Avoiding The Gen AI Hype Machine


By embedding innovation into the planning process, leaders can navigate the challenges of introducing new technologies and drive meaningful business outcomes.

Nikhil Sarathi, Eric Lu, Mofeed Sawan, and Shreya Gandhi-Gupta

5 min read

Generative AI has caused seismic shifts in nearly every aspect of our daily lives. The natural language processing tool makes it easier for people to develop code, write research papers, or come up with new dinner recipes. How to use — and control — generative AI is a hot topic of conversation from boardrooms and classrooms to the halls of power in Washington, D.C., and everywhere in between.

Healthcare is not immune from the hype. Generative AI is positioned as a game-changer for every industry segment. Life sciences companies hope it will accelerate drug discovery. Technology vendors and health systems see generative AI as a tool to enhance electronic health record systems and create more predictive care. For insurers, it may improve risk management and member engagement. The adoption of generative AI in healthcare is expected to soar over the coming years, with market value projected to hit $22 billion by 2032, up from $1 billion in 2022.

While generative AI and other innovative solutions open the door to new possibilities in healthcare, they risk being a siren song if technology leaders aren’t careful. Leaders must operate at two speeds simultaneously — maintaining existing technology all the while assessing, testing, and learning how and when to adopt a groundbreaking innovation like generative AI.

Adopting a business leader mindset

Technology leaders often face internal and external pressure to pursue an attractive, disruptive new technology. That’s certainly the case with generative AI where there’s been an arms race across industry sectors to experiment with various use cases for the technology.

Instead of falling into this trap, technology leaders must adopt a business leader mindset. Working with other parts of the organization, they need to establish a set of outcomes and key performance indicators that they want to hit. At that point, they can decide which — if any — new technology is right for the job. This approach allows leaders to keep their eyes on business objectives, effectively manage a portfolio of existing and new solutions, and drive accountability for themselves and their teams. To gain buy-in from the rest of the organization, technology leaders should be explicit about this process and make it clear that everything links back to organizational goals.

Setting an organization’s level of risk tolerance

Leaders also face a risk-reward tradeoff when deciding if a new and untested technology solution is worth the sizeable investment. Having a firm understanding of and process for evaluating risky initiatives is critical. But let’s be clear, managing risk does not equate to avoiding risk. There will always be opportunities and reasons to go into the relative unknown. The point is that leaders must weigh the opportunity cost of dedicating resources towards a new technology like generative AI against doing nothing or focusing more on existing solutions.

This is where taking a portfolio view comes into play. Leaders should map out all the opportunities that lie ahead and be ruthless in prioritizing which should be acted on in the near- and long-term. A prioritization process also allows leaders to choose what not to do, which is especially necessary during times of budget constraints.

Embedding innovation in process planning

As leaders map out a planning process for innovative technologies, they should embed these three foundational capabilities to drive success.

  • Embrace product-based development: A product-operating model allows organizations to set up business processes that address gaps in the usual project-centric execution. Under this approach, organizations can move away from funding individual ideas in favor of a holistic, enterprise-wide product vision that enables more solutions to cut across multiple use cases. This leads to more stable funding. Support for projects like generative AI cannot be flipped on and off; they require a steady stream of funding for a consistent period. That’s why starting with the use case is so important.

    Stable resourcing of design and execution teams is crucial here, particularly when customer needs are unclear, which is occurring with generative AI as overarching business goals are uncertain. A cohesive team that works on a unified problem can foster innovation and co-creation, which is not possible with standard project-centric execution. Integral to this team approach is setting clear expectations and holding everyone accountable for outcomes.

    Business and IT leaders need to be in lockstep on what the operating model looks like — who are the business stakeholders, who make up the product organization, and what day-to-day processes look like. This takes commitment, dedicated resources, and permission of joint teams to evolve and overcome obstacles that are bound to arise.
  • Evolve data governance processes: Healthcare organizations should already have a data governance process in place, although some are more evolved than others. Deploying and scaling a tool like generative AI necessitates additional considerations, especially as lines continue to blur on who owns the data.

    Generative AI’s ability to quickly structure unstructured data is one of its superpowers. But it can be unwieldy and organizations will see a muted impact if they start by looking at the data. Instead, begin with use cases and have a strong partnership between business and IT to ensure the data used is clear and well-governed. Beyond that, it is critical that the data being fed into the model is accurate and reliable to avoid potentially misleading or incorrect outputs. Security will also be paramount as organizations scale their use of generative AI and start pulling data from other sources. Going hand-and-glove with security is accountability and the ability to trace data back to its original creators. Technology teams must ensure that they can source data to ensure its accuracy and legitimacy.
  • Modernize the data environment: Organizations need to be far along in their broader data infrastructure and capabilities modernization efforts before attempting to deploy and scale generative AI applications. Benefits like computational scalability, cost-effective deployment of resources, and a data environment that’s accessible across the enterprise are all critical to unleashing the full benefits of generative AI.

    Beyond those data infrastructure musts, organizations need to establish strong data tracking processes, especially the large amount of unstructured data that is already in use. That means identifying where it comes from, how it has evolved, and where it is being used.

    However, one of the benefits of generative AI is its ability to create synthetic data outputs that complement existing data environments. Generative AI enables business users to quickly test ideas without waiting for proper data capture and cataloging typically done in a production run.

Don’t get distracted

This is a watershed moment for healthcare. Generative AI has the potential to push drug development, care delivery, and even back-office functions to new levels. Technology leaders will play a pivotal role in ensuring that their organizations do not get swept up in the hype machine. They must work on consort with business and clinical leaders to strike a balance between innovation and effective business operations. By embedding innovation into their planning process, leaders can successfully navigate the challenges of introducing new technologies and drive meaningful business outcomes.

In this video, we delve deeper into some of the unsexy parts of digital transformation that are essential for successfully deploying generative AI.

  • Nikhil Sarathi,
  • Eric Lu,
  • Mofeed Sawan, and
  • Shreya Gandhi-Gupta