The price you start with is not the price you keep

Fifty pounds per seat per month is a defensible starting price. It sits in the right neighbourhood: above the moodboarding tools like Milanote, which charge around ten pounds a month, and below Figma's enterprise tier at roughly seventy-five pounds. The comparison that matters most is Figma Organisation at fifty-five dollars, because that is the tier where agencies start getting workflow controls, governance, and admin features on top of the core design tool. If you position First Concepts as a creative workspace with governance and audit built in, rather than as a fancier moodboard, the fifty-pound price holds up. If the market perceives you as a moodboarding tool, it does not.

But per-seat pricing has a structural problem when your product runs on inference. Every image generation, every video evaluation, every brief analysis costs you real money on the backend, and those costs scale with usage, not with the number of names on the account. A heavy user running forty concept generations a day and a light user who logs in once a week to check a board cost you wildly different amounts to serve, yet they pay the same. At low volumes this is fine. The generous margins of early SaaS absorb it. As you scale, and especially as agencies start leaning hard on the generation capabilities during pitch crunch periods, it becomes a margin problem. The agencies that love you most will be the ones that cost you the most to serve.

This is why the seat price should be treated as a starting position, not a permanent architecture. It gets you in the door. It makes procurement straightforward. It lets you quote a number in a conversation without reaching for a calculator. But behind it, you need to be building toward something more sophisticated.

Seats plus credits: the hybrid that protects your economics

The recommended evolution is a hybrid model: a base platform fee per seat, plus a usage layer priced in credits. Each plan includes a generous allocation of credits, enough that most users never think about it, with clear overage rates for teams that exceed the bundle. The base fee covers the workspace, the collaboration features, the context and memory layer, the governance tools. The credits cover the expensive part: model inference, image generation, video processing, transcription.

This is not a novel structure. Most AI-native tools are converging on some version of it, because the underlying cost structure demands it. The important thing is how you package it. Agencies hate surprise bills. The fastest way to get thrown out of a procurement process is to hand someone a usage-based invoice that is three times what they expected. So the credit bundles need to be generous enough to feel like flat-rate SaaS for normal usage, with overages positioned as a power-user feature rather than a trap. Pre-paid credit blocks and hard spending caps for enterprise accounts give procurement teams the predictability they need while letting your margins breathe.

The transition from pure per-seat to hybrid does not need to happen on day one. You can introduce credits as an internal accounting mechanism first, tracking usage costs per workspace without surfacing them to the customer. This gives you the data to set the right bundle sizes before you ever change the pricing page. When you do introduce the credit layer publicly, frame it as giving teams more control and flexibility, not as a price increase.

Pitch packs: pricing in the units agencies actually think in

There is a more radical packaging option that deserves serious consideration, especially for your go-to-market in the first twelve to eighteen months. Agencies do not think in seats. They think in pitches. A new business director does not wake up wondering how many software licences the team has; they wake up wondering how to win the Unilever review. When someone in your workshop said they sometimes spend up to two hundred thousand pounds on a single pitch, they were telling you something important about how budgets flow inside agencies.

A pitch pack bundles everything needed for a defined number of pitch projects: seats, generation credits, the workspace, the audit trail. You charge per pitch or per block of pitches per month. This makes the value proposition concrete and immediate. The new business lead can say: "This pitch costs us fifteen thousand pounds less because we used First Concepts," or "We got three concepts to presentation-ready in the time it used to take us to do one." That is a sentence that moves budget. "We have twelve seats on a SaaS tool" is not.

The downside of pitch-based pricing is that it can make First Concepts feel like a point solution, something you spin up for pitches and ignore the rest of the time. If the long-term vision is to be the persistent creative workspace where context and taste accumulate over years, you need habitual daily use, not just pitch-week surges. So pitch packs work best as an entry-level offering or as a named tier alongside a broader workspace subscription. They are a wedge, not the whole architecture.

