Although it is still only February, agentic AI is already making a strong case to be considered the Oxford University Press’ word of the year, with Google searches for the term surpassing the previous week throughout this year. But what exactly is agentic AI? In simple terms, agentic AI refers to systems that do more […]
Although it is still only February, agentic AI is already making a strong case to be considered the Oxford University Press’ word of the year, with Google searches for the term surpassing the previous week throughout this year.
But what exactly is agentic AI? In simple terms, agentic AI refers to systems that do more than just respond to user input; they possess the ability to make autonomous decisions, take action on their own accord, and actively pursue defined goals without direct human intervention.
A traditional AI cannot act beyond the data it’s given, although it can often make recommendations when asked.
Take, for example, a wearable health device. Instead of simply recording health data and waiting for the user to seek out recommendations, an agentic AI could autonomously analyse this data and make decisions based on detected patterns. It could recommend lifestyle changes, suggest specific exercises, or even schedule a health check-up for the user without being explicitly instructed to do so.
Putting an agentic AI in a smart home environment could see it autonomously adjust lighting and heating based on occupants’ habits and preferences.
In a workplace environment, an agentic AI could analyse an employee’s emails, calendar events and project deadlines and automatically reschedule meetings or send reminders based on urgency.
That sounds useful. But despite the growth in AI adoption in recent years, scepticism over hallucinations and other failings remain a barrier for many potential users.
According to research conducted by Pegasystems, around one-third of workers (33%) express doubts about the reliability and quality of AI agents.
“It’s a relationship, like anything else,” says Nora Jones, senior director of product management at digital operations management firm PagerDuty. “We build relationships with humans, and now we’re building relationships with our technology.”
“A big way we’re building trust is by focusing on areas that are tedious or annoying, areas where developers would prefer to spend their time elsewhere,” she adds.
By automating repetitive tasks and eliminating bottlenecks, agentic AI aims to free up time and reduce operational risks.
PagerDuty’s agentic AI is helping organisations achieve this by automating mundane tasks and offering solutions that streamline complex processes.
One example is their ‘Agentic Site Reliability Engineer’, which autonomously identifies operational issues and guides responders to a resolution.
They’ve also recently launched an ‘Agentic Operations Analyst’, which analyses patterns within an organisation’s ecosystem to improve efficiency and inform decision-making.
PagerDuty believes that the true value of agentic AI lies in its ability to uncover hidden issues within an organisation’s systems – whether they are in the software, workflows, or operational processes.
Rethinking the workplace with agentic AI
For new organisations, AI presents a unique opportunity to rethink traditional development approaches: “Starting a new organisation today gives us a chance to think differently. Maybe AI can help us learn from what others have done wrong in the past,” explains Jones.
By using AI, organisations can potentially avoid costly mistakes that have been made by others in the past, leading to more efficient and agile business practices from the outset.
David Williams, senior vice president of product at PagerDuty, adds: “Software developers are the ones adopting this faster than anyone else.” He believes that developers are embracing AI to tackle lower-level coding tasks and automate the more tedious aspects of their work.
“Today those lower-level coding tasks are being done by AI already,” he says. “If they believe in AI to write code, they will believe in AI to run their operations.”
This mindset, according to Williams, demonstrates that if they believe in AI to write code, they will warm to it taking over operational tasks.
“They would rather use AI to write more code, and so they will be the ones the least sceptical,” he notes.
As PagerDuty sees it, the widespread adoption of AI in the workplace will happen from bottom up. They anticipate that once CIOs and CTOs begin considering AI tools for their organisations, teams will respond with, “We’ve already been using them.”
The adoption of AI is not limited to tech start-ups; PagerDuty believes that legacy industries, such as banks, fintechs, and certain cloud-native companies, are leading the way in AI implementation.
They also believe legacy banks, fintechs, and some cloud-native companies are leading in AI adoption.
Overall, the firm says it’s about building the trust, first, seeing the output and gradually firms will begin to warm to its new digital colleague.
