01 4–8 weeks
RAG in your AWS account
Ingestion, index, a Bedrock model with source citation and an evaluation set, deployed in your account.
AI inside your own AWS account
Amazon Bedrock gives you Claude, Llama, Mistral and Amazon’s own models without taking the data out of your AWS account. We build agents, RAG and LLM applications on top of it, with your IAM roles and the usage billed to you.
2,000,000 active users served · 10 countries in operation
Built on
An engineering team that has already been on the other side.
Ninety seconds: who we are, how we work and what you get at the end.
An agent that takes the event, decides and leaves a record
An email, an order or an alert comes in. The agent queries your systems, decides by your rules, asks for approval when needed and executes. Pick a case and watch it run.
Example: a customer asks what their policy covers. The system answers with the exact page of the contract and opens the case.
01Event arrives
Does it cover theft from a public car park?
context
product = motor_premium
02Plans
Queried · policy 88213
Read · clause 4.3
Decided · open a claim
Queried · CRM
Queried · price list
Drafted · proposal
Read · invoice 9902
Cross-checked · PO-4471
Detected · discrepancy
03Queries the systems
Policy_wording_2026.pdf84 pp
Motor_exclusions.pdf12 pp
Claims_email1,240
ERP · policies38,400
Price_list_2026.xlsx118 rows
CRM · opportunities2,310
Proposals_202556 PDF
Sales_email4,800
Supplier_invoices3,900 PDF
ERP · purchase_orders4,110
Framework_contracts.pdf22 docs
Payments_email2,100
84 pp → 612 chunks
04Human approval
Asked for approval · human
ApproveReject
05Executes and logs
- Queried · policy 88213
- Read · clause 4.3
- Decided · open a claim
- Asked for approval · human
- Executed · created the case in the ERP
- Queried · CRM
- Queried · price list
- Drafted · proposal
- Asked for approval · manager
- Executed · sent and logged
- Read · invoice 9902
- Cross-checked · PO-4471
- Detected · discrepancy
- Escalated · human
- Logged · case opened
log · 5 steps · who and when
Swipe to follow the flow →
Why AWS
If your operation already lives on AWS, the model can too
When the documents sit in S3, the database in RDS and the permissions in IAM, sending every query out to a third-party API opens a conversation with security that Bedrock avoids: the model is called from the same account, under the same roles and the same logs.
The useful question is not which cloud has the best AI, but where the data and the team that will maintain the system already are. If the answer is AWS, we build there.
The pieces, inside your account
01
RAG over your own data
With Bedrock knowledge bases when the case is standard, or with pgvector on the PostgreSQL you already run when you need more control. Either way, measured with recall@k before launch.
02
Claude and other models with per-task routing
An inexpensive model classifies and summarizes, a capable one writes the final answer, and the cost per query stays visible.
03
Agents that query your systems
With IAM roles scoped to what each tool needs and human approval on steps with consequences.
04
Input and output filters
Bedrock Guardrails to block topics and personal data, on top of the defenses against prompt injection.
01
Formats
Scope and timelines
02 2–4 weeks
Moving a pilot to Bedrock
A prototype built on a provider’s public API, moved to Bedrock with quality measured before and after.
03 1 week
Diagnostic of what you already run on AWS
Retrieval and answers measured separately, cost per query, and failures ranked by impact with their fix.
When not to hire us
If you are not on AWS, we will not move you there for this. The same architecture runs on Azure or on the provider’s direct API.
If the work is the infrastructure itself —network design, landing zone, server migration— we are not the team. We build the AI application that runs on top, alongside the infrastructure team you already have.
Related services
RAG systems
RAG that answers well — and knows when to stay quiet
Chunking, reranking and hybrid search, evaluated with recall@k and NDCG.
Language models
LLM applications that survive real users
LLM applications with structured output and tool calling.
AI agents
Agents that do the work, not just talk about it
LangGraph orchestration, durable state and human approval on steps with consequences.
AI on Azure
AI inside your own Azure subscription
Microsoft Foundry, Azure OpenAI and AI Search inside your subscription.
Key takeaways
Four things before the call
- 01
Amazon Bedrock is the AWS service for using language models —Anthropic’s Claude, Llama, Mistral and Amazon’s own— inside the company’s AWS account, through a single API.
- 02
What is sent to Bedrock is not used to train the models and is not shared with the companies that build them.
- 03
Quarl uses Bedrock behind the same abstraction layer as OpenAI and Azure OpenAI: switching models is configuration plus an evaluation run, not a rewrite of the application.
- 04
Resources, credentials and usage stay in the company’s name, billed by AWS, with no Quarl account in the middle.
This is where the data comes in
An AI system is worth what its sources are worth. These are the standard connectors; anything with an API or a database connects the same way, and what has no API is handled by file.
SAP
Enterprise ERP
Oracle
ERP and database
NetSuite
Cloud ERP
Salesforce
CRM and service
HubSpot
CRM and marketing
PostgreSQL
Database and pgvector
Microsoft SQL
Database
Snowflake
Data warehouse
BigQuery
Google data warehouse
Databricks
Data platform
Redshift
AWS data warehouse
Synapse
Azure data warehouse
Supabase
Managed Postgres
Workday
Payroll and HR
QuickBooks
Accounting
Sage
Accounting and ERP
Xero
Cloud accounting
Shopify
Catalogue and orders
WooCommerce
Catalogue and orders
Magento
Catalogue and orders
Stripe
Payments and subscriptions
Google Drive
Documents and folders
CSV y Excel
Flat files
Nothing in this category
A RAG system in production: three years, two million users
We built and operated the assistant for a loyalty platform serving more than two million active users across ten countries.
We built the full pipeline: document ingestion and normalization, chunking, embedding generation, vector store on Azure AI Search and Pinecone, and retrieval with grounded generation on LangChain.
We held it above 99% availability for three years.
RAG systems →Frequently asked questions
01 Which models are available on Amazon Bedrock?
Anthropic’s (Claude), Meta’s (Llama), Mistral, Cohere and Amazon’s own, among others, behind a single API. The catalogue varies by region, so before proposing anything we check what is available in the region where you will deploy.
02 Is our data used to train models?
No. AWS does not use what is sent to Bedrock to train models or share it with the model providers. The documents, the index and the logs stay in resources in your account, under the IAM permissions you already manage.
03 Is Claude on Bedrock the same as on the Anthropic API?
It is the same model, served from AWS infrastructure. Two things change: some new features reach the direct API first, and quota and billing are those of your AWS account. For most enterprise cases that difference weighs less than having everything under one contract.
04 Who pays for the AWS usage?
You do, directly to AWS, in your own account. Quarl does not resell usage or put an account of its own in the middle. What we do deliver is the measured cost per query, so the invoice is not a surprise when usage grows.
05 Can you work on what we already have on AWS?
Yes. If a system on Bedrock already exists and answers badly, we start by measuring it with real questions and their correct answers, as in a rescue, and the fix is made on your resources, without rebuilding what works.
Fifteen minutes. A concrete answer.
It gets fixed, it gets built, or it is not worth it. And if the diagnostic does not reach three actionable findings, it is not charged.