EP11: Executive Assistants and AI: How Carve is Redefining the EA Role

From Leadership Development to AI Training

Fiona founded Carve in late 2022, originally intending to build leadership development programmes for executive assistants, drawing on her own background as an EA and 13 years in corporate L&D. As she explored the space, it became clear that AI adoption, not traditional leadership training alone, was the real key to unlocking career progression for EAs — freeing them from repetitive administrative work like calendar and inbox management, and creating capacity for more strategic, high-value contributions.

Confidence and Capability: The Two Pillars of AI Upskilling

Fiona is clear that effective AI upskilling requires both confidence and technical capability. Beyond simply learning tools, EAs need to build genuine comfort with experimentation — adopting a mindset of curiosity rather than fear of "breaking something." On the technical side, she highlights prompt engineering as the essential skill for working effectively with large language models, alongside a realistic understanding of what today's AI tools can and can't yet do well, cutting through hype and unmet promises.

How the AI Landscape Has Shifted Since 2022

Fiona notes a clear evolution in how organisations approach AI since Carve launched. In the early days, there was significantly more freedom for EAs to experiment with no-code tools and workflows independently. Over time, organisations have increasingly settled into specific ecosystems — such as enterprise ChatGPT or Microsoft Copilot — which, while understandable from a governance perspective, has narrowed the scope for informal experimentation.

At the same time, appetite has shifted from generic "intro to AI" sessions — often delivered by well-meaning IT teams — toward highly specific, role-relevant use cases. Fiona argues that only a small fraction of generic AI training is genuinely useful to any individual audience, and that real value comes from clearly demonstrating specific, practical efficiencies relevant to a person's actual role.

Real-World Impact: From Meeting Minutes to Custom-Built Tools

Fiona shares two standout examples of AI transforming EA productivity:

• AI-powered meeting minutes — Using AI transcription and chatbot tools to convert recorded meetings into polished minutes in minutes rather than hours, cutting a task that once took three to four hours down to under twenty minutes, while also helping senior EAs stay across multiple meetings without attending each one in person.
• A custom-built compliance tool — One EA working in a family office, tasked with manually cross-checking new investment opportunities against hundreds of existing holdings for conflicts of interest, used ChatGPT's guidance to learn basic Python and build a bespoke tool — despite having no prior coding experience — ultimately saving an estimated several weeks of work per year.

Both examples illustrate a recurring theme: curiosity and willingness to experiment, supported by the right guidance, can unlock significant productivity gains even for non-technical users.

Common Mistakes Organisations Make with AI Training

Fiona identifies several recurring pitfalls in how businesses approach AI adoption:

• Generic, one-size-fits-all training that fails to address role-specific use cases, leaving most content irrelevant to any single audience.
• Overly restrictive tool access, which can unintentionally push top-performing employees toward "shadow IT" — or even out of the organisation entirely — in search of better tools elsewhere.
• A lack of time investment from the busiest employees, who are often the ones who could benefit most from AI-driven efficiency gains, but who are least able to prioritise the learning required to achieve them.

Why Culture Is the Real Differentiator

Fiona frames organisational AI learning culture as a combination of shared values and shared behaviours — including the often-overlooked need to actively "unlearn" old ways of working, not just learn new tools. She highlights the value of combining top-down leadership messaging with grassroots, peer-led initiatives such as internal AI steering groups, where employees can openly share what's working, what isn't, and practical prompts or use cases.

A significant barrier to open AI adoption is the persistent perception that using AI is somehow "cheating", leading employees to hide their use of AI tools out of fear it might undermine perceptions of their value or job security. Normalising open discussion of both AI successes and failures, she argues, is essential to building genuine organisational confidence.

EAs as Agents of Change

Perhaps the most compelling theme in the conversation is Fiona's observation that executive assistants are uniquely positioned to act as organisational change agents for AI adoption. Working across every level of a business, many of Fiona's EA clients have taken on informal leadership roles — launching AI steering groups, championing best practice, and even evolving into broader "chief of staff"-style positions as their organisations recognise the strategic value they bring beyond traditional EA support.