GPT needs this creative fix
Plus, 🤖 Create smart agents for free

Hello Readers 🥰
Welcome to today's edition, bringing the latest growth stories fresh to your inbox.
If your pal sent this to you, then subscribe to be the savviest marketer in the room😉
In Partnership with Insense
⏳What Could You Do With 40 More Hours and Zero UGC Chaos?
You’re not short on strategy. You’re short on bandwidth.
Most marketers lose 40+ hours a month managing influencer logistics, rewriting briefs, and chasing DMs, just to get UGC that’s late or off-brand.
Insense gives you all that time back. It’s the UGC execution engine trusted by 2,000+ DTC brands, including Victoria Beckham Beauty, Solawave, and Woxer.
Here’s what those teams no longer touch:
✅ Creator sourcing + outreach to 7M+ vetted influencers
📦 Product seeding, tracking, and communication, all handled
🧠 Licensed, ready-to-use UGC delivered in under 14 days
You can even plug in a dedicated platform manager to run every campaign for you.
You stop micromanaging creators. Your team starts launching faster. Everyone wins.
💸 Get up to 20% off managed services, offer ends June 20.
Book your free strategy call now!
Create Smarter Ads with GPT
Most GPT-generated ads fall flat. They often miss the mark on layout, tone, and visual branding, leaving you with content that looks amateur at best. But with the right structure and prompts, you can flip GPT into a powerful creative engine. Here's how to systematize the process and get 95% of the way to polished ads.
Steps to Generate Better Ads Using GPT:
1️⃣ Build a Brand JSON File:
Don’t rely on vague prompts. Instead, create a structured JSON file that includes your brand name, tone, fonts, color palette, value props, audience, and more. GPT performs much better with this level of specificity.
2️⃣ Ask GPT to Write Its Own Prompt:
Give GPT your JSON and an example ad. Then ask, “How would you prompt yourself to recreate this ad using this data?” GPT will generate a clear, detailed prompt you can use directly or tweak for better results.
3️⃣ Provide Visual Inputs:
Match the reference ad’s structure. Whether it’s a clean product shot or lifestyle image, the closer your input matches the layout, the better the generated output.
4️⃣ Combine Everything for a Draft:
Feed GPT your JSON file, prompt, and visuals. The result? A rough ad layout with copy, brand styling, and value props baked in.
5️⃣ Final Touch by a Designer:
Take GPT’s draft and hand it off to a junior designer for polish. With the concept nearly complete, this step is fast, affordable, and keeps design quality high.
The Takeaway
GPT won’t replace your creative team. But it will cut your ad creation time and cost dramatically, when used strategically.
Build a Browser Agent for Free
You don’t need to spend $200 to create your own AI browser agent. Thanks to open-source tools like DeepSeek R1 and Browser-Use, you can build a fully functioning web-based AI agent without touching a line of code. Here’s how to get started in minutes.
No need to pay $200 to use Operator
— Paul Couvert (@itsPaulAi) January 24, 2025
You can create an agent that uses a web browser without writing a line of code.
Combine DeepSeek R1 and Browser Use (free and open source) and you're good to go.
(Links and prompt below) pic.twitter.com/NBhzv7AeE3
Steps to Create Your Browser Agent:
1️⃣ Use DeepSeek R1:
Start by installing DeepSeek R1, a powerful open-source LLM that provides the intelligence for your agent. It's fast, efficient, and entirely free to use.
2️⃣ Add Browser-Use:
This tool acts as your agent’s eyes and hands on the web. It gives your AI the ability to search, click, and extract content directly from browser windows.
3️⃣ Combine the Tools:
Once you have DeepSeek R1 and Browser-Use set up, it’s time to connect them. You’ll do this by writing a prompt that clearly explains what you want your agent to do—think “search Google and summarize results.”
4️⃣ Use the Provided Prompt Template:
To make life easier, use this ready-made prompt structure:
Using Browser-Use (documentation attached), create an agent that [insert task].
Write the full setup for a beginner using [your IDE] on [your OS]. Prefer ‘uv’ for Python if applicable.
Include an example from the Browser-Use repo.
5️⃣ Copy-Paste the Docs:
Drop in the full Browser-Use documentation so your LLM knows what tools it’s working with.
The Takeaway
You don’t need expensive subscriptions to build powerful AI agents. With the right open-source tools and a bit of structure, you can deploy browser automation smarter, faster, and for free.
We'd love to hear your feedback on today's issue! Simply reply to this email and share your thoughts on how we can improve our content and format.
Have a great day, and we'll be back again with more such content 😍