AI Agents - Netics

Sign in Subscribe

AI Agents

A collection of 21 posts

Netics editorial card on invoice intake automation, with the checks between extracted fields and a posted entry, from supplier identity to a named approver above the threshold.

Automating Invoice Intake With Checks a Person Can Verify

Invoice intake automation reads a supplier invoice, extracts the fields and routes each one with rules you can inspect. The design question sits in the checks between extraction and posting:

Netics editorial card showing a website assistant answering from a catalogue and handing a qualified conversation to a person.

Designing an AI Concierge That Qualifies Visitors With Facts You Publish

A website concierge answers, asks and recommends from your own catalogue and rules, quotes prices instead of inventing them, and hands the conversation to a person with a summary that can be

Netics editorial card for the private RAG assistant pillar article, contrasting retrieval, citation and refusal.

How a Private RAG Assistant Answers From Your Own Documents

A private RAG assistant has three separate jobs: retrieve the right passage, attach the passage to the claim, and refuse when retrieval comes back empty. The refusal is an engineering output

Netics editorial card for the AI agent harness pillar, showing the loop around the model and the model layer kept replaceable.

What an AI Agent Harness Contains, and Why the Model Layer Stays Swappable

An agent harness is the loop around the model: context, tool calls, session state and permissions. The model inside it is the part that changes on somebody else's calendar, so the harness is

DeepSeek DSec elastic sandbox platform article card with a DeepSeek DSec text identity and Netics branding.

DeepSeek Just Published the Blueprint for Training Agents at Scale

DeepSeek's DSec report shows what agent training actually needs from infrastructure: an elastic sandbox factory with tiered isolation, on-demand image delivery, and RL co-design.

n8n Agents Change the Automation Contract: From Steps to Goals

n8n Agents shift the automation contract from laying out steps to describing an outcome. The session log becomes the audit trail; one turn becomes the budget.

Netics feature card for the OpenAI Agents API article with the official OpenAI identity

The OpenAI Agents API Makes the Harness a Managed Product

OpenAI's Agents API turns the Codex harness into a managed service with hosted sandboxes. Netics on why the interesting boundary is the sandbox, not the model.

Netics feature card for the Salesforce and Google Cloud partnership article with the official Salesforce and Google Cloud identities

Salesforce on Google Cloud: The Agent Stack Gets a Shared Foundation

Salesforce and Google Cloud are unifying infrastructure and agents: Hyperforce on Google Cloud, Gemini Enterprise with Salesforce data via MCP, and Zero Copy data sharing. Netics on what is

Gemini 3.8 Live voice-agent continuity concept with a Google Gemini identity and Netics editorial treatment.

Gemini 3.8 Live turns voice agents into continuity systems

TL;DR Google’s September 15 announcement introduces Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking as near-real-time reasoning models for more natural voice interaction. The announcement highlights interruptions, language switching, real-time visual context, background tools and API calls, and an Extended Thinking mode that can keep

Google ADK for Kotlin 1.0 official banner showing agent development across Kotlin and Android.

Google ADK for Kotlin 1.0 Makes Agents Portable — Not Operationally Identical

Google’s ADK for Kotlin 1.0 unifies server and Android agent development around Kotlin Multiplatform. Netics examines where that parity ends.

LaunchDarkly AI SDK analysis visual showing agent access boundaries and runtime evidence.

LaunchDarkly’s AI SDK makes the control plane the runtime boundary

LaunchDarkly’s new AI SDK simplifies provider routing, agent graphs, judges, metrics, and traces, but convenience can deepen platform lock-in.

Cloudflare internal AI engineering stack control plane with identity, gateway, context, and review layers

Cloudflare's internal AI stack is a lesson in control-plane design

Cloudflare's internal AI stack shows why identity, routing, context, and review must be designed as one control plane rather than assembled as disconnected tools.

Google Managed Agents hooks and budgets article card with cloud infrastructure and Netics branding

Google Managed Agents Turn Tool Hooks Into an Operating Control Plane

Gemini Managed Agents add hooks, budgets, model selection, and scheduled triggers. Netics examines what this changes for production governance.

Diagram comparing MCP tools, resources, and prompts as three distinct server-exposed primitives with Netics branding

"MCP Tools vs. Resources vs. Prompts: Choosing the Right Primitive for Enterprise Integrations"

MCP tools, resources, and prompts serve different jobs — this comparison helps enterprise teams pick the right primitive for each integration.

Illustration of bounded identity and wallet controls for agentic commerce.

Agentic Commerce Needs a Bounded-Authority Model

Cloudflare's new Wallets and cloudflare.pay give AI agents a stable identity and a spending cap. That solves payment mechanics — it doesn't settle who an agent is allowed to act for, or who

Netics visual showing a multilingual agent skill and the tests needed to verify its behavior.

Multilingual Agent Skills Need Evaluation, Not Just Translation

The non-English share of agent skills is rising. For teams, the issue is not only language; it is proving that an agent follows the procedure with the same precision.

AWS unified agent architecture thumbnail with a licensed AI-compute hardware photo background and readable Netics title.

Agent Retrieval Needs More Than One Table

AWS combines operational records and vector search in DynamoDB, but the architecture trades duplicated stores for stricter schema and event-pipeline discipline.

OpenAI Agents SDK concept showing a sandbox, manifest, credential boundary, recovery path, and visible OpenAI identity.

Agent Sandboxes: Durable State for Long-Running Work

OpenAI’s updated Agents SDK makes sandbox execution and durable state easier to compose, but production safety still depends on manifests, credential separation, approvals, and recovery evid

AI Studio vs. Gemini Enterprise decision matrix for prototype-to-production boundaries.

AI Studio vs. Gemini Enterprise Agent Platform: When to Move to Production

Google AI Studio or Agent Platform? A practical migration guide for credentials, users, spend, security, and capacity.

Dark Netics card showing an AI sales agent connected to tools, policy, state, and observability

AWS AgentFlo Shows What Production AI Agents Actually Need

AgentFlo's AWS architecture is less interesting as a chatbot story than as a permissions, state, and operations story. The useful pattern is the separation between agent reasoning, tools, po

Stateless MCP article card with the official MCP symbol and MCP name on a dark Netics card

What Stateless MCP Simplifies and Leaves Exposed

Cloudflare's stateless MCP revision removes mandatory sessions — but the trade-offs (MRTR, relocated state, headers as data) are real, and its enthusiasm is not neutral.