A Note Before We Begin: I wrote this because I genuinely admire what Aravind Srinivas and the Perplexity team are building. I’ve been following their work closely, reading interviews, watching social media content, and tracking their product evolution. The technical execution is remarkable. The team is world-class. The growth trajectory speaks for itself.

But I’m fascinated by a different question: What makes a business truly connect to identity? Not just what it does, but what it means to the people who choose it. This is what I love exploring — the invisible dynamics between positioning, identity, and gravitational pull that explain why certain companies succeed in ways their founders can’t quite articulate.

I’ve tried my best to get the facts right, drawing from public statements, user reviews, media coverage, and the company’s website, socials, etc. If I’ve misrepresented anything, that’s entirely unintentional. My goal isn’t critique, it’s revelation. Aravind, I’m not claiming to know your business better than you do. You have data, customer conversations, and lived experience I’ll never have.

What I’m attempting is something you literally cannot do yourself: read your own label. No one can. We’re all inside our own stories, explaining our success through the frameworks that make sense to us, using the language that fits our identity. That’s not a weakness. It’s human.

This entire piece asks only one question: What business are you really in?

Not what you sell. Not what you build. Not even what you say you do. But what concept you own in the minds of the people who can’t imagine going back to life before they found you.

The answer might surprise you. It surprised me.

PS: The CEO Clarity Starter Kit uncovered all the insights you’ll read in this perspective.

Part 1: The Story They Tell

Aravind Srinivas has an elegant answer to why Perplexity works. Ask him about his company, and he won’t talk about search algorithms or AI models. He’ll tell you about human emotion.

“Every company should stand for a core human emotion,” he says. “Ours is curiosity.”

This isn’t marketing fluff. It’s genuine philosophy driving product decisions. The tagline, “Where knowledge begins,” reinforces it. The mission frames democratizing access to knowledge as more valuable than wealth accumulation, a principle Srinivas absorbed growing up in Chennai, where intellectual pursuit trumped material gain. “You can probably focus on wealth and your net worth, but at some point, it taps out. On the other hand, there is no end to knowledge.”

When Perplexity introduced Auto Mode to simplify their increasingly complex interface, Srinivas framed it as helping people “ask better questions.” When they launched Deep Research, an agentic mode conducting multi-step analysis autonomously, he framed it as removing friction between curiosity and answers. The Comet browser? More curiosity infrastructure, embedding AI capabilities directly into web navigation.

The competitive narrative flows from this framing. Google corrupted search with advertising incentives. Link-based results force users to sift through clutter. Traditional search engines became librarians pointing to books rather than reading them. Perplexity solves this by being an “answer engine,” a fundamentally different category delivering direct, cited responses. Citations aren’t just features; they’re a transparency serving intellectual rigour.

Srinivas credits his team’s elite AI pedigree (OpenAI, DeepMind, Google Brain) and their proprietary RAG (Retrieval-Augmented Generation) architecture. The technical moat, he explains, lies in sophisticated retrieval: completeness, freshness, speed, and fine-grained content understanding. They’re model-agnostic, leveraging third-party LLMs while investing heavily in the infrastructure feeding those models. It’s an AI-native approach to conversational search.

The competitive strategy follows naturally. Against Google: ad-free, direct-answer focused. Against ChatGPT: real-time, citation-mandatory. Growth comes from curious users discovering a better way to learn. 10 million to 15 million active users in 2024, powered by “curiosity-driven engagement loops.”

He references Sun Tzu: “Make the enemy’s weakness your strength.” Google’s weakness? Any answer reducing link clicks threatens their revenue model. Perplexity’s entire architecture exploits this structural vulnerability.

The metrics reinforce the narrative: 92% query accuracy, 5-7 second response times, retention prioritized over volume. The $18-20 billion valuation with only 52 employees? Proof that a focused, technically elite team can challenge incumbents by redefining categories.

This story is coherent, inspirational, and intellectually defensible. It’s also missing what’s actually happening.

Part 2: The Hidden Position

Here’s what users actually say when describing Perplexity:

“Verified answers.”
“Reliable sources.”
“I can trust it.”
“Finally, cited information.”

