AI is making it easier than ever to produce content, analyse information and get communications out the door... but just because we can automate something doesn't mean we should. When it comes to communication especially, the consequences can extend well beyond saving a few hours.
For communicators, the challenge isn't simply figuring out how to use AI - it’s actually understanding when to use it, when to keep it out of the process and what our audiences expect from us. So, what did our research tell us?
Efficiency is only part of the equation
AI has quickly become part of the communications toolkit. Research, first drafts, content development and strategy planning are all common use cases, with studies showing widespread adoption across internal communications and PR.
However, our comfort with using AI doesn't necessarily reflect how comfortable our audiences are receiving AI-generated communications.
Our research involved 124 respondents across 15 countries, which found that people were generally more comfortable with AI supporting lower-risk tasks, such as developing content and plans. Comfort levels dropped when AI was used for more personal or high-risk processes, including recruitment, payroll, performance management and complaints.
There's a fairly obvious distinction here; using AI to help structure a communications plan is very different from using a bot to conduct a job interview or respond to an employee raising a sensitive concern. One helps someone do their job, while the other could fundamentally change how a person experiences an organisation.
Though that's entirely the point because it shows that efficiency for the organisation doesn't automatically translate to a better experience for its audiences.
Trust is part of the deliverable
Communications is centred around helping people understand what's happening, why it matters and what they need to do next. All of which requires credibility, context and, sometimes, a little humanity. AI can produce polished content, but polished doesn't always mean effective and when you have generic language, repetitive phrasing and content without a distinctive voice - audiences can become disengaged. Worse, if people feel they're being fed automated messages when they expect genuine human communication, it can undermine trust.
This matters particularly when communicating change, responding to employee concerns or delivering messages from leaders. AI can help shape the words, but it can't replace a leader's accountability or a communicator's judgement about how a message will land.
Human oversight needs to mean more than giving an AI-generated draft a quick scan before publishing. Someone needs to check the facts, challenge assumptions, consider the audience and make sure the content reflects the organisation's voice and values. If AI is involved, be clear about where human judgement sits and if a communication requires empathy, discretion or a meaningful personal connection, consider whether automation is appropriate at all.
Don't assume everyone feels the same way about AI
One of the more interesting findings from the research was that overall comfort with AI increased with age among respondents. That challenges the familiar assumption that younger employees will automatically be more comfortable with new technology. Of course, the findings reflect this particular survey rather than a universal rule, but they reinforce an important point: familiarity with technology doesn't necessarily translate to trust in it.
People's attitudes towards AI can be shaped by concerns about privacy, fairness, job security, accuracy and how the technology is being used. Someone might happily use AI to organise their workload but object to an automated system assessing their job application.
For communicators supporting AI adoption, assumptions about age, role or digital confidence can get in the way of an effective strategy. Find out what your audiences know, what concerns them and what they need before deciding how to communicate the change.
There's also an opportunity to explore how AI could make work more accessible. Respondents living with disability or neurodivergence identified benefits including reducing overwhelm, supporting organisation and helping with communication. These benefits shouldn't be overlooked, but neither should organisations assume that one solution will work for everyone. Involve the people affected in decisions about how tools are introduced and used; it's a much better starting point than making assumptions on their behalf.
Give AI a job description
Rather than treating AI as either the answer to everything or a technology to avoid altogether, organisations need to be clear about where it belongs. Use it to support research, structure ideas, develop first drafts and reduce repetitive work. Put appropriate checks around accuracy, privacy and the use of sensitive information. Be particularly cautious when AI is involved in decisions or communications that could materially affect an individual.
Plus, don't confuse human involvement with human accountability. If a person is responsible for the final output, they need the time, capability and authority to question what AI produces rather than simply approve it. The same applies to content - copying and pasting an AI-generated post might save time, but editing it to reflect your organisation's voice, audience and purpose is what makes it useful communication. The goal is to use it where it adds value without compromising the quality of the outcome.
The question communicators need to ask
AI will continue to change how communications teams work. The opportunity is to spend less time on repetitive tasks and more time on the thinking, relationships and judgement that make communication effective, but that only works if we keep the audience in the equation.
Before introducing AI into a communications process, ask what it will improve, what risks it introduces and how the people on the receiving end are likely to experience it. Consider whether the task is routine or personal, whether human judgement is essential and whether the approach aligns with audience expectations.
Just because AI can do something doesn't mean it's the right tool for the job. The best communications decisions won't be the ones that maximise automation, but the ones that use technology thoughtfully, protect trust and make life better for the people we're communicating with.
Read the rest of the AI report here.
