Hi everyone and welcome to NY Tech AI, where it is all about AI - AI Events, AI Insights, and the AI Ecosystem in NYC. If you know a friend or colleague, who might benefit too, please send it over. And with that, let’s begin.

What I cover today

📅 Upcoming AI Events in Town

Next Week:

The Week After:

After That:

💡 AI Highlights, Insights & Ideas

Nr. 1: AI allows you to go back to the basics

Tech tools are amazing as they improve workflows, save us time and make our work more effective. But within the current landscape, we see an overflow of tools popping up on each corner and industry every day.

My Linkedin stream is full of SDR AI Agents and the next big career tool to help me write my CV in such a way as to make my job applications more successful. It’s wonderful that we see this amount of innovation and development and I love it, as it allows us all to be more productive through the tools we use and the AI that democratizes our ability to bring those products to life.

And yet, many tools are simple copies of AI models like OpenAI, Anthropic and Google Gemini. A team connects an AI via an API, creates a beautifully looking front-end interface and sells it to a higher price for that specific use case.

For those tools, those simple tools, a bit of prompting may bring you to a much better solution than using one of the dozen tools out there on the market:

I think, you get the gist. 😉 In those 4 cases, Claude or to some degree another AI model provides you with the same if not better results than using any other tool that costs you at least the same $20 that you pay with a professional AI solution. You are faster. And your results are often better.

And now to the Smart Prompting: In my experience, smart prompting means “precise iteration” → Tell the AI what you need and how you need it.

Here is an example for the case of your Pitch Deck.

Nr. 2: AI creates your pitchdeck in 30 min

You need a pitch deck fast and most of the other tools take too long, have boring designs or are tools that put constraints on you that keep you from being as flexible as you would like to be?

Then here is the solution. Use Claude and create yourself some nice 16:9 landscape SVG files - perfectly designed and smoothly structured:

Step 1: Tell Claude what you need

The key here is to make that as specific as possible. Write down all the details that are important for your presentation. These can be placeholders for images, a time size of 3 specific types of slides, graphs etc.

Once you are done, review the resulting SVG files.

Step 2: Review & Make Improvements

It is unlikely that your first prompt leads to the perfect result. And likely also not the second. But over time, you will figure out what works best to use as prompts.

Here are some improvement prompts I often use with Claude while creating pitch deck, sales and QBR presentations:

Learning: Don’t expect AI to give you the perfect answer right away. Instead, let it create a first draft, and then start to iterate, 2x, 3x, and you are done.

Step 3: Create your final presentation

Once you are done, you may download your slides as SVG files. I personally prefer to quickly put them in Figma, make tiny adaptations in the design if needed, and then download them as PNG files or PDF. Online conversion tools or Canva work well.

One other reason, why I put them in Figma and the code in Notion is for automation purposes. The reason? See Nr. 3 below:

Nr. 3: First build your workflow manually, then automate

Connecting to people in the AI industry as with any other product in tech, I see that it is often better and more successful to understand what problem you are trying to solve before you are recreating an existing workflow before you automate that workflow before you use AI for your new workflow.

As AI is this new shiny tool and everyone loves to invest in AI and build a product with AI, we often forget to ask ourselves:

Answering these questions is extremely helpful to understand first, before jumping in and automating a workflow with tech and AI.

Becuase what happens when you skip that step? You may just create an amazing product that no one is interested in, as it does not solve a real problem that is urgent enough for people to invest in your product.

Thus, first figure out what you want to improve, why you want to improve it, and then use AI and tech to automate it. And ask yourself: does it really need AI for that?

Coming back to Nr. 2 - Step 3 - that is the second reason why I use Figma + Notion to save my slides. Figma for the designs & Notion for the SVG Code.

Once I have done the process several times, I can build my own automation out of it.

This worked not because I used AI right from the start.

But it worked, because I first used and tested all possible Pitch Deck apps on the market.

While I did that I consistently pushed my head against the ceiling of possibilities until I ended up with Claude as my best option so far.

Now, the next one will be to automate that workflow.

And with that, we are back at Nr. 1 of the AI Insights:

The more you follow your way of improving the workflows you are using in your day-to-day, the more you realize that there are ways to improve your own workflows with AI. Once you do, you can automate them, and once you do that, you may as well bring your own product to market.

Till that point that someone else comes, and finds a better and faster way, and then the cycle repeats.

With AI that just works much faster than with other tech products.

Because nowadays everyone has access to a tool as powerful and human-similar as AI.

And with that, let’s see what AI will bring us in 2025!

🧠 AI Innovation Insights

Research Spotlight: AI's Impact on Investment Decision Inequality

A large-scale experimental study examining how generative AI affects financial decision-making reveals critical findings. Research shows:

Key Takeaway: While AI makes financial information more accessible overall, it paradoxically widens rather than narrows the performance gap between sophisticated and novice investors. This challenges assumptions that AI democratizes financial markets. The research suggests AI's effectiveness depends heavily on matching its outputs to user expertise levels, and highlights how different user groups interact with and benefit from AI in fundamentally different ways.

✨ Miscellaneous Stuff

🚀 Community Highlights

🤝 Get Involved

Your NY Tech AI Team

Want to be featured? Have feedback? Reply to this email or reach out on LinkedIn. Let’s build it together! 🗽🚀💡