Every Business Now Needs an AI Strategy. Most Don’t Have One

By John Samuel, Founder and Director of Seventh State

AI is moving beyond the experimentation phase and becoming embedded in how business is done. Even companies without an explicit strategy are using it, whether they’re investing in systems and tools, or employees are using them independently without any top-down initiatives in place.

With many organisations now building AI into customer service activities and operations, businesses without a defined strategy run the risk of falling behind their competitors. Not just from a technological perspective, but from the negative impact of fragmented decision making, caused by the lack of defined business framework.

AI is now a critical strategic concern

AI deployment has passed the stage of being a technology initiative – it’s migrated into a critical strategic concern for every type of business. For many observers, it’s gone beyond IT and become a question of leadership, with boardrooms needing to understand AI implications even if they aren’t technical experts. A PWC report found that only 35% of boards have incorporated AI and GenAI into their oversight roles, meaning the majority are still overlooking something that’s becoming as fundamental as cloud, cybersecurity and digital transformation strategies.

The importance of organisational fit

As with any decision, technology can become expensive when adopted without a clear, rational understanding as to where it fits into the wider architecture of the organisational model. Businesses mistake AI activity for AI strategy. They launch isolated pilots, encourage experimentation without effective governance in place and chase trends driven by what they’ve heard their competitors are doing, or what’s currently occupying media headlines. And the outcome of this unstructured activity? Disconnected investments, potential security and compliance risks, and little or no value for money. This fragmented approach makes it difficult to scale AI initiatives that have shown promise. While these unstructured implementations aren’t necessarily failures, they can limit future options and seed the ground for unwanted organisational complexity further down the line, according to Deloitte’s State of AI in the Enterprise study.

An effective AI strategy requires a more coherent approach, one that’s fundamentally grounded in overriding business objectives, not technology selection. Creating an AI framework focused on identified business outcomes and high-value use cases is critical. For instance, deployments targeted at growing revenue, improving operational efficiency, enhancing the customer experience and reducing compliance risk can deliver tangible business benefits. But for them to do so, the organisation must have the correct structures in place.

Effective governance is key

Establishing robust governance is a vital first step. Defining who owns data and specifying responsible policies for use will help to improve ongoing security, and assist with compliance obligations in the long term. That’s why the US National Institute of Standards and Technology has described governance as ‘a continual and intrinsic requirement for effective AI risk management over a system’s lifespan and organisational hierarchies’.

Working to improve data quality and ensuring the right skills and infrastructure are in place are the foundations on which good governance is built. Accenture’s AI ROI Report, describes how resilient structural discipline facilitates effective knowledge management, which is essential for accurate forecasting, implementing process automation and improving customer support. Taking the time to define success metrics is also important. It’s advisable to distil tangible criteria to help teams focus on identifying where value such as productivity gains can be found, and where cost savings can be achieved, or time-to-market improvements gained.

Laying the foundations to support AI implementations

Of course, the value of AI is dependent upon how it interacts with existing business systems and workflows. AI implementations tend to create new data and messaging flows that existing processes aren’t equipped to interpret or leverage for operational improvements. This requires businesses to think beyond existing tools and operational models which can quickly erode any value gained. Significant work is required to ensure infrastructure and processes are modified to fully support AI implementations.

Without a concerted effort to orchestrate information architecture capable of underpinning AI-driven transformation, all the cost, time and effort that’s gone into the project will be wasted. Poorly connected solutions implemented without this essential background work will simply create new operational silos that hamper business efficiencies.

Successful AI adoption doesn’t really have much to do with technology selection, according to Harvard Business Review’s Stop Trying to Pick the Best Technology. It’s much more about engineering a model that facilitates reliable information flows between systems, processes and teams and – above anything else – adds value to what the business does. And that’s exactly why a robust and comprehensive strategy is essential for every company, regardless of business size or industry sector.