In partnership with

Good morning,

Claude and ChatGPT have had another upgrade. But the bigger opportunity is in the work you can now hand over.

You may still be doing the planning, moving information between tools and telling AI what to do next. Today’s story will make you reconsider how much of that job still needs you.

Today:

Black Friday sales are already going live. Plus, Frøya’s offer gives shoppers three reasons to add another item.

In AI Launch Codes: AI has grown up. Are you still giving it one instruction at a time? We show three complete business assignments you can hand over, with the prompts to get started.

Presented by Bezel

30+ staff in the US or Canada? AI labs will pay for your data

AI labs are spending $15B this year on real business data: the workflows, docs and Slack threads your team already creates.

For established companies, that’s a potential $100k–$3M revenue line hiding in your SOP.

In the Headlines

Best Buy starts its 10-day Black Friday sale today. Daily doorbusters lead into a limited $100 gift card for $60 in stores on November 27. The campaign gives shoppers fresh reasons to return. Check whether your sale has a second reason to visit after the launch email.

Target sets its main Black Friday event for November 22–29. Its holiday plan combines deals, exclusive products and a gift card giveaway. For smaller brands, pick one reason your offer deserves attention alongside the biggest retailers, and make it visible in the first screen.

Google adds promotion-aware bidding to Demand Gen. Seasonality Adjustments can now account for an expected conversion-rate increase during a short sale. If you use Demand Gen for BFCM, base the adjustment on your expected conversion lift and review the actual result.

On Socials

1/ Your discount can cut profit twice as fast. Ordoro shows how a 25% discount changes a sample order’s economics once shipping costs stay fixed. Before extending your BFCM offer, calculate what remains per order after fulfillment.

2/ Holiday shoppers want AI’s advice before they trust it to buy. Salesforce’s new survey covers more than 2,500 shoppers across four countries. Make product comparisons, delivery details and returns easy to find for the shopper who researches with AI and completes the purchase themselves.

Idea Worth Testing

Give the next item a reason to enter the basket

Frøya’s early Black Friday offer pairs discounts with gifts that accumulate as the order grows. Its ad copy sets out three thresholds:

  • $80+ — bamboo face cloth.

  • $125+ — add a tote bag.

  • $150+ — add a full-size hand balm.

Frøya early Black Friday ad advertises up to 60 percent off plus free gifts

The ad leads with the sale and free gifts. Its accompanying copy explains the three thresholds.

The next reward gives a shopper a concrete reason to add another item. Test one gift threshold just above your usual basket value, show the amount left to unlock it, and compare contribution profit per visitor with your current offer. Include the gift and extra shipping in the calculation.

Make spending a little more feel worthwhile. See Frøya’s ad and page in the Funnel Library.

AI Launch Codes

Are You Giving AI Too Little to Do?

The biggest AI upgrade this week changes what you can hand off.

At six, she stands on a stool to reach the kitchen counter. Her mother shows her how to crack an egg. A piece of shell falls into the bowl, and they fish it out together.

Next comes the whisk. Then the pan. Then the reminder to turn down the heat.

For years, breakfast works like this: one instruction, one action, the next instruction.

A child learns to cook step by step and later cooks independently

Then the stool disappears.

She makes breakfast before anyone else wakes up. Eventually, she leaves home, cooks for other people and learns recipes her mother never taught her.

One morning she visits. She is already making breakfast when her mother walks in and says, “First, crack two eggs.”

She stops and looks up.

That is how many of us still talk to AI.

We ask it to pull a number. Then summarize it. Then write a chart. Then suggest an action. We manage every step, even when the work is one connected problem.

And the models have been growing up.

Anthropic’s Fable 5.2 release describes a research task where the model took on the roles of planner, coordinator, code writer and job submitter. One brief led to a chain of work.

Excerpt from Anthropic’s Fable 5.2 announcement about handling several roles in a complex task

Source: Anthropic’s Fable 5.2 announcement.

OpenAI’s GPT‑6.2 release notes describe Sol as able to work through tasks that take several steps, while Astra is aimed at more demanding professional work. GPT‑6.2 is currently an API release, so access in ChatGPT depends on the product you use.

OpenAI evidence for GPT-6.2 performance on complex work

Source: OpenAI.

For a founder, the change is simple: the next prompt can describe the outcome, the sources to inspect, the checks to make and the finished decision you need.

Here are three jobs that make that shift concrete:

  • Meta ads: move from ad metrics to an investigation that reconciles Meta spend with Shopify sales, inspects creative and landing pages, then shows where the next dollar should go.

  • Creative and landing pages: research what customers actually say, study the ad-to-page argument of a relevant brand, then draft a new concept and page for your own product.

  • BFCM offers: model discounts, bundles and thresholds against margin, acquisition cost and inventory before you choose an offer.

Which ad gets your next $500?

Say Meta shows you two ads. One gets a purchase for $24. The other costs $36. If you’re deciding where to put another $500, the $24 ad looks like the easy choice.

Now look at what people actually bought.

  • The $24 ad: First orders leave $18 each after discounts and variable costs. After ad spend, that’s a $6 loss per order.

  • The $36 ad: First orders leave $42 each before ad spend. That’s $6 left per order after acquisition.

Illustrative numbers. Both ads are measured on the same first-order basis.

That’s the first pass. The second is where it gets interesting. If the $36 ad brings in buyers who choose a higher-margin bundle, can you get more of them with a new creative? Or is the landing page doing the work?

Meta ads analysis across Meta, Shopify, landing pages and an interactive report

One assignment. Several sources. A decision you can act on.

The brief I’d give AI starts with the ads you’re ready to scale. It checks which ones still make money after costs, traces each one to its offer and page, and comes back with a short list of what to review and what to test next.

Start with the ad you’re most tempted to scale. Which orders actually made it look like a winner?

logo

This post is for paid subscribers

We built AI Launch Codes to help you scale smarter. We test AI tools, build workflows, and write prompts—so you don’t have to. But testing takes time, money, and effort. This paid upgrade helps us keep experimenting, while you get the winning playbooks delivered straight to your inbox.

Upgrade

Already a subscriber? Sign in

Why You’ll Want It:

  • Use AI to scale faster, cut costs, and boost efficiency
  • Ready to use workflows, automations and prompts
  • Real world use cases, no theory