🗞️ AI-Magic Inside Google Apps! 💥💪

Welcome to the 15th edition of the AI for E-Commerce newsletter, your trusted resource for actionable AI strategies in eCommerce, especially on Amazon! OMG, we are about to hit 1600 subscribers 💥🔥! Let’s go!

I am super excited about the latest AI enhancements that have come to Google Workspace. Meet Gemini AI, Google’s native AI platform. Once you turn it on, you can easily operate AI from the comfort of your home base.

Gemini AI can read ALL of your Google Workspace documents and create new documents based on information in your documents, sheets and slides. That’s next level.

When you open a Google Doc, you will Gemini right there. It can help you write based on any prompt.

It can also help you create a cover images, take meeting notes, draft an email, all while you are within your Google doc.

What’s more it can generate images right inside of slides, with no additional software!

Gemini AI can also help summarize what was said in a Google Meet, on the fly.

How much you ask? Gemini for workspace costs $20 per user per month. This plan provides access to Gemini in Google Workspace apps like Gmail, Docs, Sheets, and Slides and is the most comprehensive integration of AI into your daily workflow. I’m loving it!

OpenAI just gets better and better! There is a cool new model update to ChatGPT. It’s called “CPT-4o with canvas”. It brings next-level sophistication to both writing and coding. It’s quite interactive and has some nice bells and whistles.

Here are some of the cool things GPT-4o with Canvas can now do:

• Suggest edits just like a collaborator in a doc

• Adjust the length of the response on the fly, with a simple drag from Shortest to longest

• Change reading level on the fly, from Kindergarten to Graduate level with a simple drag

• Add final polish which enhances the formatting, titles and subtitles

• Add emojis on the fly (!!!) in all the appropriate places. Look how effortlessly it sprinkles emojis throughout your text.

If you’re not using the paid version of ChatGPT, consider it. I am saving so much time and gaining so much productivity with each new model that OpenAI comes out with.

💡 Do you know how much energy we consume each time we ask OpenAI a question? Here is a comparison of different models of popular AIs. Looks like GPT-4o Mini is the most efficient while GPT-4 was the worst, consuming 0.35 KWh of energy per query.

To put 0.35 kWh into perspective, here are a few comparisons with everyday activities:

1. Powering a 100-watt light bulb for 3.5 hours:

   • 💡A 100-watt light bulb consumes 0.1 kWh per hour. Thus, 0.35 kWh would keep it lit for 3.5 hours.

2. Running a microwave oven for 20 minutes:

   • Most microwave ovens use around 1 kWh per hour. So, 0.35 kWh would be roughly equivalent to using a microwave for 20 minutes.

3. Charging a laptop for around 10 hours:

   • Laptops typically use 30-40 watts, which means 0.035 to 0.04 kWh per hour. Using 0.35 kWh would power a laptop for 9 to 10 hours.

4. Making 17 cups of coffee with a coffee maker:

   • A typical coffee maker uses around 0.02 kWh per cup, so 0.35 kWh would make approximately 17 cups of coffee.

5. Running a refrigerator for 7 hours:

   • A refrigerator typically consumes about 0.05 kWh per hour. So, 0.35 kWh would keep it running for around 7 hours.

This makes me feel bad about my energy consumption, but I am hopeful that with each new model advancement these numbers will go down.

For the Thought leader section this week we asked our friend Matt Kostan from ProductPinion to share his highlights of the INBOUND conference:

"While Amazon had their Accelerate conference, INBOUND 2024 was happening in Boston with over 12,000(!) marketers in attendance. Sometimes it’s nice to step outside the Amazon bubble and get a fresh perspective.

At inbound, I had the chance to hear from amazing speakers like Dharmesh Shah, founder of HubSpot (obviously, it’s their event), Neil Patel, the godfather of marketing, Dane Vahey, Head of Strategic Marketing at OpenAI, and Jay Schwedelson, founder of Subjectline.com and host of the #1 marketing podcast in the U.S.

Here are some of the big takeaways from what was shared on stage:

From Neil Patel:

SEO has evolved. It’s not just “Search Engine Optimization” anymore; it’s Search Everywhere Optimization. For Amazon sellers, this hits home. We now need to optimize for RUFUS, COSMO, and AI searches like Perplexity and SearchGPT. I’ve been preaching for a while about how to get AI search to recommend your brand over others, and it all starts with being listed on high-profile product roundups. You can grab a SOP to get listed on the sites here.  

Neil also stressed that AI is a tool, not a replacement. Most AI-generated content still flops, with low engagement and shorter read times. Authentic content remains key, with engagement being the best metric. He emphasized that nothing drives that engagement quite like short-form videos, followed closely by long-form.

From Dharmesh Shah (HubSpot):

With AI advancing, Dharmesh mentioned that search volume is expected to drop by 25% by 2026. This is something Amazon sellers should pay attention to, especially since many buyers may start their research on Google and eventually land on Amazon.

HubSpot was able to increase their conversion rates by 82% using AI-driven personalization. This is pretty mind-blowing, considering HubSpot already teaches best practices in conversion rate optimization already. So for them to be able to increase conversions that much via personalising content, it’s staggering. This trend towards personalization is something Amazon's SERPs are moving toward too, with personalised search results being the future of e-commerce.

Dharmesh is also all in for AI Agents and plugged their new agent.ai site which allows users to hire AI agents to perform tasks. He also announced a beta where you could build an AI agent with a no-code interface.

From Dane Vahey (OpenAI):

Dane shared how most brands are underutilizing OpenAI, often limiting it to just content marketing. He pushed for more diverse use cases like web personalization and deeper analytics insights.

A big reveal was OpenAI’s latest model, “o1”, which can think and reason across multiple models like GPT-4, DALL-E, Whisper, and more. This unified approach is designed to let brands leverage AI across different tasks. 

From Jay Schwedelson (Subjectline.com):

Jay’s session was a highlight for me, particularly when he discussed the impact of Apple iOS 18 on email deliverability. With 47% of iPhone users on the Mail app, this is going to be huge. He warned that AI will soon categorize emails differently, and this won’t be like Gmail tabs. The main trick he suggested to overcome this was to make sure your newsletters come from a person’s name (not info@ or support@) to avoid being dumped in the promotions tab.

For Amazon sellers who are sending newsletters, switching to a personal name could lead to a 48% higher open rate compared to branded newsletters. Definitely something worth testing.

Love these insights? Follow me on LinkedIn where I’ll be sharing more (this week you’ll discover an all new AI tool that can almost instantly optimise your main image for more clicks)”