Use Case Intelligence RSS Search use cases Search
AWS Software / SaaS United States

Perplexity

Enterprise Knowledge Search

Perplexity wanted to offer a powerful alternative to the traditional online search engine, so it created an interactive search companion that provides personalized and conversational answers that are backed by a curated list of sources. To simplify access to proprietary models, such as the popular cutting-edge LLM Claude from Anthropic , and to fine-tune open-source LLMs, Perplexity needed a powerful global infrastructure for its search engine, Perplexity AI.

Read the original story Explore all use cases
And Anthropic made further improvements in reducing hallucinations and increasing accuracy with the Claude 3 family, which has even further improved accuracy over Claude 2.1.
Anthropic’s previous model, Claude 2.1, had already reduced their rate of hallucination by half.

Reported impact

Business outcome

Perplexity AI continues to offer users a selection of models that suit their needs, automatically accessing the recent iterations of Claude and driving the availability of new features for users. “On AWS, we have a highly reliable experience with all the pieces of infrastructure that need to come together to make our complex product work,” says Heydari. Key reported metrics include Anthropic’s previous model, Claude 2.1, had already reduced their rate of hallucination by half.; And Anthropic made further improvements in reducing hallucinations and increasing accuracy with the Claude 3 family, which has even further improved accuracy over Claude 2.1..

  • Customer Experience
  • Risk Reduction
  • Scalability

Solution

Learn how Perplexity’s AI-powered search engine uses Amazon Bedrock and Anthropic’s Claude 3 for accurate, comprehensive answers to user queries. Perplexity wanted to offer a powerful alternative to the traditional online search engine, so it created an interactive search companion that provides personalized and conversational answers that are backed by a curated list of sources. Users can choose among multiple high-performing large language models (LLMs) for relevant, accurate, and understandable information. To simplify access to proprietary models, such as the popular cutting-edge LLM Claude from Anthropic , and to fine-tune open-source LLMs, Perplexity needed a powerful global infrastructure for its search engine, Perplexity AI.

Data flow

Source data enters the Perplexity workflow through its application services, processed with Amazon Bedrock, Claude 3, Stability AI.

Solution architecture

7 components · 1 layer
  1. Compute
    • Amazon Bedrock Hosts the foundation models and inference APIs that power the generative AI features.
    • Claude 3 Serves as the primary large language model for generation or analysis tasks.
    • Stability AI Supplies the image generation model used to create creative assets or ads.
    • Mistral AI Supplies open-source or commercial foundation models used in the solution.
    • Amazon EC2 P4de Instances Provides GPU infrastructure for large-scale training workloads.
    • Amazon SageMaker Supports model training, fine-tuning, evaluation, or managed inference for the customer workload.
    • Amazon SageMaker HyperPod Provides clustered infrastructure for large-scale foundation model training.

Architecture clues

  • In addition to running its own models on AWS, Perplexity offers its users access to Claude through Amazon Bedrock , a fully managed service that offers a choice of high-performing FMs from leading AI companies such as AI
  • “Amazon SageMaker HyperPod has been instrumental in driving our AI innovation.” Perplexity AI continues to offer users a selection of models that suit their needs, automatically accessing the recent iterations of Claude
  • “The infrastructure was simple to manage, and the hardware-related failures dramatically reduced.” To learn more about how Perplexity accelerates foundation model training by 40% with Amazon SageMaker HyperPod, read this
  • “We stay on the cutting edge of AI capabilities, use powerful models, and are open to anything that enhances our user experience.” Using a high-performing service such as Amazon Bedrock means we are tapping into Anthropi

Evidence from the source

And Anthropic made further improvements in reducing hallucinations and increasing accuracy with the Claude 3 family, which has even further improved accuracy over Claude 2.1.
Anthropic’s previous model, Claude 2.1, had already reduced their rate of hallucination by half.
Because Claude provides information in concise, natural language, users can arrive at clear answers quickly.
Perplexity AI continues to offer users a selection of models that suit their needs, automatically accessing the recent iterations of Claude and driving the availability of new features for users.
Perplexity wanted to offer a powerful alternative to the traditional online search engine, so it created an interactive search companion that provides personalized and conversational answers that are backed by a curated list of sources. To
Users can also quickly upload and analyze large documents because Claude 3 models feature a context window of 200,000 tokens, the equivalent of roughly 150,000 words or more than 500 pages.
Transparency Evidence & methodology 95/100 reference value · 90% confidence · 2 sources

Review how this reference was scored and trace each normalized claim back to its documented source.

01

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value95/100
  • Problem clarity20/20
    Measures whether the customer's operational problem is specific enough to compare.
  • Implementation depth20/25
    Rewards named components, data flow, integrations, and architecture detail.
  • Outcome specificity20/20
    Measures whether the reported benefit is concrete rather than a generic claim.
  • Quantified impact25/25
    2 metrics describe measurable impact.
  • Deployment scale10/10
    Rewards evidence of a live or scaled deployment and meaningful integration scope.
Evidence confidence90%
  • Source authority20/25
    The main evidence is an official vendor-published customer story.
  • Claim traceability23/25
    11 of 12 key fields link to a named source.
  • Direct evidence20/20
    6 direct source quotes retained.
  • Date provenance8/10
    The publication date comes from publisher-controlled metadata.
  • Source transparency9/10
    2 source references document publisher, URL, authority, retrieval, and supported fields.
  • Implementation specificity10/10
    Rewards source-backed components, data flow, quantified impact, and a specific outcome.
02

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

  1. AWS

    Perplexity customer story

    First-party
    https://aws.amazon.com/solutions/case-studies/perplexity-bedrock-case-study/
    Source type
    Vendor-published customer story
    Published
    Apr 15, 2024
    Supports
    • Customer
    • Business problem
    • Solution
    • Products
    • Solution components
    • Data flow
    • Integrations
    • technical area
    • use case category
    • Business outcome
    • outcome category
    • Metrics
    • Architecture
    • evidence quotes
    • Publication date
  2. LinkedIn

    Perplexity | LinkedIn

    Third-party
    https://www.linkedin.com/company/perplexity-ai/

    The public company page includes numberOfEmployees value 1408 and a PostalAddress with streetAddress 115 Sansome St, addressLocality San Francisco, and addressCountry US.

    Source type
    structured profile
    Supports
    • Region

Date evidence

Not separately sourced Industry, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Related

Similar use cases

See all like this
AWS Feb 14, 2025 MANZ Software / SaaS Austria MANZ , an Austrian company that provides information services for legal and tax professionals, sought to revolutionize how its registered users searched its vast repository of legal documents. 30% expected savings Enterprise Knowledge Search Amazon Neptune · Amazon OpenSearch Service AWS Jan 28, 2025 Bynder Software / SaaS Global Bynder initially acquired a company to incorporate visual search capabilities into its software. The fast rate of customer adoption encouraged Bynder to build on its success. One of Bynder’s customers states that the time spent searching for assets for a typical campaign task has… Enterprise Knowledge Search Amazon Bedrock · Amazon Titan Multimodal Embeddings AWS Dec 11, 2024 Goodnotes Software / SaaS Global Founded in 2011, Goodnotes has been a leader in digital note-taking and productivity, now serving over 24 million monthly active users. As a pioneer in digital handwriting, Goodnotes was among the first apps to support… Founded in 2011, Goodnotes has been a leader in digital note-taking and productivity, now serving over 24… Enterprise Knowledge Search Amazon Bedrock · Amazon CloudWatch · Amazon EKS