Notice what’s absent? Curiosity.

One G2 review captures the pattern: “I appreciate Perplexity for its reliable and verified answers.” Another: “Perplexity is a research powerhouse when you know how to prompt it.” The word appearing constantly isn’t “curious” or “exploration” or “learning.” It’s trust.

The noun Perplexity actually owns isn’t curiosity, it’s reliability. More precisely: Research Confidence. They’ve created a perceptual monopoly around the idea that AI-generated answers can be trusted if they’re properly cited and grounded in real-time sources. That’s the mental territory they occupy. That’s what customers buy when they subscribe to Pro at $20/month.

Watch how competitors can’t touch this territory. Google’s AI Overviews? Users complain about hallucinations and incorrectly synthesized information. ChatGPT? Brilliant for creative work, terrible for factual accuracy without constant verification. Claude? Safe and thorough, but no real-time web access without workarounds. Perplexity stands alone in the space of “I need a correct answer right now with sources I can verify.”

The positioning isn’t “Ask me anything because you’re curious.” It’s “Ask me anything because you can trust the answer won’t be made up.”

This creates a fascinating disconnect. Srinivas talks about Deep Research as enabling curiosity at scale, “conducting dozens of searches, reading hundreds of sources, reasoning through material.” But users describe it differently: “saves hours of research,” “generates confident business intelligence,” “provides credible analysis for high-stakes decisions.”

The job-to-be-done isn’t satisfying curiosity. It’s producing reliable work product without the anxiety of being wrong.

Look at the B2B layer where this becomes crystal clear. Enterprise tier pricing ($40-325/seat/month) isn’t sold on curiosity. It’s sold on audit logs, data privacy, internal knowledge search, and collaboration features. These are trust infrastructure, not curiosity tools. The Crunchbase and FactSet integrations? Financial professionals don’t use those because they’re curious about market data. They use them because their careers depend on getting valuations and competitive intelligence right.

The position they actually own (Research Confidence) chose their distribution path. You can’t charge for curiosity (it’s abstract and universal), but you can absolutely charge for reliability (it’s concrete and valuable). The pricing ladder makes sense only through the reliability lens:

• Free tier: enough to verify citations are real and answers are grounded
• Pro ($20/mo): unlimited, reliable research for knowledge workers
• Enterprise ($40-325/seat): reliability infrastructure for professional teams

This explains why the copyright lawsuits matter existentially. If publishers successfully argue that Perplexity represents “sophisticated plagiarism” or “unauthorized copying,” the trust position collapses. Users tolerate AI hallucinations from ChatGPT because they know it’s a creative tool. They will not tolerate unreliability from Perplexity because reliability is the entire promise.

Wired’s investigation claiming Perplexity “closely paraphrases articles with minimal attribution” and Cloudflare’s CEO comparing their crawling methods to “North Korean hackers,” these aren’t just legal problems. They’re existential threats to the position.

Here’s the invisible dynamic Srinivas may not fully see: his philosophical commitment to the infinite value of knowledge created the conditions for owning trust, but he built a trust engine while believing he was building a curiosity machine.

Part 3: The Identity Layer

Who is a Perplexity user? Not demographically, but identity.

They are people who identify as rigorous thinkers. They’re the person in meetings who says “Let me verify that” before accepting a claim. They’re the analyst who won’t present findings without sources. They’re the product manager who needs to understand the competitive landscape before making recommendations. They’re the investor who must differentiate signal from noise in market research.

Perplexity doesn’t sell to “curious people.” Curious people use Wikipedia, podcasts, YouTube rabbit holes. Perplexity sells to people whose identity requires being correct. The professional class whose self-image revolves around intellectual credibility.

This creates powerful identity resonance. When someone chooses Perplexity over Google, they’re not just choosing better results. They’re expressing: “I am the kind of person who verifies sources.” When they choose it over ChatGPT, they’re saying: “I am the kind of person who won’t risk hallucinations in professional contexts.” When they pay for Pro, they’re declaring: “My intellectual standards justify premium tools.”

