Focal

How does Granola run performance creative ops to ship more winning ads? Join Creative Strategist Naomi Peh Haeger on Thu, Sep 24 to find out: - How Naomi structures sprints and routines to keep output moving - Naomi's take on the optimal creative mix on Meta and beyond - How Claude and other AI tools fit in the process, where do they bring value and where they fall short - What a practical performance creative operation actually looks like in 2026 If you want a hype-free look at how one of the sharpest creative operations out there gets it done, this is the session for you.

How is Granola running creative strategy and building winning performance creative - and how you can do it too? Find out on Thu, Sep 24. Seats are going fast so secure yours today: https://lnkd.in/dqjFu-RD Granola's Naomi Haeger will be sharing her thoughts on optimal creative sprints and processes, where AI tools add true value, the best creative mix on Meta (is there such a thing?) and much more. See you there! 😎

Time to bust out my acting chops (x2) for today's launch! Focal Workflows make the annoying chores of creative asset management a breeze: no more manual ad platform uploads, writing pings to media buyers by hand, and more. See for yourself ⬇️ (This is a simple demonstration with a few assets, but imagine what this can do for you if you're launching 100+ ads every week across multiple platforms). Unlike 3rd party task automation like Zapier, Focal Workflows actually understands your creative production: projects; tasks and associated fields; creative assets and their tags, metadata and visual content. You can hand any item to an AI model for summaries or deeper analysis, update values, send Slack alerts, upload to ad platforms, and it can even send and receive data to other platforms (Zapier included). You can learn more from the link in the comments. And if you want to take this into use, just reach out and I'm happy to help.

Hook analysis is of my favorite Focal MCP use cases. When new assets are first uploaded to Focal, we start by auto-tagging your hooks based on visuals, voiceover, in-asset copy and music. This is fully customizable by brand, so you can tag the message, copy and audio hooks separately for more granular analysis. We then tie performance data back to the assets, making it easy and quick to use Claude to generate summaries and reports.

One of the best ad inspiration and intel sources is new ads other relevant brands are launching. So every Monday morning I get a structured breakdown of new creative angles others launched last week. There's too much material to go through manually, so you need something to digest it all in a meaningful way though. What most people end up doing is zooming in on the few examples that catch their eye the most, but that doesn't really help you get the big picture. The Meta Ads MCP helps quite a bit already for stuff like understanding ad copy strategies at scale, but it's still limited in what you can do in terms of visual analysis. That's why I've found Focal MCP so helpful. I can dissect the ad creative (and copy) using taxonomy I care about the most, and can get a detailed analysis of what others in my space are doing with visual examples. This makes it much easier to adapt interesting angles into my own briefs or adjust our positioning for example. What are some of your favorite ad inspiration or intel sources and how do you analyze them in a meaningful way?

Higgsfield AI is one of my favorites for video storyboarding. If you need a concept with 20 shots, just use the MCP and give it a spreadsheet with the shots. It's still a lot of setup and juggling reference material though, so I combined it with Focal: - I write the creative concept shot by shot in Focal, each with its own brief and generation instructions. - When a shot needs a specific kind of reference, I can just look for one in my library. - AI search finds the exact scene in seconds e.g. based on visual content, or I can even filter for what actually performed best in ads (so you only use winners to generate new stuff). - Then I run Claude: it reads the brief from Focal, generates the shot with Higgsfield, and pushes the finished clip back into Focal automatically. There's no manual exporting or spreadsheets to manage. - From there, the AI generations become storyboard references for the real shoot, or could even go straight into the final cut. - Once every shot has the right material, it's easy to move to final edit Happy to share in more detail. DM me or drop a comment if you'd like to see more.