The New Communicator: Human-Led, AI-Supported
Human Led - AI Supported. Credit: Rich Baker
Edition 1 in a three-part series on AI and the evolution of communications in organisations.
The first time I used AI to draft an all-company email, I felt two things: relief and unease. Relief that the blank page was filled in seconds, and unease wondering,
“Am I about to be replaced by a robot?”
In internal communications circles, this mixed feeling is everywhere. Depending on who you ask, generative AI is either going to make your job way easier or quietly edge you out.
The truth, as ever, sits somewhere in the middle.
“You’re not likely to lose your job to AI… but you very well may lose it to someone who knows how to use it.”
— Frank Wolf , Staffbase Co-founder
The new internal communicator — and really, insert any job title here — isn’t a robot. It’s the human who knows how to work alongside one.
Beyond the hype: AI as a comms partner
AI has officially entered the chat. ChatGPT alone now supports over 400 million users a week (OpenAI, 2025). Meanwhile, nearly 30% of comms leaders say generative AI is one of their top priorities this year (Gallagher UK , State of the Sector, 2025).
This isn’t theoretical. Communicators are already experimenting with AI — using it to brainstorm, edit, research, personalise, analyse, and automate. Not to replace the work we do, but to shift how we do it.
“Generative AI is your assistant, not your replacement. The human element is critical.”
— staffBase , 2024
Used well - and again this applies to most jobs - AI doesn’t steal our value; it gives us more time and space to deliver it.
Augmented, not automated: Where AI adds value
Here’s how internal comms teams are already putting AI to work:
Faster first drafts. Tools like Amazon Q, Microsoft Copilot, and ChatGPT can generate content from prompts or bullet points in seconds. Staffbase reports that some teams have cut drafting time by 50% using AI assistance.
Personalised communication at scale. Enterprise platforms like Poppulo and Cerkl are using AI to deliver smart segmentation and dynamic messaging based on role, location, or language preferences ( Tim Vaughan Poppulo , 2025). It’s not one-size-fits-all anymore — and the bar is rising.
Listening at speed. AI can flag broad sentiment patterns across survey data or live chat threads. But tools like Workvivo caution that nuance is still a struggle — sarcasm, local context, or subtle cues often trip AI up (Workvivo by Zoom , “AI in Internal Comms: What Not to Do”, 2024). So treat these insights as indicators, not absolutes.
Always-on answers Internal chatbots like Amazon’s excellent Ask A to Z are already reducing HR team load by answering FAQs 24/7 in natural language — a huge help in high-volume, multilingual environments.
Human + AI in action: A real-world example
Last year, I partnered with a colleague in our European People Experience and Technology (PXT) team at Amazon to explore how generative AI could support policy comms.
The challenge: translate key policy summaries into six languages — quickly, and in a way that felt ‘human’ but was also transparent in how it was created. We piloted the use of AI-generated video using real-word video as the template, to deliver multilingual, audience-tailored summaries across our frontline population.
One of the key outcomes — and one that mattered to me — was understanding whether people were ready for this. When we scoped the pilot, that was at the heart of the conversation: would the value for employees — clarity, speed, access — outweigh any hesitation about AI being part of the process?
The results were clear:
Over 90% of employees rated the videos helpful or very helpful
Understanding increased across all six language groups
HR helpdesk queries about the policy dropped noticeably
When asked directly how they felt about the use of AI to create the videos, almost no one expressed concern — the value of the message mattered more than how it was made.
We didn’t hand it all to the machine — human involvement and oversight was essential. The AI wasn’t perfect. But AI helped us scale faster, cut through complexity, and meet people where they are. And that’s the point: AI can accelerate delivery, but it’s still human judgement that ensures communication actually lands.
What AI can’t do (and shouldn’t)
Despite its shiny speed and fluency, AI still lacks the qualities that make communication matter.
It doesn’t know the political nuance of that reorg announcement. It doesn’t sense that this update will hit differently in Manchester than it will in Madrid. It won’t tell you that the phrasing in paragraph two will upset frontline teams — because they’ve heard it all before.
That’s our job.
AI might help with the how. But we still own the what, the why, and — most importantly — the feel.
As Sonam Jindal put it in The Verge:
“Human intelligence is the basis of artificial intelligence, and we need to be valuing these as real jobs in the AI economy that are going to be here for a while.”
And there’s something else: trust.
Internal channels are where employees go for clarity and truth — especially when external information is messy or unclear. If we start publishing unfiltered AI copy, or using models we don’t fully understand, that trust erodes. We risk sounding synthetic. Or worse, bland.
We don’t just need to be in the loop — we need to be at the heart of it. Not just for quality control, but to shape the message’s foundation: its cultural tone, lived experience, and empathy.
It’s what makes our work human. And effective. The more technology we bring into the mix, the more vital it is that we bring ourselves with it.
Where to start: Try, learn, iterate
If you haven’t started experimenting with AI yet, now’s the time. Start small. Stay curious. Stay human.
Here’s a few good entry points:
Use approved platforms. Stick to your organisation’s official tools. For example, Amazon offers a range of enterprise-ready GenAI services — including Amazon Q, Bedrock, and Titan — designed to support secure, scalable use across industries. Avoid uploading sensitive content to public platforms.
Start low-stakes. Try AI to rewrite a draft newsletter intro or summarise a meeting for your own use, before diving into high-stakes messages.
Review and refine. Always. No matter how good the first pass looks.
Build prompts with your tone in mind. AI will echo what it’s given. So make sure your voice is part of the process.
Across the industry, comms teams are being encouraged to play, test, and share what works. The skill now isn’t knowing all the answers — it’s knowing what to try, and how to learn from it.
Final word: This is the job now
Communication is evolving — and so are we. The new internal communicator doesn’t just write content, or create strategy. We design systems. Shape experience. Use data. And increasingly, we orchestrate people and technology in step with one another.
We’re not being replaced. We’re being repositioned — from creators to enablers. And that shift starts with getting comfortable letting AI take the first pass… while we always take the final one.
Coming soon: “Don’t Let the Robots Write Alone” — Edition Two of this series will explore how to use AI for writing without losing tone, trust, or taste. We’ll look at the lines you shouldn’t cross — and how to use AI without sounding like one.
