Just a marketer trying to master AI in 2024 🦾 | Host of AiDrivenMarketer.com

Marketers everywhere are being told “learn AI.” The question is: learn it for what?

The real opportunity isn’t learning every tool... it’s understanding how AI changes how content is created, distributed, and consumed.

I’ve spent the last few years helping marketing teams adopt AI—from podcast workflows to full content systems and teaching thousands through The AI-Driven Marketer.

I help marketers turn AI from chaos into clarity by automating the busy work, amplifying their creativity, and clarifying their message.

👉 Subscribe to The AI-Driven Marketer newsletter for hype-free lessons on using AI in your content strategy at https://AIDrivenMarketer.com/

👉 Or start with my free 5-Day AI Fundamentals course at https://danchez.com/course/

Dan Sanchez

Too many people overemphasize AI automation.

Most of my AI work is still manual.

Automation is great and I love setting up automations where I can.

When the process is right, when it's repeatable, when there's clear trigger points.

But most of our work is figuring out new things, solving unsolved problems, and launching new projects. 

You can't automate that. 

At least.. not yet.

That's why I still spend most of my AI time between ChatGPT/Gemini/Claude. They are coworkers and thinking partners helping me push things forward.

I'm also always hunting for things that are daily, weekly, or monthly tasks I can automate but, at a marketing leadership level, those things are few. 

#Danchez

You Can Be AI-Driven AND Human-first. Join me & 5k others to leverage both at AIDrivenMarketer.com

Honestly, all the free AI apps are so good right now.

There's no reason why you shouldn't be using them.

ChatGPT, Gemini, Claude. 

Just download and create accounts with all of them.

Use all of them.
They are all fantastic. 
I'm flipping back and forth often now.

You?

An AI strategist, operator and builder, who is passionate about helping businesses realize the tangible benefits of artificial intelligence. He has led AI adoption from enterprises to start-ups through education, strategy and implementation. 

With a background and expertise in building businesses, demand generation, and keen ability to drive AI implementation company-wide, Karl has led initiatives resulting in substantial ARR growth and operational efficiencies.

Karl Yeh

JUST A REMINDER: Anyone posting a variation of "I tested XXX AI tool so you don't have to" or gives their take on a AI tool ALWAYS test it yourself. 

Your use cases, your business use cases, under your processes with your data, in your context, with your benchmarks. 

There are too many variables when people are testing or providing their own conclusions, including ones that I provide. 

From personalization to memory to prompting styles to examples, whether the model was grounded or not grounded, how they measured it and on and on.

Don't get me wrong, there's a lot of great posts and takes.  Using their testing methods may get similar outcomes, or you may not. 

And they also provide some great examples on how to test, but again make sure to test it yourself in your context because you've probably built out a different way of working and your AI tool probably has a history with different personalization.

And how you interact with one model prompt for one model has different outcomes on how you prompt for other models as well. 

And our advice to many businesses is to never get siloed into one tool. Always be open, always be testing if you can as it will never be a one model to rule them all, especially for different tasks, there are definitely different tools that can provide the output that you're looking for better than the others.

Helping business use the full benefits of artificial intelligence | AI strategist and operator | Co-founder 0260.AI

Here's an AI business implementation case study involving presentation decks with a one-day turn around I wanted to share. 

The client already created a couple of decks for stakeholder presentations but they weren’t aligned with their brand guidelines.

The main one was a 78 slide quarterly performance deck and the other was for partners. 

The task was:
1. Review the company brand guidelines and a pre-built deck template
2. Rebuild the deck inline with the brand guidelines. Use the pre-built template but make adjustments to it.
3. Make sure the numbers are consistent with the original deck
4. Leverage AI to turn some of the numbers into infographics

So I thought I could use multiple tools to test this out:
- Claude in Powerpoint
- Genspark
- Regular Claude
- ChatGPT pre-built Claude skill based on the brand guidelines
- Microsoft Co-Pilot (I'm just joking. There's no way I would use this for this type of task)

Didn't use Gemini because client is a Microsoft shop and had to be a easily replicable process. Just added effort converting Slides to PP. 


The results:

CLAUDE FOR POWERPOINT - FAIL
I uploaded the brand guidelines and  pre-built template in Claude for PowerPoint and tried to have it rebuild the deck. However, I found it very lacking.  It’s better at editing an existing deck or building something new.

individual slide edits are eat tokens, and outputs weren't looking for no matter the model used. 

For example, client wanted dark backgrounds converted to light backgrounds but the fonts and infographics needed to be inversed to make it readable….Claude in Powerpoint had a hard time with that. 

Also there's was a lot of context compacting and if you're not on that max plan you will get rate limited. I had to restart the task a lot.

GENSPARK - WIN
Genspark was the best tool for this use case. It was able to review the brand guidelines and apply the original template the best.

It had no problem with the light and dark change and intuitively inversed the fonts and graphics. 

The ability to edit individual slides both with AI and manually just put it over the top. Once the deck was approved, it became the new template and we rebuilt another deck that was on point.

one downside is rebuilding a 70-page deck will eat your credits.  I had to buy some more

CLAUDE - FAIL
Using Claude was able to build a little bit better deck but not much. Trying to edit it slide by slide was difficult the same problem  fonts and graphics.

CHATGPT - FAIL
 I knew going in ChatGPT didn't have the design capabilities as Claude or Genspark, but to try to make it on par, I broke down the brand guidelines, turned it into a skill, and uploaded that skill as part of the ChatGPT thread. 

 it took a long time editing, and the output was okay, but it did recognize and inverse the fonts correctly. Just wasn't as good as Genspark. 

One thing I took away from this exercise was how real business processes benefits from both AI capability and human judgement.

taplio

The Battle

Dan Sanchez

Karl Yeh

Followers

Posts

Engagements

Likes

Performance ScoreDan Sanchez's LinkedIn Performance

Performance ScoreKarl Yeh's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Dan Sanchez

Karl Yeh

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Dan Sanchez's LinkedIn Performance

Performance ScoreKarl Yeh's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreDan Sanchez's LinkedIn Performance