The Age of AI in Business – Do Our Expectations Meet the Reality?
These days, it is almost impossible to go online without getting some notification around the growth of AI. However, 2025 has been rumoured to be the ‘year of the AI agent’ with hopes in the business world for AI agents to transform the way that work is done. However, AI developers and tech experts have mixed feelings about whether we are jumping ahead of ourselves in our expectations of what AI will be able to do.
Hot Topic in AI News- Is It Stealing Jobs?
There have been plenty of reports and newsletters released throughout 2025, suggesting that there is a growing gap between the reality and expectations of Generative AI. There has been a growth in AI usage in areas such as consulting, tech, media and STEM areas. Entry-level jobs available for graduates have been on the decline; however, it has been said that this is not only an AI issue but a wider economic one, too.
Graduates have been noted to be particularly vulnerable in the growth and adoption of AI systems in employment. Many entry-level roles, such as basic data entry and form-filling, are considered drudge work that AI systems can do instead. Over the last 2.5 years, these types of positions have decreased by a third and June 2025 saw the toughest job market in the UK for graduates since 2018.
Employees & AI
In response to the upheaval in the job market and trends of AI, the Stanford Institute for Human-Centred AI and Digital Economy Lab conducted a study to discover if what workers wanted from AI aligned with current AI capabilities. The study discovered that there are areas ripe for research and development after surveying 1,500 workers from 104 occupations to discern where AI agents could benefit work and where they may cause harm.
The biggest concerns found in this survey were around the accuracy and reliability of AI systems, at 43% of respondents feeling this was a big issue. Job loss concerns were reported by 23,% and the lack of human oversight concerned 16%. However, there was an overarching concern regarding AI systems encroaching on communications with vendors and clients or creative tasks.
The areas in which they felt AI would be most impactful were:
- Automation freeing up time for higher value work, 69.4%.
- Reduction in task repetitiveness, 46.6%.
- Improving work quality, 46.6%.
The most welcomed automations that have been highlighted include scheduling appointments, rectifying record errors and maintaining information files.
In Local Government, employment is yet to be determined. As more councils adapt to the use of automation and artificial intelligence, much of the work being conducted this way is in administration.
Many AI experts highlight the importance of recognising AI as a tool, rather than worker replacement. Most current Local Authority AI usage is tool-focused; however, many report a need for training and expertise. The outlook for Local Authority employment going forward may be in training for AI applications and AI-specific roles. However, this is still a developing area with much potential to grow and to change.
There is also an influx of cybersecurity and protection in AI use to help safeguard data and systems.
The Impact Upon the Job Market
AI automation is going to redefine work whether we like it or not, so being aware of the way employment, business and skills need to align with the AI shift is critical. The Stanford Institute for Human-Centred AI and Digital Economy Lab looked at the value of various skills and how comparable they were with those that are least susceptible to AI replacement.
The analysis from this data would suggest that some jobs considered high-wage and high-skill, such as process monitoring and data analysis, may decrease in value. Alternatively, roles that require effective communication, teaching, and organising work may grow in importance as AI evolves.
AI Agents & Expectations
AI agents are the newest and ‘hottest’ AI tool in many industries now- they are software programs capable of autonomous actions to plan, understand and execute tasks. These agents are powered by LLMs (Large Language Models) and are capable of interfacing with other models, tools and aspects of a system to fulfil the goals of the user. This is beyond the typical scope of AI chatbot functions, or the automated customer experience emails and bots.
AI agents are also different from typical AI assistants, which require a prompt every time a response is required. The theory behind these agents is that a user provides a high-level task and the agent figures out how it should be completed. Yet, the offerings on this are still in early stages, and generalisations in technology about 2025 being the year of the AI agent are said to be mostly hype. AI researcher and Senior Research Scientist Marina Danilevsky believes that the new ‘agent’ technology is nothing new and is just orchestration dressed up with a new name. She also stated regarding 2025 being the year of the AI agent, “We haven’t even figured out ROI on LLM technology.
Danilevsky and other experts on AI technology share the same tone that there may be a disconnect between reality and our expectations. Considerations around the required effective communication from users to allow for the agent to do what is necessary and interpret the prompts correctly have been voiced. The hype around imagining the agent could think for the user and make decisions, then act, which is quite a harrowing concept, has been a notable flaw in expectations.
However, there is still a great deal of hope and promise around the development of AI agents. It has been highlighted that this development should be a tool, not a replacement for human workers.
The future has AI interwoven in it, and while the technological background of the economy is changing, perhaps AI training and integration of automation into education is the forward movement needed to keep pace with ever-changing technology. How effective AI agents will be for businesses is yet to be seen.
Find out more in our magazine: https://issuu.com/iesecic/docs/the_public_catalyst_issue_1