The B2B identity layer is even richer. When a company buys Perplexity Enterprise, what are they really buying? Yes, they get Internal Knowledge Search, audit logs, and collaboration features. But what they’re actually purchasing is identity protection for their entire organization. They’re buying insurance against the professional embarrassment of AI-generated misinformation. They’re buying permission for employees to use AI without putting the company’s reputation at risk.

This explains why the Publisher Program (revenue-sharing with TIME, Fortune, Der Spiegel, 200+ outlets) matters strategically beyond resolving copyright disputes. It’s about maintaining identity alignment between Perplexity’s position and its users’ needs. If Perplexity is the tool for rigorous thinkers, it must maintain relationships with rigorous sources. Users need to believe the synthesis they’re getting comes from places they’d cite themselves.

Now consider Srinivas’s identity and how it unconsciously shaped the position.

He’s a UC Berkeley PhD who worked at OpenAI, Google Brain, and DeepMind. He contributed to foundational AI research, the Perceiver architecture and FNet. His co-founders are similarly credentialed: Denis Yarats (NYU PhD, Meta AI), Andy Konwinski (Databricks co-founder, Apache Spark creator). This is not a team that would build unreliable technology. Their identity as elite AI researchers made citations and source provenance non-negotiable from the start.

But here’s what’s invisible to them: they think citations are about intellectual honesty (curiosity), when users experience citations as risk mitigation (trust). Srinivas genuinely believes that showing sources helps people learn and explore. Users actually experience it as: “Thank god I can verify this before I stake my professional reputation on it.”

The founder’s philosophical framing (knowledge as infinite value) provided moral high ground for challenging Google. But the actual market resonance came from something more primal: professionals drowning in the misinformation crisis who needed a lifeline they could grab with confidence. Srinivas built that lifeline while talking about curiosity.

This also explains the copyright crisis’s psychological complexity. From Srinivas’s perspective, he’s democratizing knowledge access, making elite research capabilities available to everyone. Facts are free; he provides citations; publishers should be grateful for traffic. This genuinely feels consistent with the curiosity mission.

But from publishers’ perspectives (and increasingly users’ perspectives), the behaviour looks different: unauthorized use of copyrighted content, sometimes “virtually identical to original sources,” with spoofed user-agent strings to bypass robots.txt restrictions. That’s not curiosity infrastructure. That’s ethically questionable behaviour, contradicting the trust position.

The identity conflict is profound: Srinivas experiences himself as a knowledge democratizer, while external parties increasingly perceive Perplexity as a content appropriator. And because the actual position owned is trust (not curiosity), the perception problem threatens the business in ways the curiosity-framing doesn’t help him see.

Part 4: The Success Mechanics

What’s actually working at Perplexity isn’t what Srinivas thinks.

He believes success comes from superior RAG architecture, citation transparency, category creation around “answer engines,” and avoiding Google’s advertising conflicts. These are all true and important. But they’re symptoms of something deeper.

The real success mechanic: Perplexity’s position arrived at exactly the moment when the misinformation crisis made “reliable AI answers” worth paying for.

Consider the timeline. Perplexity launched in 2022, right as ChatGPT was demonstrating both the promise and terror of generative AI. Users discovered that AI could produce convincing-but-wrong information at scale. “Hallucination” entered the common vocabulary. Initial excitement about AI assistants gave way to anxiety: “This is amazing, but how do I know it’s not making things up?”

Perplexity’s entire architecture answers that anxiety. Not philosophically, but practically. Every answer has inline citations. Every claim links to verifiable sources. The system doesn’t just tell you something; it shows you where it learned it. In an environment where trust in AI-generated information was collapsing, Perplexity offered the only AI tool that let you trust-but-verify in real-time.

This explains why freemium works when many AI tools struggle with monetization. Users pay $20/month because the alternative (using unreliable AI or doing manual research) carries real professional costs. Eight hours of research time saved per week (reported by API partners) isn’t about satisfying curiosity. It’s about reducing the anxiety of being wrong while maintaining productivity.

