Nº 062 · Companies
Real cases of AI and AI agents at Brazilian companies
What companies are doing with AI, with which tool and with what result. Every case has a public source, a verification date and the numbers as published, with no estimates.
- Cases
- 62
- Sectors
- 22
- Sourced
- 100%
- Updates
- Daily
Sectors
22 sectors · tap to filter
Case index
62 of 62 cases
01
Itaú Unibanco
ia.i reached about 300,000 customers in its first phase; the legal solution runs more than 2 million tasks per month with efficiency gains of 20% to 80%
ia.i conversational assistant for customers and the IARA governance platform; legal solution covering 90 million documents
02
Nubank
The assistant resolves 55% of Tier 1 requests, handles more than 2 million chats per month, and cuts response time by 70%
AI customer service assistant, call center copilot, internal search and fraud QA with computer vision
03
Bradesco
82% first-level resolution rate and 89% first-week retention; BIA usage grew 8x
BIA, a virtual assistant, integrated with generative AI for customers and managers
04
Bradesco
10% to 20% gains in credit recovery capacity
AI collections agent that transcribes and analyzes agent conversations and suggests improvements
05
Bradesco
65% more efficient audit planning
AILA, an AI assistant built by the internal audit team itself
06
Magalu
More than 20 agents in production; results detailed in the Bain case
Lu assistant on WhatsApp with more than 20 orchestrated AI agents for search, recommendation, customer service and purchasing
07BDM (Big Data Mining)
85% faster analyses, 98% accuracy, 40% less operational effort
LOUIS, a generative AI model for document processing used by a financial institution
08
Condor
Prototype built in weeks; greater efficiency in ticket search (no figure published)
Generative AI prototype for IT support and help desk, built on a knowledge base and ticket history
09
Gimba
Time to register each item dropped from 13 to 2 minutes; solution cost 50% lower
Platform that generates descriptions for a catalog of 30,000 products
10Grupo Elfa
Quoting time dropped from 40 to 2-3 minutes; more than R$ 100 million in new revenue in 8 months
CotAI, an AI that reads and answers thousands of quote requests received by email
11
iFood
Every order goes through more than 100 AI models (no result metric for GenAI)
Generative AI for experience personalization and developer productivity
12
Banco Inter
Investigation capacity increased by 108%
Penélope, an agent that automates financial crime investigations
13Sinosserra
12,000 bonds per month; validation dropped from 24 hours to 1 minute
Automatic document validation for capitalization bonds
14
Webmotors
8.5x higher engagement
Agents that respond to vehicle negotiations instantly and in a personalized way
15Lina Prades (joalheria)
Two management apps created; one supports daily operations
An entrepreneur used ChatGPT for content, ads, pricing and inventory, and built two management apps with no technical background
16Tapioca da Ari
A more organized and visible business (qualitative account)
Using ChatGPT to develop recipes, the menu and customer communication
17
Copel
Queries went from a day to near real time
AI agent that answers natural-language questions about SAP ERP data
18Galaxies
Research time 97% shorter; cost per study from R$ 1 million to R$ 150,000
AI-generated synthetic personas for market research
19
Genial Care
Documentation dropped from about 2 hours a day to 3 minutes per session
LIAM, which generates clinical session records from audio
20
Sami Saúde
81% of records approved without edits; more than 900 hours saved per month
Medical records generated by AI from text, images and audio
21
3 Corações
299 agents created; 390.5 hours saved; 98% adoption
Adoption of Gemini Enterprise with agents built by the teams themselves, such as the Stalled Orders Auditor
22
Sem Parar
Average handling time 9.3% shorter; accuracy above 92%
Customer Service Copilot for call center operators using RAG
23
Vivo
Average handling time 9% shorter (consumers) and 4% shorter (businesses)
I.Ajuda, a generative AI assistant for 11,000 call center agents
24
CAIXA
Processes that took 10 to 12 days now take about 30 minutes; 70% to 80% gains in back office
Copilot and agents in Copilot Studio for back office, legal, compliance and customer service
25
Tecban
Up to 35% productivity gain
Agents for pre-sales, compliance, audit, HR and field operations
26Cielo
Tickets reduced by 52.2%
AI agent for internal IT support
27A1 Inteligência em Viagens
55% increase in travel coordination efficiency
Fifi, an agent for customer service and travel policies
28
Marcopolo
Six agents built in four days
Training teams to build AI agents
29
YDUQS
Over 90% accuracy, average response time of 4 seconds
Automatic classification of enrollment letters
30Prodam
