The Startup Ideas Podcast

Host: Greg Isenberg
Episode: Perplexity vs ChatGPT "Deep Research" For Startup Ideas Shocked Me
Date: February 19, 2025

Main Theme

Greg Isenberg undertakes a hands-on, side-by-side comparison of Perplexity’s and ChatGPT’s (Pro, $200/month) "Deep Research" features, using them as startup idea generators and research assistants. The goal: discover which tool delivers better, deeper, and faster actionable insights for building a high-value, defensible AI agent startup using the same prompt. Greg evaluates the platforms on speed, depth, creativity, tactical advice, and overall value—providing listeners with practical takeaways on leveraging AI tools for entrepreneurial ideation.

Key Discussion Points & Insights

1. The Experiment Setup ([00:00–05:15])

• Greg’s Rationale: “I want to test the differences between OpenAI ChatGPT's deep research and Perplexity's Deep Research. It's almost like a junior McKinsey analyst.” ([00:30])
• Prompt Used: Both tools receive a carefully crafted prompt for startup ideation:
  • Request: Give 5 defensible AI agent startup ideas targeting a high-value niche; must cost <$5k MVP; include a zero-to-one playbook with clear growth tactics aiming for $1–5M ARR over 3 years.
  • No industry or business model constraints; viral growth preferred, but open to other strategies.
• Greg’s Transparency: He highlights his limited experience with Perplexity’s advanced features—“we're kind of learning this together.” ([05:05])

2. Tool User Experience & Process ([05:15–13:00])

• Perplexity:
  • Immediately starts aggregating from 29–34 sources, including YouTube transcripts (even Greg’s own content).
  • Does not ask follow-up qualification questions like ChatGPT; Greg notes you may want to preemptively specify more constraints.
  • Speed: Finishes response faster than ChatGPT.
    • Notable Quote: “First things first is that Perplexity is faster than ChatGPT. You can see that ChatGPT is about halfway done. Perplexity is done.” ([10:38])
• ChatGPT:
  • Asks clarifying questions (e.g., industry, user type, technical resources, go-to-market).
  • Processing: Takes more time but generates a “full-on business plan.” ([13:45])
  • Greg observes ChatGPT Pro’s output is far longer and more detailed than Perplexity’s, potentially reducing the need for follow-up prompts.

3. Deep Dives: Perplexity & ChatGPT's Top Startup Ideas

Perplexity Deep Research: “Legal Contract Lifecycle Automation Agent” ([13:45–24:30])

• Problem: Corporate legal teams spend vast time on manual contract review, leading to errors and compliance risks.
• Defensible AI Solution:
  • Integrates with email/SharePoint/DocuSign
  • Learns org-specific risk thresholds
  • Auto-redlines contracts with company-specific fallback positions
  • Maintains version lineage (blockchain-inspired audit trails—as Greg notes, “Not really needed for MVP,” but interesting nonetheless)
• Why Defensible: “Develops institutional memory through continuous exposure… becomes embedded in workflows.” ([15:25])

Growth Playbook:

• Phase 1: MVP as a Chrome extension, recording contract review sessions. “Offer free redline audits for 100 AMLA 200 firms—targeting mid-market companies…” ([17:30])
• Phase 2: Integrate with legal ops tools, build a clause library, usage-based pricing ($0.25 per clause).
• Phase 3: Enterprise scaling (compliance add-ons, $8,000/mo tier).

Financials:

• Year 1: 40 x $2.5k/mo = $1.2M
• Year 2: 150 x $5k/mo + usage fees = $3.1M
• Year 3: 300+ instances = $5.4M

Greg: “This idea makes a lot of sense… I sip. This is a pretty good one.” ([19:15])

ChatGPT Deep Research: “AI Sales Outreach Assistant” ([24:30–36:45])

• Problem: Scaling personalized outreach without hiring more SDRs
• Defensible AI Solution:
  • Virtual SDR automates lead research and personalized messaging
  • Builds data network effects (“…from every interaction, it learns which messaging yields replies… building a unique data set.” ([26:30]))
  • Embeds in sales workflows, improving with use

Zero-to-One Playbook:

• Focus: Automate follow-up emails to warm leads; simple web app/extension.
• Low-cost MVP: “Use existing AI APIs and low-code tools… skip complex analytics at first…” ([28:15])
• Viral Growth: Product-led free tier, internal referrals (“Whenever the AI assistant books a meeting or gets a reply, prompt the user to share the success internally… Salespeople love sharing hacks at work.” ([30:15]))
• Monetization: Per-seat or usage-based pricing ($49–$199/mo tiers for individuals/teams).
• VC scale: Suggests the path toward $100M ARR would require fundraising and expanding the product suite.