Enterprise adoption follows the same pattern. Companies aren’t buying Perplexity so employees can be curious. They’re buying it because the alternative is either banning AI (losing productivity) or allowing unrestricted AI use (assuming reputation risk). Perplexity Enterprise splits the difference: you get AI productivity with attribution infrastructure protecting the organization.

What else is working:

The position chose distribution, not vice versa. Srinivas thinks they chose freemium to democratize access. Actually, freemium was inevitable once you own “research confidence.” You can’t charge for curiosity (too abstract), but you can charge a premium for reliability (concrete value). The free tier proves the citations are real. The Pro tier is for people whose professional identity requires unlimited, reliable research. The Enterprise tier is for organizations that are institutionalizing risk mitigation.

The technical moat reinforces the position, but the position came first.

The sophisticated RAG infrastructure (hybrid retrieval, fine-grained content understanding, real-time freshness) is genuinely impressive. But these technical choices were predetermined by the position. If you’re going to own “research confidence,” your retrieval system must be complete (comprehensive answers), fresh (real-time reliability), fast (professional workflow compatible), and granular (precise attribution). Technical excellence flows from position requirements, not the reverse.

Competitive isolation is perceptual, not technical.

Google could build citation-heavy AI Overviews. ChatGPT could add mandatory web searches. OpenAI’s SearchGPT is literally attempting this. Yet Perplexity maintains its position because it owns the mental territory of “trusted AI research,” and competitors fight uphill against perception. When users see Google AI Overviews, they think “ads and SEO manipulation.” When they see ChatGPT web browsing, they think “creative tool, not research tool.” Perplexity = reliable research is already established.

What they’re missing:

The curiosity framing is diluting, not strengthening, the position.

Every time Srinivas talks about curiosity, he adds cognitive load. Users must translate: “He says curiosity, but I know he means reliability.” This translation tax makes the position less sharp. If he leaned fully into “Research Confidence” or “Knowledge You Can Trust,” the position would strengthen because language and reality would align.

The copyright crisis is a feature, not a bug.

Publishers only sue threats, not irrelevancies. The lawsuits from BBC, NYT, Dow Jones, and Condé Nast prove Perplexity has achieved position strength. They wouldn’t bother suing if Perplexity didn’t meaningfully compete with publisher traffic. However, and this is crucial, the lawsuits expose that the trust position is fragile because it depends on publisher relationships. The spoofed user-agent allegations directly contradict the trust narrative.

Position evolution is inevitable, and they’re not steering it intentionally.

The expansion into Comet browser, Shopping Hub, and Labs automation is a logical extension of “answer engine” but a potentially confusing extension of “research confidence.” If Perplexity owns trust in the research domain, does that trust transfer to e-commerce recommendations? To automated workflow execution? Maybe. But without consciously managing the position, they risk becoming “an AI tool that does lots of things” instead of “the reliable research AI.”

Testing the Position (What Product Extensions Reveal)

The real test of positioning isn’t what you say, it’s what you can extend into without breaking the mental model users have of you. Perplexity’s product portfolio beyond core search is revealing something critical about the position they actually own.

The Email Assistant: Unexpected Position Proof

In October 2024, Aravind tweeted something striking: “We haven’t seen a product that has had this level of onboarding to retaining users conversion yet.” He was talking about the Email Assistant.

A user responded: “Perplexity Email Assistant is pretty sick. This week was busy, and I had missed replying to many emails. While going through them now, I saw the draft it created for one email. All the info, details, and numbers were perfect, exactly how I normally reply.”

That phrase, “exactly how I normally reply,” is the positioning tell.

The Email Assistant isn’t helping you be curious about email. It’s not teaching you to write better. It’s doing something more profound: it’s producing professional output that protects your identity. Users trust it because it sounds like them, includes accurate details, and maintains their professional voice. They’re willing to send emails drafted by AI because Perplexity has earned trust in a different domain (research) that transfers to adjacent professional tasks.

This reveals the position might be broader than “research confidence.” It might be: professional output you can trust with your reputation.

