AI agents
Taking AI agents from a demo to production on AWS: the architecture, the guardrails and what they cost.
15 posts, newest first. Every subject
2026
-
Armakuni Achieves the AWS Agentic AI Specialization
In January 2026 we earned the AWS Agentic AI Specialization, a new specialization within the AWS AI Competency program.
News. Armakuni. A two-minute read.
-
How Agent Core Will Redefine the Future of AI on AWS
Enterprise AI is past the experiment stage. Why Amazon Bedrock AgentCore, with governed and audited agents, is becoming the standard way to deploy AI on AWS.
Armakuni. A five-minute read.
2025
-
Stop Buying AI, Start Building AI Readiness
The advantage comes less from which AI you buy than from how quickly your organization can experiment with it and ship it, with good governance around it.
Armakuni. An eight-minute read.
-
Adoption of AI does not translate to more profits
Plenty of companies have adopted AI and few are seeing the profit. The post argues that strategy decides it, more than the technology does.
Armakuni. A six-minute read.
-
Armakuni Signs Strategic Collaboration Agreement with AWS to Accelerate Generative AI Adoption at Scale
We signed a Strategic Collaboration Agreement with AWS to help customers adopt generative AI faster and at scale.
News. Armakuni. A two-minute read.
-
Armakuni in the AWS Summit Los Angeles 2025
We were at AWS Summit Los Angeles 2025, talking to teams about modernizing their workloads and putting AI to work on AWS.
Event recap. Armakuni. A two-minute read.
-
Deploying LLM-Powered Android Agents on AWS
UI scripts break whenever an app moves a button or spots the automation. This post covers Android agents on AWS that use an LLM to read the screen and act on it instead.
Justin Mitchell. A three-minute read.
-
Revolutionizing Android Automation: From ADB Commands to Real-Time Data Interaction
ADB can launch apps and tap buttons, but it can't see what's on the screen. Adding Android's accessibility APIs gives automation that awareness, and fewer flaky tests.
Justin Mitchell. A four-minute read.
-
Integrating AI With Existing Infrastructure: Easy AI Solutions for Small Businesses
A demo showing a small business adding AI to the systems it already has with Amazon Nova Sonic, without a complex integration.
Justin Mitchell. A six-minute read.
-
Generative AI in Production: How We Use Amazon Nova to Test Voice Agents at Scale
Voice agents are hard to test because they answer speech in real time and users behave in all sorts of ways. How we use Amazon Nova to test them at scale, with a demo of our CLI tool.
Justin Mitchell. A four-minute read.
-
Armakuni in the AWS Summit New York 2025
At AWS Summit New York 2025 we talked about the PUSH framework, our way of checking that a generative AI project is solving the right problem before it's planned.
Event recap. Armakuni. A two-minute read.
-
Simplifying Vector Database Scaling with AWS
For large vector datasets that grow quickly, Amazon OpenSearch, Amazon RDS and Bedrock's vector store each have strengths and trade-offs. How to choose and scale them.
Armakuni. A three-minute read.
-
Qualifying AI Projects with the PUSH Framework
More than 80% of AI projects fail, according to RAND, often because nobody checked whether they were worth doing. The PUSH framework for judging feasibility, uniqueness, scalability and ethics before you start.
Armakuni. A six-minute read.
-
Simplifying RAG Implementations on AWS: Step-by-Step Guide to Smarter AI Solutions
Generative AI works better when it can look things up. A step-by-step guide to building retrieval-augmented generation (RAG) on AWS.
Armakuni. A three-minute read.
Bring us the problem these posts are about.
A free workshop of a few hours with our architects, on your own systems. No cost.