San Francisco-based artificial intelligence (AI) startup Perplexity recently closed a monumental $73.6 million Series B funding round to scale up its conversational search and information retrieval platform built specifically for the enterprise.
Perplexity‘s Mission to Transform Enterprise Knowledge Management
Founded in 2022 by former AI researchers Aravind Srinivas, Raghu Chittarvu and Sam Skjonsberg, Perplexity has developed a cutting-edge natural language search engine designed from the ground up to deliver accurate, transparent and fully-cited answers to user questions.
"In an era where misinformation and AI hallucinations are causing increasing concern, accuracy and transparency are fundamental prerequisites to making AI-powered search ubiquitous," emphasizes Aravind Srinivas, Co-founder and CEO of Perplexity.
The startup is on a mission to enhance how knowledge workers and enterprises find, analyze and share information. Its long-term vision is to catapult organizations into an age of accelerated productivity by creating ubiquitous conversational interfaces to workplace knowledge.
Perplexity‘s vision to lead the paradigm shift towards conversational experiences for enterprise search and information management
How Perplexity‘s Conversational Engine Works
Many conversational AI platforms today rely on brute-force large language models that are prone to hallucination under the guise of human conversation.
In contrast, Perplexity has built an enterprise-ready search framework from first principles focused on accuracy and reliability to meet the needs of specialized workplace scenarios:
Structured Knowledge Graph
At the core of Perplexity‘s engine lies a vast curated knowledge graph interlinking domains, entities, concepts and relationships within enterprise datasets. This structured representation of organizational knowledge builds the foundation for in-depth reasoning.
Neural Information Retrieval
The startup combines graphical knowledge representation with deep natural language understanding models including BERT, RoBERTa and T5 transformers. Jointly, these neural networks empower the platform to parse user queries, map intent to relevant enterprise data nodes and surface insightful answers.
Recursive Question Answering
Unlike standard chatbots, Perplexity‘s conversational flow allows recursively asking and clarifying questions to extract comprehensive, multilayered answers from the knowledge graph without losing context. This drives more productive human-AI interactions.
Explainable Rationale and Sourcing
Perplexity also stands out with its ability to provide complete audit trails for answers indexed back to the underlying structured data sources. This explainability around answer provenance builds crucial transparency into the process to maintain trust.
This powerful combination of knowledge representation, deep NLU and recursive question answering is what unlocks Perplexity‘s capability to deliver a transformational change in enterprise search and analytics, tailored to organizational data environments.
Perplexity Product Portfolio Targets Enterprise Scale
On the heels of a successful $25.6 million Series A funding round in March 2022, the Perplexity team has been heads-down building out its product portfolio to serve enterprise customers.
It recently launched Copilot – an AI Research Assistant for knowledge workers and academics that can comprehend complex questions and provide citations and links to original sources alongside answers.
Other products include real-time analytics for market, customer and product research and customizable conversational interfaces for enterprise search applications.
Within a year since its public launch, Perplexity has impressive traction:
- 10+ million monthly active users on its consumer search engine
- 500+ million queries processed per month
- 131% month-on-month growth in 2022
- Partnerships with leading academic publishers and research institutions
"We have seen exponential pull from global enterprises in regulated industries like financial services, life sciences and energy where data integrity is paramount," reveals Co-founder Sam Skjonsberg. "This underscores a major mainstream breakout on the horizon for enterprise conversational AI."
Key metrics showing Perplexity‘s explosive early growth
Marquee Investors Fuel Perplexity‘s Rise as Conversational AI Leader
Perplexity‘s remarkable early traction is also evidenced in investor confidence, with its Series B round massively oversubscribed at $73.6 million.
The round saw participation from premier Silicon Valley firms, top global venture capitalists and entrepreneurs and industry specialists:
- Led by IVP, one of the Valley‘s most reputed late-stage investors
- Existing investors include NEA, Databricks, Elad Gil and GitHub CEO Nat Friedman
- New individual investors: Jeff Bezos via Bezos Expeditions, Shopify CEO Tobi Lutke, angel investor Naval Ravikant, Balaji Srinivasan and others
- Institutional investors: NVIDIA, Bessemer Venture Partners, IVP, B Capital Group
"The overwhelming investor interest reflects growing conviction that the future of search lies in conversational AI," notes Somesh Dash, General Partner at IVP, who also joins Perplexity‘s board. "Moreover, accuracy and reliability will prove the biggest differentiators as hands-free interfaces become ubiquitous."
