The result first: after we took over The Outfields’ social media and rebuilt it around an AI workflow, social interactions rose 45% and qualified leads grew 20%. Here is exactly how we did it, the tools, the steps, and where a human still made every call that mattered.
The 60-second version
- The Outfields is a brand-new outdoor venue with almost no audience. The job was reach and enquiries, fast, on a lean budget.
- We built a repeatable system: Claude for strategy, captions and repurposing, Veo3 for short-form video, with a human approving everything before it shipped.
- Outcome: +45% interactions, +20% qualified leads, and a content pace a one-person posting schedule could never match.
- The point of AI here was throughput, not replacing judgement. The brand voice stayed human because a human owned it.
Most “we use AI” claims are a homepage badge with nothing behind them. This is the opposite: the actual pipeline we run, written up so you can judge it. The Outfields makes a good case study precisely because it started from zero, a new venue, a tiny following, and a need to look established before it was.
The problem: a beautiful venue nobody had heard of
The Outfields is an open-air event venue in East London. Gorgeous in person, invisible online. A new venue faces a brutal chicken-and-egg problem: you need content and social proof to win bookings, but you have no events yet to photograph and no audience to show them to. Hiring a traditional social agency to fill that gap is slow and expensive. Posting once a week from a phone does not move the needle.
So we treated content like a production line, and used AI to run the line at a pace a lean budget could afford.
The workflow, step by step
1. Strategy and calendar (Claude)
We started by giving Claude the venue’s real facts, capacity, the multicultural niche, the booking calendar, and had it help map a month of themes: open-sky weddings, Nikah and cultural celebrations, corporate summer parties, supplier spotlights. Claude is good at turning a messy brief into a structured plan quickly. We edited it down to what actually fit the brand.
2. Captions and repurposing (Claude)
For each theme, Claude drafted captions in the venue’s warm, sensory voice, then repurposed one idea across formats: a long caption for the feed, a short hook for a reel, a few lines for a story. The first drafts were never the final ones. We rewrote anything that sounded generic, because generic is what kills a brand account.
3. Short-form video (Veo3)
A new venue with few real event photos has a content hole. Veo3 let us generate atmospheric short-form video, golden-hour fields, festoon lights, long tables, to carry the feed while real event footage built up. Every clip was checked to make sure it looked plausible for an East London field, not a generic stock dream.
The guardrail that mattered most: nothing published without a human read. AI set the pace; a person set the standard. That single rule is the difference between an AI content engine and AI slop.
4. Schedule, publish, learn
We batched a month at a time, scheduled it, then watched what actually landed. The posts that drove saves and enquiries told us what to make more of. Claude helped spot the patterns; the decisions stayed ours.
The results
+ Veo3
AI did not write the brand. It let one small team run a content engine that usually takes a department.
The interactions number matters because it is reach the venue did not have before. The leads number matters more, because for an enquiry-only venue, a lead is the whole game. The content looked consistent and on-brand the entire time, which for a new business is half the battle: looking established before you are.
What this means for your brand
If you are a lean team or a founder who needs to look bigger than you are, this is the model: AI for throughput, a human for judgement, a system you can actually repeat. It works for a venue, a fintech, a café, a fashion label. The tools change per brand; the discipline does not.
Want a content engine like this?
Our Growth package runs exactly this workflow for your brand: AI-built content, human-approved, on a system that scales. Book a call and I’ll show you what it would produce in your first month.
Frequently asked questions
How do agencies use AI for social media?
The strong version is a system, not a gimmick: AI drafts strategy, captions and short-form video at speed, while a human edits for voice and approves everything before it publishes. AI handles throughput; people handle judgement and brand.
Does AI-generated content hurt your brand or SEO?
Only if you ship it raw. Unedited AI content reads generic and erodes trust. Edited, fact-checked, on-voice content performs well. The deciding factor is whether a human owns the final word.
What did The Outfields actually gain?
Social interactions rose 45% and qualified leads grew 20%, alongside a consistent, on-brand feed that made a brand-new venue look established. For an enquiry-only business, the lead growth is the headline.
Can a small team run this without a big agency?
Yes, that’s the point. The workflow is designed so one or two people can produce department-scale output. We set it up and either run it for you or hand it over.
Written by Uzair Amjad, founder of THETOGLAB, a London design and marketing studio. We build AI-powered content engines that let lean teams punch above their size.