AI Summer School - Deep Research

Your guide to getting the most out of generative AI tools

Author

Innovation Profs
July 19, 2025

Welcome to Gen AI Summer School

We’re spending the summer teaching you the essentials you need to succeed in an AI-forward world.

Here’s the plan:

Introduction to Deep Research

“How can I increase the reach of my biweekly newsletter on generative AI?”

When I passed this prompt to ChatGPT (using the model 4o), I received a total of seven helpful suggestions we could employ to expand the reach of this newsletter, with 2-3 bullet points elaborating upon each suggestion. But what if we want to go deeper, with a more nuanced discussion informed by authoritative references on the subject?

Cue ChatGPT’s Deep Research tool, which, according to OpenAI’s introduction to the tool, is an “agentic capability that conducts multi-step research on the internet for complex tasks.”

Taking Deep Research for a Test Drive

Let’s see if Deep Research can do any better than the response given by 4o. First, Deep Research can be found in the “Tools” dropdown:

Upon selecting “Deep research,” I’m told that I have 25 uses of the tool over the next month (this with a paid ChatGPT Plus account; unpaid users can use the tool five times per month).

Let’s give it a whirl with the same prompt. The immediate response was a series of follow-up questions, each of which I answered.

Upon receiving my answers (including a link to past issues of the newsletter), Deep Research got to work, performing web searches and consulting various sources, with updates given to me over time:

Finally, after six minutes, Deep Research concluded its work.

The Fruits of Deep Research’s Labors

What I received was a report of roughly 10,000 words in length (almost 50 times longer than what 4o originally gave me in response to my initial prompt!), which was highly informed by five Beehiiv blogposts I had not previously encountered (Beehiv hosts our newsletter) that were referenced nearly seventy times throughout the response. In addition, all of the external references in the document were hyperlinked to the original sources.

Of the two Innovation Profs, Snider is the one with the experience in digital media strategies, so I (Porter) learned quite a lot from the report about SEO best practices, lead magnets, and lead nurturing campaigns, as well as Beehiiv’s own growth tools. As I typically rely heavily on Snider for his digital media expertise in our partnership, the report I obtained not only has summarized a massive amount of information for me, but it has given me many leads to follow. My expertise isn’t being replaced here, as I didn’t have much in this space to begin with. Instead, I now have a point of entry into a relatively new area for me, which might allow me to start building a semblance of expertise on the topic.

Takeaways

ChatGPT’s Deep Research tool offers far more than a longer answer—it offers a smarter, more targeted one. While the 4o model provided a solid overview, Deep Research unearthed highly specific, actionable strategies tailored to our newsletter platform and goals, including insights from sources we hadn’t previously seen. For someone like me, with limited background in digital growth tactics, the tool didn’t just return information—it helped me start learning the terrain. It complements existing expertise, fills in knowledge gaps, and does so in a way that invites follow-up and further exploration.

One last note: Do other LLM platforms offer similar tools? Yes, indeed they do. Google’s Gemini has its own Deep Research tool, Claude has a Research tool in beta mode for paid users, and Microsoft 365 Copilot features two research-oriented agents, Analyst and Researcher. Give these tools a shot and see what you can learn!