This perspective is corroborated by Elad Gil, prolific investor in category-defining startups like Airbnb, Coinbase, Instacart, Stripe and Square.
"Perplexity‘s rapid growth since inception shows incredible product-market fit. Their team is poised to lead the transition from keywords to conversations as the next paradigm for search and analytics."
Breakdown of Perplexity‘s investors signaling market enthusiasm for Enterprise Conversational AI
Outlook and Future Plans
Perplexity will deploy the new capital infusion to intensify its go-to-market push across North America, Europe and Asia-Pacific – building on its technical leadership in conversational search for the enterprise.
The company has swelled from 15 to over 80 employees since its Series A as it densifies its engineering, product and customer success teams. It aims to double headcount by mid-2023 with key hires in AI research and enterprise sales. Satellite sales offices will also be launched in strategic global tech and finance hubs.
With Microsoft deploying OpenAI across its stack and Google recently unveiling its own conversational search projects, competition in the still-nascent space is mounting.
However, Perplexity‘s differentiation lies in its exclusive focus on precision, transparency and trust for enterprise search use cases. This gives it a unique edge as the market matures and conversational interfaces become ubiquitous while large language models receiving growing backlash over accuracy and ethics issues.
Perplexity‘s accuracy-first approach for Enterprise Conversational AI contrasts with Microsoft and Google‘s chasing raw scale without guardrails
As per Emergen Research, overall Conversational Artificial Intelligence market revenues are projected to hit $32 billion by 2030 expanding at an explosive 22% CAGR, underpinned by demand across healthcare, retail, automotive and banking sectors.
Within this segment, Perplexity‘s purpose-built stack for enterprise search, specialized domain expertise, rocketship growth and elite funding position it at pole position to consolidate market leadership over the coming decade.
Challenges and Limitations in Conversational AI
While conversational interfaces represent an intuitive paradigm shift, open-ended dialog with machines on arbitrary topics at human levels of understanding remains an AI-complete challenge. Several tough problems plague the field:
Curse of Dimensionality
Language complexity grows exponentially with longer dialog context, diverse domains and complex entity relationships – requiring prohibitively high data and compute scale. This drives inconsistencies and fragility when generalizing models.
Interaction Challenges
Conversational systems like Perplexity also need balanced human-AI collaboration, explainability and clarification dialogs to productively resolve information requests. Transparent rationale around answers augments trust in recommendations.
Data Integrity and Compliance
For enterprise use within regulated industries like Financial Services, Healthcare and Energy – accuracy, provenance, explainability and transparency are prerequisites. This demands specialized frameworks beyond raw language model scale.
Societal Risks
As AI assistants grow more pervasive, risks around user privacy, manipulation, misinformation and addiction also raise ethical and regulatory oversight questions. The space continues to rapidly evolve on technology, governance and standards fronts.
While challenges remain, Perplexity‘s rigorous focus on precision Conversational AI for Enterprise use cases provides a robust platform for innovation while proactively mitigating risks from broader application of AI technologies across consumer and internet platforms.
The Future of Enterprise Search – Conversational and Visual
Looking ahead, the next frontier for knowledge discovery leverages synergies between conversational querying and data visualization to intuitively navigate complex information spaces using natural language and visual interfaces.
Imagine analysts conversing with an enterprise search engine to explore various business scenarios. The system surfaces insights drawing on historical data while dynamically generating interactive data visualizations to illustrate key patterns, trends and outliers. Users fluidly refine parameters and recommend actions by iterating through generations of data-driven charts and recommendations using speech and conversational gestures.
This multimodal investigative search process supercharges how enterprises make data-driven decisions – a transitional step until immersive augmented analytics via virtual and mixed reality headsets gains maturity.
As a pioneer building the foundational blocks for next-generation enterprise search with Conversational AI, Perplexity sees voice, vision and immersion driving the future of business intelligence and knowledge discovery.
