Master Copywriting in the Age of AI

This Week

Michael Spencer questions if the iPhone’s AI features are groundbreaking, OpenAI releases “deep thinking” models, Uber execs reveal how genAI streamlines customer support, and Scott Martin predicts 80% of copywriters will vanish if they don’t master AI. Plus, BCG’s framework for driving business success with AI.

Read on.

Is Apple Intelligence Actually Useful?

At Apple’s latest ‘Glowtime’ event, Apple Intelligence took center stage. But technologist Michael Spencer isn’t impressed. He argues that despite the AI buzz, Apple’s latest upgrades feel overdue and lack some key innovations. So, is Apple Intelligence a leap forward or playing catch-up?

Apple Intelligence can draft emails and texts, summarize emails, create custom emojis, and search for photos. There are improvements to Siri, too, which can now pull information from text messages and tell me when I’m meeting with my boss, but it still can’t tell me when I’m traveling next (even though I have the flight info in my calendar and email).

I agree with Spencer — Apple’s AI game still has some catching up to do.

While the iPhone 16 offers a glimpse of what’s possible, it’s not the breakthrough he hoped for. “The real AI revolution for Apple might not happen until 2025,” he says. Until then, we’re left wondering if these updates are enough to keep Apple in the AI race.

Read the full article at AI Supremacy.

Why You Need OpenAI’s New “Deep Thinking” Model

OpenAI’s latest models are here, and they’re doing more than just answering questions — they’re thinking.

The new models — o1-preview and o1-mini — are designed to think deeper and tackle more complex tasks. Whether you're working in science, coding, or math, these models are built to handle tougher problems than before.

One of the big improvements?

The models now take more time to think through your prompts. This extra effort helps them break down complex tasks, refine their approach, and catch their own mistakes.

In tests, the o1 performs up to PhD student level on benchmark tasks in physics, chemistry, and biology. As an example, in an exam for the International Mathematics Olympiad, ChatGPT-4 correctly solved 13% of problems, while o1-preview scored 83%.

ChatGPT-4 remains the go-to for most of my use cases, especially those involving multimedia inputs (text, images, videos) and tasks that rely on browsing the web, memory, and any of my uploaded files.

o1-preview is also slower and doesn’t have features like web browsing or file uploads (yet). Still, its ability to break down complex issues makes it particularly useful for strategic planning or when I want to analyze something complex (like an 85-page report on airline integrations).

If you’re interested in taking it for a spin, OpenAI also has a few prompting tips for the o1 models:

• Keep it simple and direct

• Don’t use chain of thought (it’s already built in)

• Use punctuations and tags

• Don’t overload it with information

Note that you need the pro version of ChatGPT, and you’re limited to 30 prompts per week, so pick your uses wisely.

Want to hear more about the new models? Kevin Pereira & Gavin Purcell from AI For Humans break it down in their latest video, and Ethan Mollick covers it on his blog, One Useful Thing. You can also go to the source at OpenAI.

Inside Uber’s AI-Powered Knowledge Revolution

Can AI streamline global support for millions of users? Uber is proving it can.

Hadley Ferguson, Director of Global Community Operations, and Peter Tu, Global Knowledge Production Lead at Uber, share how they’re creating a unified knowledge ecosystem by using generative AI to automate their support resources.

They’ve implemented custom AI apps using the Writer platform to help them manage 1 million+ saved replies and more than 50,000 Help Center articles, which helps them maintain consistency and accuracy across all support content.

By integrating Writer into tools like Google Sheets, where their teams already work, they’ve made it easier for teams to collaborate without introducing new platforms. The AI system is designed to account for regional legal and regulatory differences.

Their approach uses AI to identify duplicates and inconsistencies, automatically update resources, and tailor responses for user segments while giving their team members high-quality support content.

When getting started, Ferguson and Tu recommend embracing the learning curve and tackling small pilots first. Don’t overthink the first use case, and commit to action early.

Ferguson notes, "You can’t look too far down the road with generative AI—focus on what's achievable in the next six months."

Peter Tu emphasizes understanding the paradigm shift that AI brings and not underestimating the challenges of adoption.

You can read the blog at Writer.

How Copywriters Can Thrive in the Age of AI

According to copywriting expert, Scott Martin, there’s some good news and bad news for copywriters. 

The good news: AI is awful at creating direct-response copy that sells products and services.

The bad news: most people in marketing don’t understand direct response copy and think AI copy is great.

“I saw an article about a marketing manager who is being praised for his amazing approach to AI,” Martin says. “Of course, there's nothing about how much revenue his department is generating. It doesn't matter: he's using AI so all is well.”

Martin predicts that 80% of copywriters will vanish and 20% will prosper. If you want to be in the 20% then you need to do four things:

1. Your copy must be better than copy that AI produces.

2. You must know how to use AI tools intelligently.

3. You must know how to write direct response copy at a high level.

4. You must be versatile and able to write across every platform.

Now, if you’re wondering how to gain the skills to make that happen, Martin has an answer for you: The Aspen School of Copywriting.

And yes, he promises an all-new video series about using AI on the platform.

If you’re interested, check it out here.

BCG’s Framework for Implementing AI Successfully

Despite high expectations from business leaders, early AI adopters are struggling with ROI. 

To help, BCG launched the "Deploy, Reshape, Invent" framework - a guide for maximizing AI's value. The approach emphasizes starting with automating tasks, rethinking workflows, and finally, creating new customer experiences.

To implement AI for B2B success, follow these steps:

• Deploy: Automate routine marketing, sales, and service tasks using AI to improve efficiency.

• Reshape: Rethink and redesign workflows and processes with AI at their core to optimize operations.

• Invent: Use AI to create entirely new customer experiences and services that were previously impossible.

BCG says transforming B2B marketing, sales, and service with AI requires a phased approach, starting with task automation and evolving toward customer experiences. If you follow this approach you’re guaranteed to succeed.

For the full article, visit BCG's website.