Generative AI is not a magic wand, it’s a strategic tool

Billed as revolutionising everything from content creation to customer support, AI is too often viewed as all-powerful. In particular, the fast incorporation of AI agents—which can perform tasks and make decisions on their own—into business processes has only heightened this excitement. 

McKinsey puts the contribution of generative AI (GenAI) to the global economy at up to $4.4 trillion. Now the biggest challenge is to find strategic and well-aligned business objectives with AI.

Debunking common GenAI myths

Despite the extraordinary capabilities of GenAI, it’s often misunderstood. A recent survey by Deloitte tells us that 78% of the companies plan to increase their AI spend, and only 20% see significant impact. 

This difference suggests a prevalent misconception that AI can solve complex business problems independently with little to no human interventions. However, GenAI models require fine tuning, supervision and integration into existing workflows to yield impactful results.

Another traditional myth: GenAI would generate value overnight. Companies could begin to invest in AI-driven initiatives in a big way and expect quick return on investment. The reality, however, is that it requires immense capacity and effort to train models, cleanse data, and manage security and compliance issues to deploy effective AI in the marketplace. 

Next is a classic one: AI will replace human jobs. Microsoft’s 2024 work trend report suggests that business leaders want to hire more talent with AI skills. 

Steer clear of ‘shiny object’ syndrome

The key factor is data readiness. LLMs (large language models) thrive on rich, high-quality datasets. However, many organisations suffer from incomplete or siloed data. Sound data governance and infrastructure is a primary investment in creating AI-generated insights that are both possible and reliable. 

Besides this, cross-functional collaboration is instrumental in AI success. Cross-functional collaboration between IT, operations, and business units holistically drives AI adoption, ensuring alignment between technical capability and business needs.

Also Read

Generative AI isn't a replacement for traditional tech, it's an augmenter, says Google Cloud's Subram Natarajan

Security, monitoring, and talent

To roll out GenAI, there is a complicated array of operational issues that can get in the way of success. First, there is model accuracy and bias. AI models inadvertently pick up biases from training data, which produces skewed and unethical outputs. Businesses need to have solid monitoring procedures for identifying and eliminating bias to ensure fairness and abide by regulation.

Then there is security. As GenAI models handle high volumes of protective data, they become susceptible to cyber attacks. It is important to have strong security controls, including encryption of data, authorisation controls, and routine audits, to safeguard AI applications from being exploited. 

Further, adherence to NIST and ISO 27001/2 cybersecurity standards is necessary to ensure the security of AI deployments. 

Lastly, headhunting and upskilling are also paramount. AI engineers are in greater demand than supply; companies must fund training in order to equip their talent pool with AI initiatives management expertise.

Make GenAI a business asset

To enable businesses to achieve the highest value from GenAI, specific metrics need to be defined to measure the ROI in terms of enhanced efficiency, cost saving, and revenue enhancement contributed by AI projects. For instance, an Accenture study attests that businesses with AI-driven operations are reaping 2.5 times the peers’ average revenue growth.

Other qualitative variables that may generate financial figures are innovation potential, employees’ productivity, and customer satisfaction. Companies need to tackle AI deployment incrementally, beginning with pilot projects that yield rapid outcomes and insights prior to scaling up throughout the business. This mitigates risk and enables incremental improvement based on actual performance.

No magic, just strategy

GenAI is not some magic book where waving the wand delivers instant success. It’s more of an elaborate recipe—a dish that needs just the right blend of strategy, data, and execution to brew real business value. The potential of AI can be tapped only when it is thoughtfully merged into the very core of the organisation’s functions. 

So, drop the hype, take a level-headed approach, and engage in ongoing refinement so that AI is a lasting value asset and not another buzzword.

The author is Data Engineering Manager, Optum, a division of UnitedHealth Group.