Think about what’s happening psychologically. When you use ChatGPT to draft an email, you heavily edit it because you don’t trust it to represent you accurately. When you use Perplexity’s Email Assistant, you trust it enough to send with minimal edits. That’s not about curiosity. That’s about delegation without anxiety.

The conversion metrics Aravind mentioned make sense through this lens. If the position is “trusted professional output,” then email drafting is a natural extension. It’s another high-stakes professional task where being wrong carries a reputation cost.

The Comet Browser: Research Infrastructure or Position Dilution?

The Comet browser is more complex. On the surface, it could be seen as category expansion, moving from “answer engine” to “AI browser” puts Perplexity in competition with Arc, Brave, and eventually Chrome/Safari with integrated AI.

But watch what actual users do with it.

They’re not using it for general browsing. They’re using it as a research workflow infrastructure. The AI capabilities aren’t novelty features. They’re professional tools. When you’re researching a topic, you can query Perplexity without leaving your browser context. When you’re reading an article, you can ask questions about it in-line. When you’re comparing sources, you can synthesize without switching tabs.

The Max plan pricing ($200/month) tells you who this is for: professionals whose time is valuable enough that a seamless research workflow justifies premium pricing. This isn’t consumer browsing. This is professional infrastructure.

If Comet stays disciplined around “research confidence infrastructure,” it strengthens the position. If it tries to become a general-purpose AI browser with entertainment features, it dilutes.

The early signal is positive: users describe Comet as making them “more productive at research tasks” and “integrating AI into professional workflows.” They’re not saying it makes them more curious or helps them explore. They’re saying it makes them more confident in their professional output.

Shopping Hub and Labs: The Position Boundary Test

The Shopping Hub presents a fascinating boundary test. Shopping recommendations don’t obviously fit “research confidence,” or do they?

It depends on what kind of shopping.

For impulse purchases and consumer goods, probably not. No one needs research confidence to buy a t-shirt. But for research-intensive purchases where you can’t afford to be wrong (B2B software evaluation, professional equipment, major business decisions), the position holds.

If Shopping Hub becomes “get AI recommendations for stuff you might like,” that’s off-position. If it remains “research significant purchases with confidence,” that’s on-position.

Labs (automation and advanced features) face similar dynamics. General productivity automation? Off-position. Automated research workflows that produce trustworthy output? On-position.

The strategic question isn’t whether these products are good (they might be excellent). It’s whether they reinforce or dilute the mental territory Perplexity owns.

What the Product Portfolio Reveals About the Actual Position

Here’s what’s becoming clear: the position might not be as narrow as “research confidence.” It might be broader: professional output you can trust with your identity.

The through-line across products:

• Search: answers you can cite professionally
• Deep Research: analysis you can present to stakeholders
• Email Assistant: drafts you can send without anxiety
• Comet Browser: research workflows that protect your credibility
• Enterprise features: organizational safeguards for AI-generated work

All of these solve the same core job: reducing the anxiety of being professionally wrong while leveraging AI productivity.

This is broader than curiosity (Aravind’s framing) but more specific than “AI assistant” (too generic). It’s about trusted professional output at scale.

The user identity remains consistent: people whose reputation depends on being right. Researchers, analysts, consultants, product managers, investors, executives and professional classes who can’t afford AI hallucinations or unreliable information because their credibility is their career currency.

The Conversion Metric as Position Proof

Return to that tweet about Email Assistant conversion rates. Why would a product have “unprecedented onboarding to retaining users conversion?”

Because it’s solving an acute professional pain point within an already-trusted context. Users who already trust Perplexity for research see Email Assistant and think: “If they can keep me from being wrong about facts, maybe they can keep me from being wrong about tone, details, and communication.” The trust transfers.

This is gravitational pull in action. The position creates permission to extend into adjacent professional domains where trust matters. It’s not about curiosity expanding. It’s about trust deepening into more aspects of professional life.

The risk is extending into domains where trust isn’t the primary currency. If Perplexity launches entertainment features, social tools, or consumer lifestyle products, they’d be moving outside the gravity well that makes their current expansion successful.