Greg’s Analysis:

• “I feel like I’ve seen so many of these, but I’m open-minded… This is a meme on X, of course you’re building an AI virtual SDR, but automating follow-ups to warm leads is a cool idea.” ([25:20])

4. Comparative Analysis: Depth, Breadth, Follow-up Interactions ([36:45–50:00])

• Perplexity: Faster, concise, “manual” in the sense that you must often ask for deeper dives or request more detail. Great for quick exploratory research; necessarily more table-based, less verbose.
• ChatGPT Deep Research: “Literally a business mind”—incredibly comprehensive, deeper initial output, stitches together competitive analysis and actionable plans. “You’re going to one-shot it more often.”
• Pricing/Access: Perplexity is free/cheaper ($20/mo) and “a no-brainer for ideation.” ChatGPT Pro Deep Research ($200/mo) shines for more in-depth, end-to-end planning.
  • Notable Quote: “Is ChatGPT Pro worth 10 times the price of Perplexity? I’ll let you decide… depends on your workflow.” ([39:30])
• Greg’s Metaphor: “It’s like driving manual instead of automatic in your car. Some people might even prefer it.” ([43:00])

Handling Competition/Further Prompts

• Greg tests Perplexity’s analysis on legal AI competitors—gets a summary of 37 funded competitors, three under-defended niches, a competitive matrix, and an MVP plan still under $5k.
  • Memorable Moment: “They're going to give you $200 not to spend it at Chat GPT Pro. Maybe that's why they chose $4,800…” ([48:10])
• Both tools require thoughtful prompting (“You still need to know the right questions to ask… the onus is on the prompter.”)

Notable Quotes & Memorable Moments

• On Perplexity’s Aggregation:
  “Funny enough, actually one of the sources is me… That's hilarious. That's actually, it's kind of ridiculous, but it's cool that it goes through YouTube…” ([13:15])

• On MVP Realism:
  “Do we need blockchain-style audit trails? Not really. But I guess we can further prompt Perplexity to ask it: what would be the MVP for this, right?” ([15:20])

• On ChatGPT’s Depth:
  “Look how long this is, like a full-on business mind, literally.” ([24:50]) “I like that it’s bringing in real examples like Jenny AI and Cursor—these are the examples I want to see in here.” ([32:20])

• On Prompt Depth:
  “On both Perplexity and ChatGPT, you definitely need a… if you have question marks, you gotta ask again… You might have to prompt it five more times or seven more times…” ([33:45])

• On Workflow Integration:
  “Having a companion like Deep Research in your workflow is a no-brainer… It’s going to help you refine the types of products you build for the highest probability of success.” ([44:15])

Timestamps for Key Segments

• Experiment Setup — [00:00–05:15]
• Perplexity’s Approach and Results — [10:38–24:30]
• ChatGPT Deep Research Output — [24:30–36:45]
• Comparative Analysis & Meta Insights — [36:45–50:00]
• Follow-up Prompting / Handling Competition — [47:00–50:30]

Conclusion & Greg’s Final Takeaways

• Both Tools Are "Scary Good": “You can't go wrong with either product… It’s really what floats your boat.”
• Big Unfair Advantage: “Use it, play with it, come up with ideas. This is an unfair advantage right now.”
• Actionable for Entrepreneurs: Both tools democratize access to high-level research and business modeling; ideal companions for zero-to-one founders if you learn to prompt thoughtfully.
• Resource: Greg is providing all 10 startup idea/plan outputs in a doc—free for listeners.

Summary Table: Perplexity vs ChatGPT Deep Research

| Criteria | Perplexity Deep Research | ChatGPT Deep Research | |-------------------------|-------------------------------|------------------------------| | Speed | Faster | Slower | | Depth (per prompt) | Concise, needs follow-up | Much deeper “business plan” | | Follow-up interaction | Needs more prompting | Usually richer from the start| | Cost | Free/$20 mo | $200/mo | | Source transparency | Lists sources, links | Includes case studies, less granular about sources | | User type recommended | Tinkerers, “manual drivers”, fast explorations | Deep planners, want end-to-end blueprints | | Example MVP | Legal Contract AI Agent | AI Sales Outreach Assistant |

Final Words

“If you aren’t using one of these platforms, you probably should. This is an unfair advantage right now. That’s why I wanted to show it to you all.” ([53:00])

Listen, experiment, and happy building!