Podcast Summary

Overview

Episode Title: ChatGPT’s new Deep Research Update: 5 Ways You Can Use it Today
Podcast: Everyday AI Podcast – An AI and ChatGPT Podcast
Host: Jordan Wilson
Date: February 18, 2026

In this episode, Jordan Wilson explores OpenAI’s latest Deep Research update for ChatGPT, highlighting the major new features, model improvements, and five actionable ways listeners can leverage Deep Research right now to transform personal and work-related research tasks. The episode dives into why this tool is moving from “just a mode” to a full research platform, offers a user’s guide with live demonstrations, discusses real-world use cases, and provides insider tips to maximize value.

Main Discussion Points and Insights

1. What’s New in ChatGPT’s Deep Research (02:47 – 13:21)

• Overview of Improvements

  • The update is live for Plus and Pro users; limited Free availability expected late February 2026.
  • Major visual overhaul: full-screen, split-view research reports with a left-side table of contents and a right-side citation/source tracking panel.
  • Real-time document upload: Upload files at the start or during research, making it more dynamic (05:45).
  • Live steering: Pause or redirect the research agent logic in real-time without restarting.
  • New ability to specify target websites (“the ability to choose which websites it does go to, which is big” – Jordan, 09:52).
  • Powered by GPT-5.2, OpenAI’s latest model in ChatGPT. Host speculates variant used is the “thinking” version, not the Pro flavor.

• Comparing Plans' Limits

  • Pro ($200/mo): 125 “full model” and 125 lightweight queries per month.
  • Plus/Teams: 10 “full” + 15 lightweight per 30 days.
  • Free users will get some access soon.

• Why It’s Now a Platform

  • Interruptible, steerable in mid-research – a “game changing option.”
  • Enhanced reasoning and synthesis across multiple sources.
  • Improved UI/UX for digesting research outputs and verifying sources.

Notable Quote

“The ability to pause, redirect and interrupt this model in the middle is such a huge game changing option or feature.”
— Jordan Wilson (09:08)

2. Deep Research User Experience and Features Demo (13:21 – 25:32)

• Switching Between Research Models

  • Select between “Deep Research (new)” and “Legacy (old O3/O4 models)” in the interface.
  • Legacy asks 3–5 clarifying questions before starting, which can yield better first results.
  • Suggests “context stacking” as a prompt engineering tip for higher quality output.

• Connecting Data Sources

  • Apps (formerly “connectors”): E.g., Canva, Salesforce, HubSpot, ClickUp, Google Drive, etc. — more than 60 options.
  • Website targeting: Choose to prioritize or restrict research to particular sites.
    “You can require Deep Research to go to your site first or a series of sites that you trust... but you don’t have to limit it, but you can if you want.” (21:19)

• Live Research Demo

  • Jordan demonstrates a query drawing from 720 Canva decks, his own website, and web sources to analyze six months of podcast content and trends.

• Real-Time Monitoring

  • You can watch Deep Research’s stepwise logic, sources considered, and active thought process.
  • Essential to monitor for connection issues with apps or stale data.

Notable Quote

“If you’re not using a deep research tool daily and connecting your data on the front end, you are absolutely missing out... It is the best way to consume synthesized, well-sourced information at scale that is personalized for you, your use case, your business’s viewpoint, etc.”
— Jordan Wilson (11:23)

3. Five Key Use Cases for Deep Research (25:32 – 40:15)

1. Memory-Powered Planning for Next Steps (27:10)

• Start with open-ended prompts that use ChatGPT’s memory and your data.
• Let the model plan your next six months or guide priorities, then follow up with extra context or data for refinement.

“One of the biggest mistakes people make with extremely powerful large language models is we think that we know the right answer. I always say: start wide, work your way to narrow.”
– Jordan Wilson (28:25)

2. RAG Company Search (Retrieval Augmented Generation) (30:21)

• Restrict Deep Research to internal company data (Google Drive, website, etc.) for focused, accurate reports.
• Provides more reliable, cited results versus generic web-wide research.

3. Competitor Deep Dive Using Company Context (37:10)

• Analyze competitors by combining your own business data and memory with targeted competitor/industry sites.
• Yields highly tailored reports as if hiring a consultant: personalized, actionable, and strategic insights.

4. Industry SWOT Analysis Built from Your Data & News (38:49)

• Run a Strength-Weakness-Opportunity-Threat analysis by merging company data and up-to-date industry/news sources.
• Use advanced URL/filtering (“Boolean URLs”) for customized, constantly-updated streams.
• Opportunity for power-users to automate sector trend reporting.

5. Follow-up Assistant for Inbox and Calendar (39:35)

• Deep Research can comb through Gmail and Google Calendar (read-only), surfacing missed opportunities and follow-ups.
• Jordan describes using markdown files about roles and company context to enhance the assistant’s relevance.

“There’s no reason for me to suck at email. It’s just more or less overwhelming, right? When I run these, it’s like here’s 85 extremely important emails you haven’t got to, right?”
– Jordan Wilson (40:08)

4. Important Observations & Pro Tips

• Read-Only Limitation
  “Deep Research, this only has read access, okay? So it’s not going to perform actions for you… not going to delete files or change statuses.” (24:49)

• Prompt Engineering
  Always “context stack” in your prompts and run Deep Research two or three times for best results (15:07).

• Source Verification Made Easy
  Full screen view with table of contents and quick-access source panel for fact-checking.

• Wishlist: Excluding/Blacklisting Sites
  Current update only allows including sites; blacklisting would further filter sloppier/hallucination-prone sources (34:41).

• Not Just for AI Power Users
  The new interface and model improvements make Deep Research accessible enough for business leaders and every-day knowledge workers to reap massive efficiency gains.

Memorable Quotes

• “It is literally sometimes—I’ve worked with consultancies in the past, right—some Deep Research queries, if you give them enough information, enough context… it is like you’ve hired a consulting company if you do it right.” (11:45)
• “Imagine doing this for your company’s website, or a competitor’s website, or 10 industry websites... This is such a good way to do simple sentiment analysis and catch up quickly.” (36:19)
• “These are things that I do all the time. And then, like, when I do them, I’m like holy frick, that’s amazing. I need to tell people, right? So you should be doing this.” (37:19)

Important Timestamps

| Timestamp | Segment/Topic | |:-------------:|:----------------------------------------------------------------------------| | 02:47–13:21 | Deep Research update overview, model improvements, experience changes | | 13:21–25:32 | Demo: Running Deep Research, using new features, monitoring chain of thought | | 25:32–27:10 | Transition to use case discussion | | 27:10–30:21 | Use Case 1: Memory-powered planning | | 30:21–37:10 | Use Cases 2–3: Company search, competitor mapping | | 38:49–39:35 | Use Case 4: Industry SWOT from your data/news | | 39:35–40:15 | Use Case 5: Follow-up assistant for inbox/calendar |

Conclusion

OpenAI’s Deep Research update in ChatGPT (GPT-5.2) marks a pivotal change—transforming it from a chatbot research feature into a robust research platform. Jordan Wilson’s walkthrough demonstrates not just new UI/UX perks and model improvements, but also strategic, high-value workflows, from company knowledge mining to industry analysis and personal task management. Regular users and power users alike can leverage these new tools for actionable insights, reclaiming hours of manual research with better quality results.

Tip to Listeners:
Try out the platform’s new features and connect data sources relevant to you. For practical takeaways and a trend report demonstration, Jordan invites listeners to repost the episode on LinkedIn for a sample output.

For all key takeaways and AI news, sign up for the Everyday AI newsletter.