Behind the Scenes: Crafting ISSEAD’s Custom GPTs
Why We Did It: Curiosity Meets Strategy
At ISSEAD, we don’t chase trends, we interrogate them. Our venture into custom GPTs wasn’t born of hype, but of a clear strategic question: how might AI elevate our precision, scale, and alignment without diluting our ethos?
While generic AI tools are adept at speeding up tasks, they often flatten tone, overlook nuance, and fail to grasp regulatory or sectoral depth. That’s not good enough, for us or for the organisations we serve. We wanted something better: digital collaborators shaped by our values, trained on our content, and fluent in our core focus areas: EU funding, sustainability, and strategic innovation.
The Challenge: Scale Without Sacrifice
ISSEAD operates lean but ambitious. We hold ourselves to high standards in three areas:
Delivering intelligent, high-impact content.
Ensuring legal and policy alignment across outputs.
Staying true to a voice that is strategic, optimistic, and fear-free.
Generic AI tools greatly help with speed, but they often miss tone, nuance, relevance or even factual correctness. That’s risky in a field like ours. We needed something better.
The Journey: It's a Learning Process!
We asked: What would an ISSEAD-native AI assistant look like? The assistant needed to understand our three core areas, EU funding & innovation, project design, and sustainability, while working with real sources and policy references rather than relying on generic internet summaries. Just as importantly, it had to communicate in a confident, clear tone, steering away from fear-based messaging and aligning instead with our optimistic, evidence-driven approach.
But beyond functionality, we wanted our GPTs to feel like part of the team, tools that could grasp not only the content we produce but also the context and intent behind it. We created three distinct GPTs to support our core areas of work: Calliope, our content and design assistant; Hanna, a seasoned business strategist; and Themis, our company compliance advisor. Each of them brings a different aspect of ISSEAD's operations to life, acting as a digital extension of our core team.
Building these assistants became an exercise in internal clarity. We had to articulate our values, strategic focus, and brand language in clear and concise terms. It wasn’t just a tech experiment; it was a mirror that reflected back how well we understood ourselves. In this way, designing our GPTs became a form of strategic alignment and organisational development.
It was less a tech build and more an internal audit. A rare opportunity to align technology with ethos.
We used OpenAI's Custom GPT Builder to define specific instructions for each of our assistants: Calliope (content and design), Hanna (business strategy), and Themis (compliance and governance). And uploaded essential documents, including our business plan, mission statement, brand guidelines, and laws and regulations, to ground each GPT in accurate, relevant context. And finally, we tested across practical and realistic scenarios. Each round of testing sharpened how the GPTs understood our tone, goals, and priorities. We experimented with specific prompts and realistic tasks, observing how each GPT adapted to different content formats and contexts. From Calliope's responses to visual briefs to Themis's handling of compliance checklists, we continuously refined their behaviours through feedback. We taught them, as we would teach a new colleague on specific tasks. We weren’t chasing perfection; we were building tools that embed seamlessly into our daily workflow and elevate the quality of our work.
What Surprised Us
Within days, these GPTs weren’t just supporting tasks, they are challenging and elevating our thinking. Hanna helped restructure an EU grant proposal in hours. Themis flagged a compliance blind spot that had been missed by humans. Calliope drafted a visual brief with fewer iterations than usual.
They also exposed holes in our documentation. Useful AI requires clear input. Where our thinking was fuzzy, their output faltered, prompting us to codify and clarify.
Lessons for Others Building GPTs
Start with your values, not features.
Train with real artefacts, such as memos, decks, templates. That’s where your DNA lives.
Test beyond content teams. Strategy, ops, compliance, each needs different AI support.
Keep a human in the loop. AI is a collaborator, not an oracle.
Final Thought: Technology as a Mirror, Not a Crutch
Our GPTs didn’t just automate. They illuminated. They pushed us to clarify, codify, and deepen our internal logic. And that’s the real win: AI that scales clarity, not noise.
So no, we’re not outsourcing our brains. We’re codifying our best thinking so we can build faster, better, and true to what ISSEAD stands for.
This post was written with the assistance of AI-Calliope, ISSEAD’s own customized GPT.