AI in the workplace: Why training matters more than the technology

Artificial intelligence is embedded within many workplaces, often in ways that employees scarcely recognise. Yet there is an apparent contradiction at the heart of workplace AI adoption. A recent survey found that 67.5% of workers had received no AI-related training or guidance from their employer, despite around one-third already using AI tools as part of their job. This suggests that AI is spreading faster than organisational governance, creating both opportunities and risks. The challenge for organisations is no longer whether to adopt AI. Instead, it is how to ensure that AI is used effectively, responsibly, and in ways that genuinely improve productivity.

The discussion surrounding AI often oscillates between two extremes. On one side are advocates who promote AI as a revolutionary tool capable of transforming every aspect of working life. On the other are critics who warn of job losses, misinformation, and declining standards. The reality is probably somewhere in the middle.

According to John Pepper, CEO and Founder of Managed247, organisations should focus less on simply introducing AI and more on helping employees understand how to use it correctly. The greatest benefits arise when AI supplements human expertise rather than attempting to replace it. In many cases, the most valuable applications of AI are not dramatic technological breakthroughs. Instead, they involve removing routine administrative burdens that consume large portions of the working day.

One of the most common workplace frustrations is starting a document from scratch. Whether drafting a report, preparing an email, or producing a project update, many professionals spend considerable time simply constructing an initial framework. Generative AI tools such as OpenAI’s ChatGPT (ChatGPT) and Anthropic’s Claude (Claude) can help users overcome this hurdle. Rather than writing an entire report, AI can suggest structures, improve grammar, refine tone, and identify areas requiring clarification. The human author remains responsible for the content and conclusions, but AI reduces the administrative effort required to produce a polished document.

Effective AI use enhances human productivity; ineffective AI use replaces thinking with automation.

Using AI to improve meetings

Meetings are another area where AI can demonstrate tangible value. Many employees split their attention between participating in discussions and attempting to capture comprehensive notes. As a result, both activities can suffer. AI-driven transcription systems such as Otter.ai (Otter.ai) allow conversations to be recorded and transcribed automatically. Key actions, decisions, and discussion points can be extracted and organised into concise summaries.

The benefit is where participants can focus more closely on the discussion itself, leading to improved engagement and decision-making. For project management, regulatory inspections, quality investigations, and technical reviews, having an accurate meeting record can also improve accountability.

Modern organisations generate vast quantities of information. Technical reports, scientific papers, regulatory guidance documents, annual reports, market analyses, and internal procedures can run to hundreds of pages. Reading and extracting relevant information from these sources can consume significant employee time. AI tools increasingly offer support by summarising content, identifying themes, and extracting key findings.

For example, a regulatory affairs specialist may need to review an FDA guidance document spanning dozens of pages. An AI system can rapidly identify sections relevant to a particular topic, helping the user prioritise attention and focus on the areas requiring detailed analysis.

The key point, however, is that AI should support rather than replace critical review. Summaries are useful starting points, but important decisions should still rely on examination of the primary source document.

Automating routine administration

Many daily business activities are repetitive rather than intellectually challenging. Examples include sorting emails, cleaning datasets, and updating spreadsheets. These tasks often consume substantial amounts of time despite contributing little direct value. AI systems can automate many of these routine functions, freeing employees to concentrate on higher-value work requiring creativity, judgement, problem-solving, and interpersonal skills.

For knowledge-intensive sectors such as pharmaceuticals, engineering, healthcare, and finance, this shift may be especially significant because employee expertise can be focused on tasks where human insight remains irreplaceable.

Despite its advantages, AI introduces important risks that organisations cannot ignore. One concern is over-reliance. If employees become dependent on AI-generated outputs without appropriate oversight, the quality of decision-making can deteriorate. Language models may generate plausible but inaccurate information, a phenomenon commonly referred to as “hallucination”.

There are also concerns relating to confidentiality. Uploading proprietary information, financial data, intellectual property, research results, or commercially sensitive documents into public AI systems may expose organisations to significant security and compliance risks. Regulated industries face additional challenges. Pharmaceutical manufacturers, for example, must consider data integrity requirements, record retention obligations, validation requirements, and regulatory expectations before incorporating AI into quality systems.

Consequently, organisations require clear policies governing what information may be entered into AI systems and how generated outputs should be reviewed.

The importance of training

The survey findings indicating limited workplace AI training should concern employers. In particular, introducing AI without guidance is comparable to introducing any other enterprise technology without providing users with instructions, procedures, or governance. Effective AI training should cover areas like appropriate use cases, data protection requirements, verification and fact-checking, and prompt construction. Employees need to understand not only what AI can do, but also where its limitations lie. Without training, organisations risk inconsistent practices, poor-quality outputs, and potential compliance failures.

Research suggests that the most common workplace use of AI involves summarising information. Many workers use AI to condense reports, project updates, performance reviews, and lengthy documents into shorter summaries. Another popular application is information gathering. AI tools can assist with research by locating relevant information and presenting it in accessible formats. Editing and reviewing written content is also common. Employees frequently use AI to improve grammar, adjust tone, and enhance readability.

Interestingly, these applications align closely with the concept of AI as an assistant rather than a replacement. Workers appear to be using AI primarily to support existing activities rather than completely delegate responsibility.