See case (development and testing automation)
Automation of development, code validation and testing
31ProBrain
AI costs 89% lower
Migration of AI workloads to Vertex AI
32180 Seguros
BigQuery queries three times faster
Experimental conversational base for partner developers, integrated with the API
33
Direcional Engenharia
Analysis time 20% shorter
Generative AI to analyze customer relationship calls
34Totaldocs
Search and analysis time 70% shorter
Document search and analysis with generative AI
35Alfaneo
Petition drafting time 94.45% shorter
Drafting legal petitions with generative AI
36
Hotmart
Support workload reduced by up to 50%; 1.6 million questions answered
Generative AI in support for digital entrepreneurs
37
PicPay
NPS up 45%; understands 95% of conversations
Customer service assistant with generative AI
38
Stone
93% positive rating
Customer service with generative AI
39
will bank
Satisfaction 37% higher; human workload 26% lower
Customer service with generative AI
40
Hospital Israelita Albert Einstein
2,350 consultations; in testing, more than 200 hours saved and 98.9% acceptance of suggestions
AI that transcribes consultations in urgent care and generates a draft of the medical record
41Bondy
Around 85% of cases resolved (according to the company)
AI assistant on WhatsApp that answers HR questions for clients such as Ambev, Raízen and Mercado Livre
42
Localiza&Co
See the article
AI for assessing vehicle damage and analyzing the fleet
43Smartcoop
20,000 hours saved per year on searches for farmers and cooperatives
Ana, a generative AI virtual assistant for farmers and cooperatives in Rio Grande do Sul
44Integrada Cooperativa Agroindustrial
An analysis that used to take 30 days now takes 20 minutes
Farmable Areas Project: satellites and AI to analyze farmable land
45
Base39
Credit analysis cost 96% lower; decision time cut from three days to under an hour; infrastructure 84% cheaper
Credit analysis with generative AI, vector search and Claude
46
Bora Brasil
62 employees using it; an average of at least 4 hours saved per week; about 120 hours per week given back to the group
An AI office powered by Claude, built on the group's own data
47Alfa Transportes
About 4,000 work hours saved per month; only 1% of documents go to manual review
Real-time validation of delivery receipts with AI
48
MBRF
Pilot with 16,000 deliveries: 5% fewer delays en route and routes completed about 7% faster
AI for resequencing delivery routes
49
Prestex
30% resolved without human intervention; response time cut from hours to under 15 minutes in six months
AI agent for customer service
50Prefeitura de Rio Verde (GO)
64,975 requests handled since January 2026; 40% of requests resolved without human intervention
Unified AI chatbot service covering 24 municipal departments
51Prefeitura de Belo Horizonte
More than 57,000 chats handled in three months
Bel, an AI-powered citizen service tool
52
Verocard
Analysis time cut from 12 hours to 30 minutes (95% faster); operating costs 30% lower
Analysis of public procurement notices with generative AI and RAG
53
Pernambucanas
80% less manual review in fraud identification; more than 10,000 documents per day
AI for data extraction and document fraud detection
54OpenCo
About 95% of alerts resolved autonomously; average resolution time 75% shorter (from 12 to about 3 minutes)
AI agents that resolve infrastructure alerts (Amazon EBS) by following runbooks
55Maribel Clean
Daily sales 5x higher; 80% of conversations resolved without a human; about 6 hours saved per day (company data)
WhatsApp agent that handles quotes, scheduling and late-night inquiries for an upholstery cleaning company in Brasília
56
Osklen
Abandoned cart conversion up from 3.39% to 18.18%; R$ 453,000 in incremental revenue in 30 days
AI agents on WhatsApp for support, post-sale service and cart recovery
57
Pit Bull Jeans
WhatsApp's share of e-commerce sales doubled; customer service team reduced from 50 to 9 people
WhatsApp AI agent that handles triage, order status and cart recovery
58Jotanunes Construtora
Time to find a procedure cut from 5 minutes to 30 seconds (90% less)
Oráculo, a Teams agent that answers questions about technical procedures and construction standards
59
Grupo Boticário
Up to 80% autonomous resolution of operational requests; 90% of direct sales leads qualified by AI
In-house platform with more than 30 agents across the reseller journey
60Nexo Contabilidade
Monthly close brought forward by two days; about 40 hours per month given back to the team
AI for accounting close, customer service and document collection
61São Caetano Assessoria Contábil
50% higher service capacity
AI for WhatsApp customer service at an accounting firm
62Syhus
660 hours per month freed up; 25% growth without expanding the team
AI to automate invoice checking and accounting routines