Machine learning operations, abbreviated to MLOps, refers to the collection of engineering philosophies that precondition ML models to scale, be secure and be maintained within a real-business context. It is the key bridge between data science and information technology operations.

Business leaders do not necessarily have to be able to write CI/CD pipelines but should be able to understand what a good deployment of ML should look like. This includes understanding:

• How models are tested before deployment
• How post-launch monitoring takes place
• How retraining workflows are initiated when performance drifts

Such awareness prevents cases where AI investments are only captivating prototypes and not working tools which provide long-term value. Knowledge of MLOps is also useful to prevent the most frequent mistake of creating a model that performs well in a laboratory and does not transfer to production.

Data Architecture Basics

Data is the fuel of modern businesses, yes, but the powerful assumption underpinning success is that it is stored, structured and accessed in the right way. Leaders are expected to understand the basics of data architecture – what a data warehouse and a data lake are, when to apply structured and unstructured datasheet formats and how columnar systems such as Parquet can perform differently from traditional row-based systems in terms of cost-effectiveness.

These choices underpin everything from real-time analytics abilities, cloud storage expenses, model training effectiveness, etc. Knowing them lets the managers consider trade-offs, advocate on the necessary infrastructure investments, and collaborate better with data engineers and architects.

Strategic Platform Selection

The cloud and AI ecosystem are tight, competitive and dynamic. The AWS, Azure, and Google Cloud have similar sets of capabilities, but with varying strengths, pricing models and paths of integration. The choice of platforms must be made strategically; the optic to assess it a business-oriented lens involving, but not limited to:

• How compatible the platform is with current systems and requirements compliance
• Long-run cost consequences of a certain provider over the other

This call is not merely a technical decision, it is the strategic imperative that can determine agility, cost-efficiency and supply of partners that will be in place over many years.

AI Problem-Framing

The skill of problem framing might just be the most underrated in tech leadership, i.e., how to identify the major business problems that can be solved with the help of AI.

While it is very easy to be seduced by the propaganda of the new algorithms or benchmarks of new up-and-coming models, AI self-sufficiency in the business context is not necessarily about the breaking of new technical advancement at every turn, but in the rates of incremental progress captured in such metrics as churn rate decrease, better forecasting, fraud detection and conversion.

The best AI leaders possess the skill of converting overall objectives (which are vague and imprecise) into a well-scoped ML problem. They can define business goals in terms of data sets, define success metrics in detail (accuracy, recall or ROI) and put all interested parties together on the same goal. In this, there is a blending of strategic and fluent technical thinking

Advanced Tooling (Optional, But Useful)

Moving even deeper into tooling is an opportunity to distinguish oneself among others in tech-oriented professions – product managers, digital transformation leaders, or startup founders. Knowledge of container orchestration tools such as Kubernetes or ML pipeline management such as Kubeflow or MLflow will carry a lot of weight. These tools are frequently employed when implementing and scaling complicated models to production environments.

Although it is not expected of every leader, understanding these marks a strong command of how ML systems are used on a massive scale and can help talk leaders have much more sophisticated and constructive conversations with engineering teams.

—

This article is the first of the two-part series, ‘Technology as a Strategy’. Read more on the Praxis Business School blog.