Honeycomb Launches Agent Observability, Bringing Full Visibility to Agentic Workflows in Production

Honeycomb.io, the observability platform built for the new shape of software, today introduced a series of agentic intelligence and agent observability features purpose-built for AI agents in production. These new and enhanced capabilities include Agent Timeline, Canvas Agent, and Canvas Skills. Engineering teams will now have real-time visibility into what their agents are actually doing without proprietary software development kits (SDKs) or framework lock-in.

New and enhanced features from Honeycomb supercharge human and agent teams by surfacing intelligence at machine speed. These include:

  • Canvas, newly rebuilt to serve as a collaborative workspace, a chat interface, and autonomous agent all in one, allows engineering teams to query issues in plain English, work with human and agent team members on investigations, and produce sharable visualization snapshots.
  • Auto-investigations set your Canvas agent to work automatically when an alert fires, an SLO burns, or an anomaly surfaces, gathering data, creating and testing hypotheses, and proposing remediation, all before an engineer even opens their laptop.
  • Canvas Skills encode your best engineers’ debugging knowledge and best practices for frameworks or services like Kafka into reusable playbooks that run autonomously.

Honeycomb has integrated the OpenTelemetry (OTel) GenAI semantic conventions (v1.40.0) into its platform. By making gen_ai.* attributes first-class citizens, Honeycomb ensures that model evaluations, tool executions, MCP calls, LLMs, and agents are all properly observed. This alignment with OpenTelemetry GenAI standards allows for the automatic generation of structured GenAI insights without requiring re-instrumentation as specifications change, custom setups, or the use of proprietary SDKs.

Canvas, Canvas Agent, and Skills are available starting next week for all Honeycomb customers. Agent Timeline is available in Early Access, and anticipated to be generally available next month.

Learn more at www.honeycomb.io.

Forward Launches Forward Predict to Unlock Autonomous Networking

Forward (formerly Forward Networks) today announced Forward Predict, which shows the impact of network changes before they are made. By running every proposed change against a mathematically accurate digital twin of the entire production network, Forward Predict ensures that costly errors never reach production, giving organizations the confidence to move faster, operate safely at any scale, and lay the foundation for autonomous networking.

“When we founded Forward more than a decade ago, we set our sights on the future of autonomous networking,” said David Erickson, CEO and co-founder of Forward. “Everything we have built since, including the world’s first network digital twin, was working backwards from that goal. An incredible team, advances in compute, and 12 years of deep collaboration with the world’s largest and most complex networks have made this moment possible. Forward Predict is the next critical step on that journey, and we built it for every organization that cares about their network.”

Forward Predict is powered by Forward’s Network Digital Twin, a complete, accurate model of behavior, including the state of every device from every vendor, from the network layer to the application layer. This model understands all possible ways the network can handle packets, identifies where policies are in conflict, and answers questions with mathematical certainty.

With Forward Predict, these same capabilities extend to future states of the network. Proposed changes are validated against a production equivalent network digital twin spanning every vendor and every cloud. Risks are discovered and resolved before anything touches the live network. The platform verifies the impact of a change through testing and delivers deterministic evidence of the outcome. 

The ability to predict enables AI agents to safely drive changes. Forward Predict delivers the pre-verification that makes autonomous networking real. Every AI-proposed change is validated against a mathematically accurate digital twin of the entire production network before it executes. Security, connectivity, and compliance risks are identified in advance, and when a proposed change creates an unintended consequence, the agent receives specific failure feedback and can iterate until a fully verified change is found. This closed-loop verification enables AI to have the same transformative impact on network change that it has had on software development.

Forward Predict will be available in fall of 2026. Learn more at www.forwardnetworks.com/predict

Automation Anywhere Launches EnterpriseClaw

Automation Anywhere today announced EnterpriseClaw, a new capability designed to make claw-style AI agents effective across enterprise operations. Developed in collaboration with Cisco, NVIDIA, Okta, and OpenAI, EnterpriseClaw enables organizations to deploy autonomous AI agents across cloud platforms, desktops, on-premises systems, and secured enterprise networks — enabling AI-driven work to run securely, accurately, and under centralized control across large-scale enterprise operations.

Organizations are increasingly experimenting with claw-style AI agents that can execute tasks directly inside applications, browsers, terminals, and local systems. That makes them powerful for executing tasks where work happens, but many are designed primarily for individual users operating on a single machine or within isolated cloud environments.

This new class of claw-style AI agents can run across enterprise teams and workflows — including cloud platforms and behind-the-firewall infrastructure — while maintaining centralized control over access, activity, governance, and observability. EnterpriseClaw agents also tap into the power of Automation Anywhere’s Process Reasoning Engine (PRE) and Contextual Intelligence Graph to give them the enhanced accuracy and process context needed to automate business critical work more reliably than any LLM on its own.

For example, organizations can investigate complex customer claims by gathering information across desktop applications, on-premises systems, internal documents, and cloud platforms — while keeping sensitive financial, healthcare, or operational data inside secured enterprise systems.

This capability is foundational to operating as an Autonomous Enterprise — where AI runs enterprise work across systems, not just within isolated tools.

“For AI to have a transformational impact on business, it needs to be able to do work where the work actually happens,” said Mihir Shukla, CEO and Board Chairman of Automation Anywhere. “Many claw-style AI agents are incredibly powerful, but most were designed for isolated environments or individual users — while enterprise operations span teams, cloud platforms, desktops, on-premises systems, and highly regulated infrastructure. EnterpriseClaw is designed to unlock the potential of these next-generation AI agents for real enterprise operations. We’re excited to collaborate with Cisco, NVIDIA, Okta, and OpenAI, whose technologies help support the security, identity, intelligence, and infrastructure requirements of enterprise-grade agentic systems.”

Strategic Collaborations Enable Enterprise-Ready Claw-Style AI Agents

Powered by Automation Anywhere’s hybrid cloud-native deployment architecture, EnterpriseClaw integrates with leading enterprise technology providers to deliver production-ready claw-style AI agent deployment:

  • Cisco AI Defense and DefenseClaw provide comprehensive security purpose-built for agents.
  • NVIDIA contributes OpenShell, an open-source runtime to build and deploy autonomous, self-evolving agents more safely. In addition, NVIDIA NIM microservices with the latest NVIDIA Nemotron open models power EnterpriseClaw AI agents for on-premises customers.
  • Okta delivers cross-agent identity management and authentication controls for policy enforcement.
  • OpenAI will enable companies to build and operate agents using leading OpenAI models like GPT-5.5. These agents will support enterprise-grade workflows within EnterpriseClaw.

EnterpriseClaw is designed to be extensible across a range of AI agent frameworks, allowing organizations to deploy internally developed agents or those built using third-party tools, and manage them alongside existing automations.

EnterpriseClaw is available in preview, with general availability expected later this year. Learn more at automationanywhere.com.