TLDR IT 2026-05-06

Build a custom AI agent that handles them so you don't have to. This on-demand webinar shows you how to create an agent that answers IT questions directly in Slack.

Watch it to quickly build an agent that:

✅ Reads your own documentation to resolve queries independently

✅ Automatically escalates complex cases to your IT team

✅ Logs Q&A on its own

Get instant answers from your knowledge base, directly in Slack.

Access the session

ServiceNow expanded its AI Control Tower to help enterprises discover, govern, secure, and monitor AI agents across both its own platform and third-party environments. The update adds integrations across Amazon Web Services, Google Cloud, Microsoft Azure, plus apps like SAP, Oracle Corporation, and Workday, along with a kill switch that can shut down compromised agents during attacks.

Hackers stole student names, email addresses, ID numbers, and private messages during a breach at Instructure, the company behind Canvas. The company says passwords and financial data weren't exposed, but the incident is another reminder that non-core platforms like LMS systems often hold far more sensitive data than companies realize.

Anthropic is launching a $1.5 billion venture backed by Goldman Sachs, Blackstone, and Hellman & Friedman to acquire AI services firms and embed engineers directly into customer environments. The move mirrors OpenAI's own deployment push and highlights a growing reality: enterprises don't need more AI demos - they need people who can actually integrate AI into messy workflows, data systems, and operations.

Linear is rapidly closing the gap with Jira, with new customer acquisitions rising 67% year-over-year while Jira's dropped 32%. Linear dominates the 11-200 employee segment and AI-native stacks, whereas Jira maintains a stronghold in regulated industries and enterprise environments. Monday.com shows a significant 41% decline in new-user acquisition.

Not all AI revenue is created equal. While OpenAI may lead in headline revenue, much of it comes from consumer subscriptions that can churn quickly, while Anthropic is seeing more enterprise-driven revenue tied to deeply embedded workflows. AI vendors integrated into core business systems may prove far stickier than consumer-first players when the market inevitably cools.

Microsoft is phasing out legacy printer drivers in Windows 11, introducing Windows Protected Print Mode to enhance security. This shift mandates a modern, driverless printing architecture, reducing reliance on third-party software and minimizing potential attack vectors for enterprise and consumer systems alike.

Airbyte Agents is a new service that pre-indexes enterprise data into a managed Context Store so agents can search business data without constantly hitting live APIs. The company says that it can reduce tool calls by roughly 40% and token usage by up to 80% by letting agents search indexed customer, billing, support, and operational data before making live requests.

SAP is acquiring Prior Labs to advance tabular foundation models, committing over €1 billion over four years. The independent research lab will focus on structured business data, aiming to improve predictive accuracy for enterprise outcomes like churn and risk, while maintaining support for the open-source TabPFN ecosystem.

AI tools accelerate IoT development but introduce systemic technical debt by ignoring hardware constraints and architectural boundaries. Research shows 15% of AI-generated commits contain quality issues and code duplication has risen to 12.3%. Mitigate risks through mandatory human reviews, strict architectural governance, and regular system refactoring.

Supply chain attacks like the Trivy-action and Axios compromises highlight the need for robust package security. Organizations should implement cooldown periods, disable install scripts, utilize lockfiles, and deploy honeytokens. These practical, scalable controls reduce exposure windows and limit the blast radius of malicious third-party dependencies in production.

Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs are launching a new company to deliver customized Claude-powered AI solutions to mid-sized businesses lacking in-house capabilities.

Google now lets Workspace admins require explicit participant consent before Gemini note-taking, recordings, or transcripts begin in Google Meet meetings.