Ready to Integrate AI Systems Into Your Workplace Performance? Here's What You Should Know.

Opinions expressed by Entrepreneur contributors are their own.

Being comfortable with change is a core competency in today’s work environment. Market disruption isn’t slowing down nor is the application of new skills and systems in our daily work lives. “This changes everything” is a stretch, even with AI, but a renewed focus on joy, satisfaction and personal expertise continues to be the greatest investment employers can make.

How might artificial intelligence systems integrate in our daily work lives? Maybe the most natural adaptation is into systems we already use to achieve consistent business outcomes and electronic performance support.

Electronic performance support systems

EPSS is hardly a new invention, nor one that needs to (or should) rely on generative artificial intelligence systems to work. In the early 1990s, industry expert Gloria Gery detailed in great length the potential cost savings organizations could receive from providing a smart electronic coach employees can use while working. EPSS can also not only coach employees on the next best steps, but preemptively provide access to data and troubleshooting steps before mistakes are made. Imagine how helpful the following might be in doing your own job:

Support systems, regardless of medium, take a great deal of design and care during implementation to be of value. The core premise of legacy EPSS is that the contained information, automation and coaching are not only truthful but also the organization’s preferred practice. EPSS implementations help ensure preferred outcomes and also model what all successful employees should do to be successful.

In today’s digitally enabled workplaces, organizations use a wide variety of systems to provide the function of a dedicated EPSS. Examples include knowledge bases, wikis, chat and tutorials, as well as automation systems bolted onto enterprise software. Today’s knowledge worker also relies on online searching to access vendor training, checklists, videos or social media for influencer tips and tricks. OpenAI’s ChatGPT, Microsoft’s new Bing, Google’s Bard and hosts of other new systems are joining the toolbox of today’s workforce.

An effective blend of external information sources and internal, proprietary best practices has no doubt been the source of many successful projects. So, how does generative artificial intelligence change this picture?

Artificial Intelligence as performance support

Training your generative AI is extraordinarily important when using it for performance support. It’s very likely your organization’s prior implementations of performance support were highly curated and crafted and contain proprietary data, policies and procedures. Only carefully developed systems generate consistent organizational outcomes.

Generative AI, trained only on proprietary corporate data sources, could likely compose support materials and even identify the development of support materials experts might overlook. But when using open-source training data, the results may be closer to really good-looking, well-written BS.

Many experts have argued EPSS alone is a substitute for formal training. This is an important debate when applying generative AI systems. Following well-formed instructions is a lot easier than judging newly presented instructions for their applicability in a situation. Experts, with experience performing a task, are suited to evaluate new information, while first-time performers may not be.

Here are a few questions to consider about your own AI implementations:

What is the source of the AI’s training data? Open-source or proprietary?
Is my organization okay with using proprietary data in and on an open-source AI system? Will the open-source system learn my organization’s secrets?
Does the system provide trusted, verifiable results, or merely a creatively generated output?
Does the variance of generated results have the potential to impact the consistency of organizational outcomes or behaviors?

Human experts and expert systems working together

What do the employees in your organization want to be experts in? What tasks do your employees want automated?

Between human expertise and automation is a potentially crowded intersection of generative AI, talent and emotion. Early experiences with generative AI systems have exposed the bizarre potential for an AI system to interact with highly emotional language and concerning suggestions. Even with an AI system tailored not to respond with emotional language, human self-image will be impacted if an organization becomes reliant on systems for idea generation or decision-making.

Joy, happiness and self-worth of employees should not be sacrificed just because a system can help make decisions at scale. Business leaders have several introspective questions to ask about how they want to work and manage teams:

Does job satisfaction erode for employees using AI systems?
When does client satisfaction erode for customers and employees interacting with and through AI systems?
What competitive advantages become commoditized by AI systems as other industry players start using AI systems?
Who do I want to make specific business decisions? Systems or people?

Crucial questions are surfacing. Call centers, for example, have started to deploy AI support systems that take customer calls, collect early information, triage urgency and severity and whisper in the ear of the live agent (if you ever get that far). The systems have a reportedly negative impact on the emotions of agents, overwhelming their sense of agency and ability to interact with other human beings.

Human beings need joy in their lives. The systems we can implement today should start with joy in mind — the joy of success, accomplishment and feeling needed and desired — and should not exclude our interactions with digital systems. Today’s business leaders have the opportunity to design work experiences that leverage the best of technology to amplify the human experience. What questions will you ask of today’s technology breakthroughs?

Source link