Short on time? Read the key takeaways :
- Three types of AI are helping organizations transform processes and better serve customers and employees: traditional AI, generative AI and large language model AI.
- One example of how integrating AI into IT support processes can equalize care for employees is the way it could handle device issues. This scenario illustrates how all three types of AI would combine to address the common situation of an employee needing help to resolve a technology issue.
- Several factors are making this scenario of AI-driven IT support a vision of the future and not a reality now. These include the need for data training and organizations’ security concerns about exposing data and IP publicly.
No one wants to be stuck in hold purgatory or chat limbo after something goes wrong with their device. Fast, reliable IT support is highly critical for employee experience and productivity, regardless of role, seniority, or location. AI tools present a transformative opportunity for IT support to act as the great equalizer of care.
This philosophy doesn’t mean everyone receives identical support but that they receive exactly the level of care they require. It brings together tools and technologies to solve problems before any end user notices or even before these issues occur. Ideally, this objective supports an equitable level of IT support for everyone, regardless of role, location or need. The concept of equalized care aligns with the goal of creating an equitable and inclusive IT environment where everyone's needs are addressed fairly, contributing to a more productive and positive employee experience.
Approximately 36% of AI-mature organizations have adopted AI, according to Unisys’ “From Barriers to Breakthroughs” report. And AI will help them speed up their decision-making. As organizations explore opportunities where AI’s capabilities can shine, IT support stands out. Because of this, smoothing the process of providing IT support can yield significant advantages.
Types of AI
Before considering AI’s power to equalize care, it’s important to distinguish among three popular flavors of AI that organizations are introducing to transform processes and better serve their customers.
- Traditional AI: Uses algorithms to solve business challenges by evaluating data to produce insights. This is often an organization’s first introduction to AI, given its integral connection to a critical business focus — data analytics. AI and data are closely linked with a bond that can always be strengthened.
- Generative AI: Has been generating (pun intended) major buzz lately. Generative AI creates new content based on the source data and prompts fed to it. For example, by entering the appropriate prompts, you could use generative AI to create code and a script to install a driver or write suggested copy for a social media post, among many possibilities. Take a deep dive into the capabilities of generative AI in Unisys’ eBook, New dimensions in AI: How generative technologies are shaping the digital workplace.
- Large language model (LLM) AI: Is a subset of generative AI that understands human language and provides human-like answers to an extensive range of queries, even those that may include colloquialisms. For example, an AI-driven chatbot like ChatGPT can recognize the nuance of how people talk (i.e., “Oh no, broke computer” vs. the more formal “My computer needs repair.”) Without the use of an LLM, a chatbot might not recognize that former comment and immediately send the employee to a human operator. For example, an AI-driven chatbot like ChatGPT or Microsoft Copilot can recognize the nuance of how people talk (i.e., “Oh no, broke computer” vs. the more formal “My computer needs repair.”) Without the use of an LLM, a chatbot might not recognize that former comment and immediately send the employee to a human operator."
In the future, strategically combining these three types of AI will deliver a superior IT support experience for employees.
AI in action: an equalized care scenario
By embedding AI and automation into an assistance mechanism, uniform support is accessible to everyone in the organization — from recent graduates fresh out of college to the senior executive in the corner office. With more data training, real-world scenarios like the one described next could feasibly be addressed with a combination of the three categories of AI outlined above.
Imagine there’s a problem with your laptop. Traditional AI, with its adept pattern detection, identifies your IT issue and detects 10 other laptops with the same issue. The IT team can then leverage this pattern detection at scale, discovering 4,800 laptops that haven’t experienced the problem yet. The AI can recognize if the devices display similar circumstances to the first laptop, signifying a problem will arise at some point. Then, automation will proactively apply a fix before that can occur, sidestepping the escalation of the issue. You, as the end user, remain none the wiser, while the IT team is freed up to focus on more pressing business initiatives. After all, the pinnacle of equalized care is negating the need for it in the first place.
Perhaps you did notice this issue and reached out for IT support. A chatbot powered by LLM AI engages you much like a human would. This involves walking you through a decision tree of pertinent questions to diagnose the issue and directing you to the right support contact or resource if further assistance is needed.
Let’s break down what this process looks like without AI. Knowledge workers (in this case, likely service desk agents) create new decision trees themselves and distribute the trees among their teams. The agents then reference the trees when interacting with end users to troubleshoot an issue—but only after an individual calls the service desk or submits a ticket.
In contrast, when leveraging AI to deliver equalized care, the knowledge workers can simply feed the AI model a wide range of source data, from which it will create those decision trees for the chatbot. The chatbot handles the interaction with the end user, only routing the ticket to a human agent if the source data provided is not sufficient to troubleshoot the issue.
Last, generative AI can write a technical fix for the issue and apply it to all digital experience tools. This offloads the burden—on both service desk agents and end users—of spending hours trying to resolve the problem, which can require as much as one to five hours a week for half the workforce. At most, you may be asked to click “Yes” when AI alerts you to the issue and asks if it’s OK to apply the fix. In doing so, the issue won’t inconvenience other users who might not even be aware of the problem.
Why hasn't AI-driven equalized care become a reality yet?
The scenario described illustrates the ultimate goal in AI-assisted care — and it’s within reach. The tools are already available, so organizations may mistakenly believe a scenario like this should be possible already, given the ongoing buzz surrounding platforms like ChatGPT.
However, there are constraints on how fully an organization can utilize open-source generative AI platforms. This is because they risk exposing their data and IP to the broader community to gain decision trees fed by their own data.
What remains to be accomplished is thorough data collection and training for the model, rigorous validation processes, and robust security measures to prevent publicly exposing any enterprise data used in this training process.
Beyond the fears surrounding data security, other fears related to AI can be alleviated with education. Debunking these myths — such as the notion that AI will replace knowledge workers — can free up your business for innovation. A related myth is that AI will completely replace the service desk, which is far from the case.
What will the future of AI and IT support look like?
AI will never fully replace people because even with the best training models, AI will never be completely predictive. Generative AI feeds off data from repeat problems, and there will continue to be unique challenges that “stump” AI. People will always need to be part of the process, checking and verifying the information produced by AI, as well as continually updating the source data informing its algorithms.
IT budgets will naturally expand as industry requirements and other variables grow increasingly intricate. Ironically, AI will be among the technologies amplifying that complexity, necessitating human expertise in AI for breakthroughs.
A more probable scenario is AI easing the work for human operators, alleviating their workload by handling routine queries and guiding users through established troubleshooting procedures. This enables human IT support to focus on unique questions and uncommon challenges. By working in concert, AI and human intelligence can optimize the efficiency and effectiveness of IT support. Expect this symbiotic relationship to become a common evolution of IT support within months rather than years.
Explore AI’s potential for your organization
Technology tools will mature over time, and LLMs will be finessed — probably in less than a year — so that ideal situation is coming into reach. One exciting outcome of so many organizations developing tools based on AI is that the growth in this technology is speeding up.
Discover how Unisys can empower your organization to unlock the limitless potential of AI to make swift, data-driven decisions, automate tasks, streamline processes and free up valuable employee time for strategic initiatives. With Unisys' AI expertise, your organization can surpass conventional tools to accelerate efficiency, boost productivity and transform operations.