The governance tier: an unusual advantage

Most creative tools treat security and compliance as a grudging afterthought, a checkbox to tick on the way to an enterprise deal. First Concepts has the opportunity to make governance a primary reason to buy, not just a procurement hurdle to clear.

The audit capability described in your workshops, where the system captures every prompt, every model used, every asset generated, every approval decision, and can export it as a client-ready disclosure report, is genuinely unusual in the creative tooling space. Agencies are under increasing pressure from clients to disclose AI usage in creative work. Some clients mandate specific models. Some prohibit certain providers entirely. One workshop participant described a client relationship where only Google models were permitted, because the client had an exclusive technology partnership. This is not an edge case. It is the direction the industry is moving.

An AI audit pack, whether sold as a paid add-on or bundled into an enterprise tier, does something rare: it makes a creative tool easier to procure rather than harder. Typically, introducing an AI-powered tool into an agency's stack creates procurement friction, because legal and compliance teams worry about data handling, model training, and intellectual property exposure. If First Concepts can flip that dynamic by being the tool that solves the governance problem rather than creating it, you have an enterprise wedge that most creative startups cannot match.

This is also the natural home for an agency-level platform licence: an annual contract with a minimum spend, generous or unlimited seats, and the full admin and security bundle. Annual contracts with committed spend are what enterprise procurement teams prefer. They are also what gives you predictable revenue and lower churn. The governance and audit capabilities are what justify the conversation.

Who you are really competing against

The competitive landscape looks crowded if you map it feature by feature, but clearer if you think about it in terms of budget and attention. First Concepts is not primarily competing against Milanote, which is a lightweight moodboarding tool at a fraction of the price. You are not primarily competing against Miro or FigJam, which own the workshop and ideation whiteboard space but have little ambition in pitch packaging or AI governance. And you are not competing against Midjourney or Runway, which are generation engines without workflow.

The real competitive pressure comes from three directions. First, the presentation and packaging tools: Figma Slides, Gamma, Tome, Pitch, and the old standbys like Keynote and Google Slides. These compress the time from concept to deck, which is part of what First Concepts does. If an agency can get from brief to pitch-ready deck using Gamma and a shared Google Drive folder, that is budget and attention you are not getting. Second, the general-purpose workspace tools: Notion, Coda, and to some extent Figma itself as it expands beyond design. These can approximate a "single source of truth" for a project, even if they are not built for creative references and visual iteration. Third, and most seriously, the platform bundling threat: Google, Microsoft, and Adobe all have the pieces to assemble something that looks like First Concepts inside their existing ecosystems. The workshop transcripts explicitly flag this fear around Gemini's expanding capabilities.

The defensible position against all three threats is the same: context and memory. If First Concepts becomes the place where an agency's creative taste, references, decisions, and institutional knowledge accumulate and compound over time, switching costs rise with every project. A presentation tool does not remember what you liked about the last campaign. A shared drive does not learn your visual preferences. A bundled AI assistant does not maintain a persistent understanding of your creative standards across dozens of briefs. That accumulated context, what has been described as the "creative memory" layer, is where long-term defensibility lives. It is also, not coincidentally, where the product needs the most investment.

What agencies will demand before they sign

Even independent agencies, once AI touches client work, start behaving like enterprise buyers. The procurement requirements are predictable, and you should plan for them rather than scramble when they come up.

Single sign-on with SAML and user provisioning through SCIM are table stakes for any agency with more than about fifty people. Figma gates these to its enterprise tier. Notion does the same. You should follow the same pattern: include SSO and SCIM in your top tier, because they are features that signal enterprise seriousness and justify premium pricing.

SOC 2 compliance, or at minimum a credible timeline toward it, will come up in almost every security questionnaire from a mid-size or large agency. ISO 27001 is the European equivalent. You do not need to have the certification on day one, but you need a story: what you are doing, when you expect to complete it, what controls are already in place. A clean data processing addendum, GDPR alignment, a published list of subprocessors, and a clear "we do not train on your data" default are non-negotiable.