Strategic Implications for Product Development

Every new product should pass this filter: Does this help professionals avoid the anxiety of being wrong?

If yes:

• Legal research tools (lawyers can’t afford bad precedents)
• Medical research assistants (physicians need reliable clinical information)
• Financial analysis automation (investors need trustworthy data)
• Regulatory compliance research (companies need accurate requirements)
• Academic research tools (scholars need citable sources)

If no:

• Entertainment recommendations (no professional stakes)
• Social media tools (different identity dimension)
• Creative brainstorming (explicitly wants unpredictability)
• General life advice (trust threshold is different)

The product portfolio is actually revealing the position more clearly than the narrative does. Users vote with retention, conversion, and willingness to pay. The Email Assistant’s metrics aren’t lying. They’re showing that “trusted professional output” has more gravitational pull than “curiosity” ever could.

The gravitational pull is real.

Developers integrate the Sonar API because they trust Perplexity’s RAG capabilities. Enterprise buyers choose Perplexity because citations protect their reputation. Users pay for Pro because reliable research is worth $20/month. Media partners join the Publishers Program because Perplexity drives meaningful traffic with proper attribution. All of this feels “obvious” to Srinivas because the position is doing the heavy lifting, creating natural alignment between what Perplexity offers and what the market needs.

Part 5: The Coaching Moment

Let’s reframe the questions.

Instead of “How do we grow adoption?” ask “What concept should we own with increasing clarity?”

Right now, Perplexity is between two positions:

1. Curiosity (aspirational, abstract, what Srinivas wants to own)
2. Research Confidence (actual, concrete, what users experience)

The strategic question isn’t “Which features expand our reach?” It’s “Which position creates more defensible mental territory?”

Curiosity is harder to own because:

• It’s universal (everyone is curious about something)
• It’s immeasurable (how do you prove you enable curiosity?)
• It’s uncapturable (Google could claim it, Netflix could claim it)
• It doesn’t command premium pricing (curiosity feels recreational)

Research Confidence is more defensible because:

• It’s professional (targets high-value users with budget authority)
• It’s measurable (citation quality, source authority, answer accuracy)
• It’s specific (requires technical infrastructure that competitors can’t easily replicate)
• It justifies premium pricing (reliability has calculable ROI)

The recommendation: Fully embrace the Research Confidence position. Stop talking about curiosity as the core emotion. Start talking about confidence, reliability, and trust. Change “Where knowledge begins” to something like “Research you can trust” or “Knowledge worth staking your career on.” This isn’t abandoning the philosophical foundation (knowledge still has infinite value), but it’s acknowledging what users actually buy.

Instead of “What features differentiate us?” ask “What identity do we enable users to inhabit?”

The current feature roadmap seems driven by “answer engine” category expansion: browser, shopping, finance, and automation. These make sense from a product perspective. But from a positioning perspective, the question is: Does each expansion strengthen or dilute the “research confidence” identity?

The Comet browser: Does it strengthen research confidence? If positioned as “research workflow infrastructure,” yes. If positioned as a “general AI browser,” probably not. That’s a different identity.

The Shopping Hub: Does it strengthen research confidence? Only if shopping decisions are framed as “research-intensive purchases where you can’t afford to be wrong.” Consumer impulse buying? Doesn’t fit. B2B procurement research? Fits perfectly.

Labs automation: Does it strengthen research confidence? If it’s “automate reliable research workflows,” absolutely. If it’s “general productivity automation,” maybe not.

The strategic filter should be: Does this make users more confident in their intellectual output? If yes, it’s on-position. If no, it’s dilution.

Instead of “Which segments should we target?” ask “Which professional identities require absolute reliability?”

Current customer base skews toward: researchers, analysts, consultants, investors, product managers, marketers, developers. What do they have in common? Their professional reputation depends on being right. They’re not using Perplexity because they’re curious. They’re using it because being wrong is career-limiting.

