Customer Service Automation
KABAM Robotics wanted to make its service robots more intelligent and useful across security, facilities, and hospitality settings while keeping experimentation fast and model deployment affordable.
Reported impact
KABAM Robotics improved robot intelligence and concierge capabilities, accelerated model experimentation, and reduced the cost of productizing AI models to one-tenth of maintaining equivalent hardware while supporting deployments across four continents.
KABAM Robotics built its Smart+ cloud robot-management platform and REMI intelligence layer on AWS to improve the concierge and operational abilities of service robots deployed across four continents. Amazon Bedrock gives the team fast access to large language models through APIs so it can experiment and choose the right model behavior for robot interactions. Amazon SageMaker is used to deploy open-source vision-capable models such as Llava, Amazon EKS provides the managed Kubernetes environment for cloud-side robot services, and Amazon Polly adds voice capabilities for more natural interactions. This stack lets KABAM combine generative AI, deployed models, orchestration, and speech into a single robot platform rather than maintaining specialized hardware for each function. The result is faster experimentation, lower model-operating cost, and more capable robots for customer-facing environments.
Robot-management workflows are coordinated through the Smart+ cloud platform; Amazon Bedrock and SageMaker-backed models provide language and vision intelligence, Amazon EKS hosts the cloud services, and Amazon Polly turns responses into customer-facing voice interactions for the robots.
REMI – a groundbreaking system equipping their robots with the power of large language models, seamlessly integrated using Amazon Bedrock API
The startup created Smart+ – a cloud-based robot management platform
With Amazon SageMaker, we were able to deploy open source model Llava with ease, at a tenth of the cost of maintaining our hardware and productizing the models.
while utilizing services such as Amazon SageMaker Amazon EKS and Amazon Polly to further enhance the overall functionality of their robots.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
The Organization JSON-LD lists addressLocality 'Singapore', and addressCountry 'SG', and numberOfEmployees value 55.
Related
AWS Sample Sep 11, 2024 Georgia-Pacific The lack of a centralized, readily accessible knowledge base often leads to reduced machine productivity, prolonged machine downtime, and higher troubleshooting and repair costs. Making phone calls to experts and… Accelerated operator onboarding Amazon Bedrock · Amazon Kinesis 
In-house Sample Aug 27, 2026 Lenovo Lenovo needed to handle customer support at very large scale across its global device support operation, where routine issues created cost and complexity and human-led resolutions costing $30 to $50 could delay simple… $30 to $50 Lenovo Intelligent Voice Agent · Lenovo Machines Fix Machines Agent · Lenovo Smart Repair Agent 
Atos Sample Jul 29, 2026 Misumi USA Misumi USA needed to handle about 2,000 inbound emails and more than 3,000 attached documents each day, but manual routing and document handling caused frequent misrouting, slowed order and quote processing, and made it… 14 years reduced to 2 years for web category transformation AWS Lambda / API gateway · Amazon Bedrock · Amazon Connect 