timothywheelerct

[timothywheelerct]

Generative AI Transforms Organizational Capabilities

One topic discussed in a symposium at the Gartner IT Symposium/Xpo 2023 in Orlando centered on generative AI (GenAI) and emerging trends in human-machine relationships. Able to learn from existing artifacts (data-driven output), GenAI generates new artifacts at scale. Such artifacts are novel and contain characteristics of training data without repetition. Beyond video, music, speech, and images, GenAI is being utilized in a diversity of business scenarios to create everything from product design to programming code.

According to a Gartner poll in early 2024, corporate interest in GenAI is strong, with two-thirds of organizations deploying GenAI across more than one business unit and 40 percent using GenAI in three or more business units. This was a nearly 20 percent increase in less than half a year, with 20 percent of organizations having GenAI solutions in production. Leading use cases include customer service (16 percent), marketing (14 percent), and sales (12 percent).

The process of transitioning to generative AI carries with it both significant risks and rewards. One major reward is the greater efficiencies and potentialities embodied in machines' shift from being tools to being integral team members. How this gets achieved varies. Many off-the-shelf AI platforms have public data as a basis, with the organization adapting the model to its own proprietary data and needs. Other GenAI systems involve creating a custom algorithm within the company that is trained on proprietary data and delivering highly specific output. While this approach costs more, it can be effective and necessary in delivering output that meets fine-tuned, highly technical needs.

As with other areas of IT, risks surrounding data privacy, hacking, and regulatory compliance are present with GenAI. In addition, outputs may be unreliable, opaque, and difficult to comprehend. For this reason, well-thought-out GenAI implementation involves exhaustive use-scenario plans, with areas such as degree of transparency and degree of automation defined and monitored. The key is effectively pairing AI solutions with human-driven functions and operations.

Practical examples of GenAI deployment include its use within one of the largest healthcare systems in Rhode Island for surgical consent conversations. Prior to any surgery, US patients must read and sign consent forms, which can be complex and challenging to understand. They are typically written at a college level, though many patients who sign them do not possess such a level of literacy or take the time to read them fully. Patients may feel overwhelmed and simply sign consent documents at an emotionally charged moment without considering potential outcomes.

The use of ChatGPT enabled the health care system administrators to take hefty three-page documents and turn them into one-page output that a middle school reader can readily understand. The health system is now applying this process in other administrative areas that interface with non-technical readers, such as patient intake forms.

This approach can be used in non-healthcare use scenarios, such as when municipalities help applicants navigate complex regulations related to new construction or the use of a public plaza for an event. Chatbots can also be integrated into 911 dispatch centers as staffing shortages persist and nonemergency calls and duplicate notifications are efficiently transferred to AI-powered customer service operations.