April, 2026.- In the fast-paced 2026 agency ecosystem, lost time isn’t found in execution, but in the reconstruction of context. Conor Hoey, a sharp mind who has identified the fragility of the creative process under pressure, presents First Concepts: the first “operating system” designed not to generate content faster, but to provide teams with coherent memory. For Hoey, the industry’s fundamental problem is the “context loss” occurring between the brief, references, and final delivery—a gap that costs agencies hundreds of thousands of pounds and thousands of hours in rework. His proposal is revolutionary: an intelligence layer that learns the “taste” and “sensibility” of a creative director, ensuring that AI tools don’t deliver the most probable algorithmic response, but the specific, intentional vision of the human team.
In this exclusive interview with Roastbrief, Conor Hoey breaks down how his platform is already being tested by powerhouses like Mother London and R/GA, achieving 70% time savings by eliminating exhausting correction cycles. Hoey challenges the notion that technology should focus solely on speed; instead, he proposes a “Creative DNA” infrastructure where past decisions inform future innovation, preventing agencies from circling the same well-worn territories. Discover how First Concepts is turning ephemeral “taste” into a structured asset and why the million-dollar pre-seed backing is just the start of a transformation aimed at giving creatives back the energy to push the boundaries of the new, without having to start from scratch every time.
- The Creative Operating System: You’re building First Concepts as an “operating system for creative work”—one that understands creative minds and how creative teams work. What does an operating system for creativity actually look like, and what problem in the creative process are you solving that existing tools don’t address?
The problem is context loss. Agencies pitch up to ten times a month, spending as much as £200k per pitch and up to 40% of creative time gets lost reconstructing context across fragmented tools. The brief lives in one place, references in another, feedback changes direction and nothing remembers any of it.
An operating system for creativity sits above the stack. It coordinates leading models and creative tools through a shared taste and context layer. Brief, references, decisions, output, all connected in one persistent space. When you move from research to ideation to execution, the thinking travels with you. When someone new joins the project, they inherit the full picture.
No existing tool does this. They’re all point solutions. We’re the connective tissue. Our mission is to transform creative context from something fragile and ephemeral into a structured, compounding asset.
- Learning Style, Taste, and Decision-Making: First Concepts learns your style, preferences, and the way you make decisions, then carries that across different tools, people, and projects. How do you train the system to understand something as subjective as “taste”? What’s the technical and philosophical challenge in teaching AI to recognize creative sensibility?
Taste isnʼt a setting, itʼs a pattern of a thousand decisions. What you keep, what you kill, what you reference, what you reject.
The platform is built on three foundational pillars: a Creative DNA taste engine that learns individual and brand judgement, a unified interface that treats context as the source of truth and a tool-agnostic infrastructure that ensures creative coherence across every output. Every time a creative director pulls a reference, rejects an output, or chooses one direction over another, that’s a signal. Creative Memory captures those signals and builds a living picture of how a team thinks. This isnʼt static, it evolves with every project. The technical challenge is making that work across different tools, team members and types of creative work.
- Not Starting from Scratch: One of the most compelling promises is that “you don’t start from scratch every time”—the system remembers how your brain works. For creative teams that thrive on fresh thinking, is there a risk that consistency becomes rigidity? How do you balance continuity with the need for surprise and novelty?
Thereʼs a misunderstanding in the question. Consistency doesnʼt create rigidity. Forgetting does. Most teams donʼt start fresh, they unknowingly repeat themselves. Same territories, same references, just reassembled under pressure. Thatʼs not novelty, thatʼs drift. Creative Memory makes that visible by tracking reasoning, not outputs. It knows what was explored and why it didnʼt land, so youʼre not circling old ground without realising it.
The balance is in the roles. The system preserves continuity. The creative decides when to break from it. We donʼt generate ideas or steer direction, we hold the context so it doesnʼt collapse as work moves. That gives teams the confidence to push further because theyʼre not wasting energy rebuilding or second-guessing. The best creative directors have extraordinary memory. Weʼre giving that same memory to entire teams.
- The Felix Richter Endorsement: Felix Richter at Mother London said the output “actually reflects your references and taste. You’re not constantly pulling it back to where it should be.” Why is this such a critical pain point for creative teams? What does it cost agencies in time, energy, and creative quality when they’re constantly fighting to align output with intent?
The pain point is correction cycles. Right now, when a creative team uses any AI tool, the output almost never matches the intent on the first pass. So senior people spend hours pulling it back. Re-prompting,
refining, overriding. That’s expensive, exhausting, and it erodes trust in the tools. Multiply that across ten pitches a month and you’re talking about hundreds of hours a year spent fighting your tools instead of using them.
What Felix described is the absence of that fight. The output reflected his references and taste from the start. That’s not a feature. That’s the whole point. If the system doesn’t understand your intent before you correct it, you’re just doing quality control on a machine.
- The Pre-Seed Backing: You’ve raised $1M from Arāya Ventures, Antler, and industry names like Jez Jowett and Nathan McDonald. What did those investors see in First Concepts that convinced them to back you, and how will their industry experience shape the product as you move toward launch?
Majority of creative tools are solving the wrong problem. Everyoneʼs racing to generate faster. Nobodyʼs fixing the context collapse underneath. Thatʼs where time is lost, ideas degrade, and teams fall out of sync. What weʼre building with Creative Memory is an infrastructure layer. It compounds. The more a team uses it, the more it understands how they think, how they decide, what they care about. Thatʼs a very different kind of value and a much harder thing to replicate than another generation tool.
The people backing us have spent decades inside this industry. Theyʼve seen every wave of creative tech come and go. They know where adoption breaks, where workflows fall apart and why most tools never become essential. They didnʼt need convincing that the problem exists. They just needed to see someone building the right layer to solve it.
- Closed Beta with Awarded Agencies: You’re already in closed beta with 40+ awarded agencies across London and New York, including Mother and R/GA. What have you learned from those early partners, and how has their feedback shaped the product ahead of your April 14th launch?
We learned that speed isnʼt the selling point. Coherence is. Teams didnʼt care that the AI was fast. They cared that it stopped losing their direction. Most of the time in a pitch isnʼt spent creating, itʼs spent rebuilding context. Rewriting prompts, re-explaining decisions, pulling work back on brief. The reported 70% time savings came from removing that rework, not making generation faster.
The other big shift was around noise. Nobody wants more features. They want something that actually holds the thinking together. Thatʼs shaped how weʼve built the product. It doesnʼt feel like another tool in the stack. It feels like working with something that remembers what youʼre trying to do and keeps you on track.
