How JAWS Group's Agency CEOs Automated Creative and Competitive Intelligence in One Day
JAWS Group is a Paris-based holding company that unites five specialized agencies — Splashr (video), Sparkle (social & influence), Blinked (paid media), Arkée (SEO), and Place To Be Media (ad space buying). Together, they serve over 1,000 brands including LinkedIn, Citroën, Tropicana, Yves Saint Laurent, McDonald's, and Cartier.
Their agency CEOs aren't developers. They're operators who manage creative pipelines, client relationships, and strategic decisions. But they wanted to understand what AI coding tools could unlock for their day-to-day work.
We spent one day with them. By the end, they had built two working automations from scratch.
The Challenge: Agency Leaders Drowning in Repetitive Work
Running a creative agency means high-volume, high-variety execution. Every client needs fresh assets. Every pitch requires competitive intelligence. Every campaign needs monitoring.
The pain points were operational:
- Generating image variations for client campaigns required manual back-and-forth with creative teams
- Competitive analysis meant manually browsing Meta Ads Library, screenshotting, and compiling reports
- CEOs spent hours on work that felt strategic but was actually mechanical
- No one on the executive team had coding experience, so automation felt out of reach
The assumption was clear: automation is for engineers. Agency leaders just manage.
We challenged that assumption.
The Solution: A One-Day AI Coding Bootcamp for Executives
The format was simple — one full day, all the CEOs of JAWS Group's agencies in the same room, laptops open, building real things.
No slides-only sessions. No theoretical overviews. We installed Claude Code on their machines and started building within the first hour.
The Mental Model Shift
The first thing we established: you don't need to know how to code. You need to know how to describe what you want.
Claude Code turns natural language into working scripts. The CEO's job isn't to write Python — it's to specify the outcome clearly and review the result. This is the same skill they use when briefing their creative teams, just applied to a different medium.
We covered:
- What Claude Code is — an AI agent that writes, runs, and debugs code autonomously in your terminal
- How to prompt it — describe the workflow you want to automate as if briefing a junior developer
- How to iterate — review the output, give feedback, refine until it works
Use Case 1: Bulk Creative Asset Generation
The first automation we built together tackled a daily pain point: generating multiple image variations for client campaigns.
Agency teams routinely need to produce dozens of visual assets — different formats, copy overlays, and compositions — for a single campaign. This work is repetitive but not trivial: each variation needs to match brand guidelines and feel intentional.
Using Claude Code, we built a script that:
- Takes a brief describing the visual style, brand colors, and messaging
- Generates multiple image variations using AI image generation APIs
- Outputs them in the formats needed for different platforms (Instagram stories, feed posts, LinkedIn banners)
What previously required a full creative cycle — brief, design, review, revise — was compressed into a workflow that any CEO could run from their laptop. Not to replace the creative team, but to generate first drafts and speed up the iteration loop.
Use Case 2: Competitive Intelligence from Meta Ads Library
The second automation was more strategic: scraping Meta Ads Library to audit a competitor's advertising strategy.
For one of their brands, we built a script that:
- Scrapes Meta Ads Library for a competitor's active campaigns
- Extracts key data: ad creative types, copy, launch dates, and engagement signals
- Generates a structured audit report showing what's working for the competitor
- Highlights patterns in creative format, messaging strategy, and spend allocation
This is intelligence that Blinked (their paid media agency) collects manually for pitches and strategy reviews. Automating it meant the CEO could walk into a client meeting with a fresh competitive audit generated in minutes, not days.
The Results: Non-Technical Leaders, Real Automations
What Changed in One Day
- Creative asset generation: From full creative cycle to first-draft automation — CEOs can now generate image variations independently
- Competitive intelligence: From hours of manual browsing to an automated audit script that runs in minutes
- Mindset shift: Agency leaders now see coding tools as extensions of their strategic toolkit, not engineering-only territory
The Bigger Picture
The real output wasn't the two scripts. It was the realization that dozens of other workflows — client reporting, content calendar generation, SEO audits, influencer outreach — follow the same pattern: repetitive, structured, and automatable.
Each CEO left with:
- Claude Code installed and configured on their machine
- Two working automations they built themselves
- A framework for identifying and automating their next workflow
Key Lessons for Agency Leaders
1. You Don't Need to Code. You Need to Brief.
The skill gap between agency executives and AI coding tools is smaller than anyone thinks. If you can write a clear creative brief, you can prompt an AI agent to build automation. The input is the same: clear objectives, constraints, and expected output.
2. Start with What Hurts Daily
Don't try to automate your most complex process first. Pick the task that eats an hour of your day and feels mechanical. For JAWS Group, that was image generation and competitive research. For your agency, it might be reporting, content scheduling, or proposal formatting.
3. Automation Augments, It Doesn't Replace
The image generation script doesn't replace Splashr's creative team. It gives them a head start. The competitive audit doesn't replace Blinked's strategic analysis. It gives them raw material faster. The best automations compress the boring part so humans can focus on the interesting part.
4. Train the Leadership, Not Just the Team
Most AI training targets individual contributors. We went straight to the CEOs. When leadership understands what's possible, they can identify opportunities across the entire organization — not just within one team's workflow.
Want to make your team autonomous with AI?
We help non-technical teams build real automations in a single day. No coding experience required. Book a call and we'll walk through your workflows to find the highest-impact opportunities.
This case study documents a training session completed in February 2026 with JAWS Group's executive team in Paris.