Model and vendor governance, the ability to restrict which AI models can be used on a per-client or per-project basis, is specific to your product category and is worth highlighting as a differentiator. Most AI tools let you use whatever model the platform has chosen. Letting an agency lock a workspace to approved models only, because a client demands it, is a procurement conversation you can win that your competitors cannot even have.

The question is sequencing. You do not need all of this before your first ten customers. But you need a clear roadmap, and you need to start the SOC 2 process early enough that it does not become a blocker when a larger agency wants to move forward. The audit and governance features should be prioritised ahead of the compliance certifications, because they create immediate product value. The certifications are overhead that unlocks access to larger accounts.

Getting to market: founder-led, event-driven, integration-aware

For a product at this stage, selling to this market, the go-to-market should be founder-led sales supported by events and community, with integration partnerships as an accelerant.

Founder-led sales means exactly what it sounds like: the founders are on every call, running every demo, closing every deal. This is not a limitation; it is an advantage. You learn faster. You hear objections in real time. You can adapt the pitch on the fly. The positioning should shift depending on who is in the room. For a creative director, lead with taste preservation and iteration speed. For a new business lead, lead with pitch economics and win rates. For an operations or risk lead, lead with audit trails and model governance. These are three different stories about the same product, and the founders need to be fluent in all three.

Events and community are a natural fit for a product that lives at the intersection of creativity and technology. The idea of an AI creative festival or "creative-thon," with shared briefs, senior judges, and brand sponsors, is a smart one. It creates social proof, generates case studies, and puts the product in the hands of exactly the people you want using it, in a context where they are motivated to push it hard. The outputs from these events become marketing assets.

Integration-led growth means building visible connections with the tools agencies already use: Figma, Slack, Google Drive, Frame.io, and the major project management platforms. A browser extension that works across the web, capturing references and feeding them into the First Concepts workspace, is the simplest version of this. Published integrations and templates lower the barrier to adoption and signal that First Concepts plays well with existing stacks rather than trying to replace them.

Product-led growth, a free tier or self-serve sign-up flow, should wait. It was discussed in your workshops, and the instinct is understandable: let people try it, let the product sell itself. But a free tier is expensive when your product runs on inference, and self-serve onboarding is hard to get right for a tool that needs context and configuration to deliver value. You can revisit this once the product is more stable, the unit economics are well understood, and you have hard usage caps that prevent a free tier from becoming a cost centre.

The recommended commercial architecture

Bringing this together, the commercial model for First Concepts should be built in layers, starting simple and adding complexity as the customer base and product mature.

At launch, keep the fifty-pound per-seat monthly price as the standard offering. It is simple, defensible, and gets deals done. Alongside it, introduce pitch packs as an alternative entry point, priced per block of pitch projects with seats and credits included, aimed at new business teams who want to try the tool on a specific, high-stakes use case.

Behind the scenes, begin tracking AI usage costs per workspace in credits, even before surfacing this to customers. This data will tell you when and how to introduce the hybrid pricing layer.

Within six to twelve months, introduce tiered plans:

• A standard tier at the current price point with a generous credit bundle, for smaller teams and lighter usage.
• A professional tier with higher credit allocations, the AI audit pack, and project-level access controls, for agencies doing regular pitch work.
• An enterprise tier with annual contracts, SSO and SCIM, model governance controls, unlimited or high-cap seats, and dedicated support.

The audit and governance capabilities should be visible from the professional tier upward, because they are a differentiator, not just an enterprise upsell. The creative memory features, which capture and compound an agency's taste and decision patterns over time, should be present across all tiers but become more powerful and configurable at higher levels.

The north star is an agency paying an annual platform licence that covers the entire creative department, with governance and audit baked in, and usage-based pricing that scales with activity without creating bill shock. That is the deal structure that produces high retention, predictable revenue, and defensible margins. Everything before it is a step on the way there.