This suggests that targeting expansion should focus on adjacent professional identities with similar needs:

• Legal research (reliability is legally mandated)
• Medical research (reliability is life-or-death)
• Academic research (reliability is publication-gating)
• Compliance and regulatory (reliability is legally required)
• Due diligence and M&A (reliability is financially material)

These are domains where “research confidence” carries premium value. Contrast this with expanding to general consumer use cases, where reliability matters less and price sensitivity is higher.

Instead of “How do we defend against Google/OpenAI?” ask “How do we deepen our mental territory?”

The competitive anxiety shows in Srinivas’s references to Sun Tzu and Google’s structural weaknesses. But here’s what positioning theory reveals: category kings don’t worry about feature parity; they worry about position dilution.

Google can’t truly challenge Perplexity’s position without abandoning its business model. They make $200+ billion annually from clicks. Any product that reduces link-clicking threatens that. Their AI Overviews attempt to straddle both worlds (synthesized answers, links, and ads), which means they can’t fully commit to either position.

OpenAI could challenge more directly. SearchGPT is explicitly attempting to own “AI research with citations.” But they face the opposite problem: ChatGPT’s identity is “creative AI companion,” not “reliable research tool.” Users have three years of experience with ChatGPT hallucinations. Overcoming that perceptual baggage is hard even with technically equivalent capabilities.

Perplexity’s best defence is position depth, not feature breadth. Instead of trying to do everything AI can do, double down on being the most reliable AI research tool. This means:

1. Resolve the copyright crisis authentically: Not through legal maneuvering, but through genuine publisher partnerships, reinforcing the trust position. The Publishers Program is in the right direction, but it needs to be expanded and made more prominent in the narrative.
2. Make citation quality the obsessive focus: Not just “we provide citations” but “we provide the best citations, most authoritative sources, most relevant passages, most transparent methodology.” This turns technical infrastructure (RAG) into position reinforcement.
3. Expand in domains that demand reliability: Deep Research for finance, specialized agents for legal, and medical research tools for physicians. Each vertical should reinforce “when it matters, use Perplexity.”
4. Create reliability metrics users understand: Not internal metrics like “query accuracy 92%” but user-facing signals like “citation authority score,” “source diversity index,” “recency timestamp.” Make reliability visible and quantifiable.

Instead of “Are we positioned correctly?” ask “Does our position align with our operational reality?”

This is the uncomfortable coaching question. The aspiration (curiosity) doesn’t match the actuality (reliability). And the actuality is threatened by operational choices that contradict it.

You can’t own “trust” while behaving in ways that break trust. The Cloudflare CEO calling Perplexity’s methods comparable to “North Korean hackers” isn’t just bad PR. It’s a direct assault on the position. Publishers claiming “unauthorized copying” and “sophisticated plagiarism” directly contradict “reliable, cited research.”

The path forward requires operational alignment with the position:

1. Ethical indexing must be non-negotiable: Not because it’s morally superior (though it is), but because the trust position requires it. Every crawler should be fully transparent. Every robots.txt should be respected. Every piece of content should be properly attributed.
   
2. Publisher relationships must be win-win: The Publishers Program can’t feel like damage control. It should be positioned as “we help premium publishers monetize their research” with revenue sharing that makes economic sense for both sides.
   
3. Internal metrics should prioritize position strength: Instead of tracking only query volumes, retention, and revenue, track:
   
   • Position awareness: % of the target market that associates Perplexity with “reliable research”
   • Position exclusivity: % that think of Perplexity first for trusted AI answers
   • Position resilience: Net Promoter Score among users who’ve experienced competitors
   • Position expression: % of Enterprise customers who cite “reliability” as the primary buying factor

What Srinivas Got Right (For Reasons He May Not Fully See)

This isn’t a critique. It’s a revelation of excellence.

Srinivas made brilliant positioning decisions, even if he explains them through a different framework. The insistence on citations? Genius. The freemium model gating advanced features? Perfect for the position. The focus on professional use cases? Exactly right. The technical investment in RAG over foundation models? Creates the moat that matters.

He correctly identified that Google’s advertising model is their structural weakness. He’s right that “answer engines” represent a category shift. He’s right that citation transparency matters profoundly. He’s right that elite technical talent creates defensibility.

