What this podcast really covers

The premise of Unprompted is deceptively simple: find IC professionals who are genuinely using AI, sit them down, and ask them what actually happened. No sanitised case studies, no sponsored talking points. The result is a series that operates as a real-time field report from a profession in transition.

The three episodes released in Season 1 already define a clear editorial arc. Rebecca Lindsay of hyperexponential tackles AI agents — software that can act semi-autonomously on behalf of IC teams, handling scheduling, content distribution, or audience segmentation without constant human intervention. Amanda Atkinson confronts the harder question of where human judgement must remain sovereign, particularly in culturally sensitive or leadership-critical communications. Erik Sebok takes the conversation furthest, treating AI not as a productivity tool but as a strategic interlocutor — a partner in thinking through communications problems, not just executing them.

What unites all three conversations is a refusal to abstract. Guests name the tools they use, describe the specific workflows they tested, and report outcomes with professional precision. This makes the content directly applicable in a way that most AI commentary in the business press is not.

Who this podcast is essential for

Internal communications managers and directors who are under pressure from leadership to demonstrate AI adoption but lack peer reference points will find in this series a practical benchmark. Each guest models what responsible, incremental AI integration looks like — including the moments it fails.

HR and People leaders responsible for employee experience and change communications will benefit from the show's recurring theme of human-AI balance. The tension between efficiency gains and the irreplaceable human quality of certain messages is examined from multiple angles, providing frameworks for decision-making rather than simple rules.

Communications technology professionals and AI product teams building tools for the IC market will find in this podcast an unfiltered source of practitioner needs, resistance points, and unexpected use cases. The guests reveal where existing tools fall short in ways that product surveys rarely capture.

What the episodes really reveal

Across the Season 1 episodes, a consistent pattern emerges: the most productive AI use cases in internal communications are not the obvious ones. Drafting assistance — the first thing every IC professional tries — quickly becomes table stakes. The more significant gains appear in earlier stages of the communications workflow: problem framing, audience analysis, message architecture, and scenario planning.

The AI agent episode with Rebecca Lindsay reveals a second pattern: automation in IC is less about volume processing and more about cognitive offloading. The value of an AI agent is not that it writes 50 emails instead of 5 — it is that it handles the low-cognitive-load tasks that fragment an IC professional's attention, freeing them for the high-stakes work that genuinely requires human presence.

Amanda Atkinson's episode surfaces a pattern that is rarely discussed openly: AI introduces a new class of error in IC. Not factual errors, which are well-documented, but tonal and cultural errors — outputs that are technically accurate but emotionally wrong for a specific organisational context. Managing this class of error requires IC professionals to develop a new editorial competency: AI output review as a distinct professional skill, not simply proofreading.

What this changes in practice

The cumulative argument across the series is that AI does not make internal communications easier — it makes it more demanding of the human practitioner. The editorial judgement required to evaluate, correct, and deploy AI outputs is more sophisticated than the judgement required to write a first draft from scratch. IC professionals who approach AI as a shortcut will produce worse communications. Those who approach it as a collaborator requiring active management will produce better ones.

This reframing has structural implications for IC teams. It argues against reducing headcount as AI adoption increases, and argues for investing in AI literacy as a core professional competency. It also repositions the IC function: if the strategic value of the role increases as AI handles execution, the profession's claim to a seat at the leadership table becomes stronger, not weaker.

For practitioners currently hesitant about AI, the series offers something more useful than reassurance: it offers a realistic picture of the learning curve, including the embarrassing experiments and the false starts, from peers who are navigating it without a roadmap.