AI Growth Marketing Lead (UK startup)

We think the next great marketer isn't a specialist on a team of eight. It's one person who understands the customer better than anyone else in the room, and who has built a set of AI tools that let them research, write, test and ship at a speed a team of eight can't match.

That's the job. You own creative strategy for paid social — the research, the angles, the hooks, the briefs — and you follow the customer the whole way through: the landing page they hit, the offer they're weighing up, the checkout they abandon, and the email that brings them back. You'll build the AI system that does the heavy lifting alongside you, and you'll do it with the founder, in the room, with nothing set up yet.

How we work: AI does the heavy lifting

This is the part of the role that makes it different from every other growth job you're looking at, so we want to be explicit about it.

We are building a high-end AI marketing system, and we expect AI tools and agents to carry a large share of the actual work — the research sweeps, the first drafts, the data pulls, the competitor teardowns, the variant generation, the reporting. Your job is not to do all of that by hand. Your job is to build the tools that do it, direct them, and be the quality bar on what comes out.

• You'll build the tools, not just use them. Research agents, creative-analysis agents, briefing agents, reporting agents — anything that removes work you'd otherwise repeat next week.

• You'll get properly good at Claude Code / Codex and agent workflows. You don't need to write code. You need to be dangerous with prompts, context, structure and tooling, and to recognise the moment a task should stop being manual.

• Output at volume, without dropping quality. Ten times the throughput is the point, but only if the work still lands. You own the judgement layer that AI doesn't have — taste, customer knowledge, and knowing what's actually true.

• It compounds. Every tool you build is leverage you keep. Six months in, you should be able to point at a system, not just a set of results.

What you own

Creative strategy for paid social

• Mine reviews, DMs, comments, support threads, Reddit and competitor ad libraries for the language customers actually use — mostly through research agents you've set up — and turn it into angles worth testing.
• Write the hooks, scripts and concepts. Brief creators, editors and designers so what comes back is what was in your head.
• Live in the ad account. Read creative-level data, spot the pattern early, cut what's dead, and know the difference between a hook problem, a creative problem and an offer problem.
• Watch the organic feed properly. Know why something is working on TikTok or Reels before it's obvious to everyone else, and know how to translate it without copying it.

Research and customer understanding

• Own the source of truth on who our customer is: the problem, what they've already tried and failed at, what they're quietly afraid of, and what they'd never admit in a survey.
• Apply real behavioural psychology — objection, identity, status, loss aversion, timing — as testable creative decisions, not as jargon on a slide.
• Make pattern recognition repeatable. A swipe library, a scoring system, a research agent — whatever means the insight doesn't live only in your head.

CRO, offers and the funnel

• Own conversion end to end: ad → landing page → offer → checkout → first purchase.
• Structure and brief landing pages yourself; a designer and developer build them. You decide what goes above the fold and why.
• Design and run offer tests. Understand how much of performance the offer carries, and how to change it without eating the margin.
• Set up and read tests honestly — sample size, what actually counts as a result, when to call it, and when to admit it was noise.
• Personalisation: different message, different page, different sequence, depending on where someone came from and what they care about.

Email, retargeting and lifecycle

• Own the flows and campaigns: welcome, abandoned checkout, post-purchase, win-back, reactivation.
• Segment on real behaviour and real signals — not a first-name token and a discount code.
• Keep one voice across every surface. The ad, the page and the inbox should obviously come from the same brand.

Who we're looking for

Must have

• 7+ years in growth, performance or creative strategy — ideally DTC, ecommerce or subscription, ideally something people buy to fix a personal problem.
• Genuinely good writing. Short-form, direct response, voice-driven. We will judge this hard.
• Comfort in an ad account. You can read creative-level data and reach a decision without a dashboard being built for you.
• Real understanding of our customer and the problem we solve — the kind that comes from having lived near it, not from reading the deck.
• A grasp of CRO and funnel mechanics. You know what moves someone from scroll to click to buy, and where they usually fall out.
• Technical curiosity. You've already pushed AI tools well past the chat box — automations, agents, custom workflows, something you can show us.
• Evidence you learn fast. We'll pay for courses and give you the time. You have to actually do them.
• Comfort with no structure. There is no playbook here. You're writing it.

Helps, but isn't essential

• Klaviyo or an equivalent ESP, Meta Ads Manager, Shopify.
• Briefing designers in Figma, or building pages in a no-code page builder.
• You've already built an agent or automation that saved you hours a week — and can show it.
• Basic HTML/CSS literacy, purely so you can talk to developers / AI coding agents

Read before applying

This role is intense, and we'd rather you self-select out now than three months in.

• It's founder-led and fast. Feedback is direct and unfiltered. Decisions get made in hours, and sometimes reversed the next day when the data says so.
• Nothing is set up. No process, no playbook, no team beneath you, no one to hand a half-formed brief to. You'll build the thing and then run it.
• The scope moves. Some weeks are creative strategy. Some weeks are rebuilding a landing page and a flow because the numbers changed.
• You'll be asked to work in ways you haven't before, with AI systems doing a real share of the work and you responsible for the quality of what they produce.

What you get in exchange: you build the system rather than inherit it, you hold the leverage that comes with that, you learn faster than you would anywhere with a process already in place, and the bonuses are tied to numbers you personally move.