Where the coaching opportunity exists is in language-reality alignment. The position is stronger than the narrative. The actual concept owned (Research Confidence) is more defensible than the aspirational concept (Curiosity). The users they’re attracting (professionals who can’t afford to be wrong) are more valuable than the users they think they’re serving (curious learners).

The gravitational pull is already there. Perplexity has achieved position-market resonance. The valuation, growth, and competitive positioning prove it’s working. The question isn’t whether it’s working. It clearly is. The question is whether it can be done more intentionally, with less cognitive friction between what they claim and what users experience.

The Invisible Choice Point

Every successful company reaches a moment where it must choose: optimize the story it tells itself, or optimize the reality it’s created. Perplexity is at that inflection point.

Path 1: Continue framing around curiosity, answer engines, and democratizing knowledge. Keep explaining success through technical architecture and category creation. Maintain philosophical commitment to infinite knowledge value. Handle copyright issues as necessary, but secondary concerns.

Path 2: Acknowledge that users buy reliability, not curiosity. Reframe narrative around “Research Confidence” or “Knowledge You Can Trust.” Make publisher relationships and ethical sourcing central to brand identity. Position-first product development, every new feature filtered through “Does this deepen trust?”

Path 1 isn’t wrong. The company is successful. The technology is impressive. The team is world-class. They’re solving real problems.

But Path 2 is more powerful because it aligns positioning (what you own in your customer’s minds) with operations (how you actually behave) with narrative (how you describe yourself). That alignment creates the strongest form of gravitational pull — when what you say, what you mean, and what you do are the same thing.

The irony is that Path 2 still honours everything Srinivas values. Knowledge does have infinite value. Citations do enable learning. Transparency does serve intellectual honesty. But the market buys reliability first, curiosity second. And in positioning, you must own the concept the market is actually buying, not the concept you wish they’d buy.

The Real Question

This isn’t “How should Perplexity grow?” That’s tactics.

It’s not “What features should they build?” That’s product.

It’s not even “How do they compete with Google/OpenAI?” That’s competition.

The real question, the one determining whether Perplexity becomes a category king or a well-executed feature, is:

What should Perplexity own in the mind of every professional who can’t afford to be wrong?

If the answer is “curiosity,” they’re building for the market they wish existed.

If the answer is “reliable research,” they’re building for the market that actually exists.

And if they choose the latter, they stop competing with Google for search dominance and OpenAI for AI ubiquity. Instead, they own an entirely different territory: the trusted layer between professionals and information. That’s worth far more than $18 billion. That’s worth whatever price professionals will pay to protect their reputation.

That’s the position.
That’s the gravity.

The rest is just tactics.

---

Uncover your position

Before you hire a messaging consultant to wordsmith your homepage, or an agency to “refresh your brand,” or someone to fix what they’ll call positioning (but is really just tactical framing), try this first.

The CEO Clarity Starter Kit

It does exactly what we just read. It helps you find and own your noun.

What you do:

• Run the Position Audit (reveals what noun you might already own without knowing it)
• Complete the 8-Question Advisor (the same questions that would surface “control” for 37signals)
• Feed the output into ClarityGPT (included)

What you get:

• Your noun. The concept you can actually own, not just claim
• A 4-Level Positioning Canvas showing how to move from saying it to OWNING it
• ClarityGPT translates your position into landing pages, offers, and LinkedIn profiles (written in your buyer’s voice, not consultant-speak)
• A 30-day positioning course so you can apply this method without me

Time required: About an hour (less time than reading three more case studies about tactics that won’t work without position)

Who’s used it: 200+ CEOs and founders who were tired of pushing uphill

Investment: $249 USD

Most realize they don’t need the consultant or agency after this. Or they need far less than they thought. Because once you know your noun (your position), the tactics become obvious. The distribution chooses itself. The customers explain you better than you explain yourself.

And yes, if you buy the kit, it nudges me closer to that Porsche in the photo. Thanks in advance for supporting excellent positioning and questionable life choices.

Stop competing on features. Start owning concepts.

Get your CEO Clarity Starter Kit