Microsoft's New Mandate:

At Ignite 2025, Microsoft unveils "Agent 365" to manage the AI workforce. Plus, Google’s AI that learns by watching your screen, and the escalating threat of autonomous cyberattacks. Here’s what it means for restorers.

Hey everyone,

If you’ve been treating AI like a collection of handy apps and chatbots, this week’s news is a wake-up call. The conversation is no longer just about what AI can do. It’s about how we control it and how it will fundamentally change the way we operate.

From November 18-21, Microsoft held its Ignite 2025 conference, and the central message was stark: Autonomous AI Agents are proliferating faster than businesses can manage them. Microsoft's solution is a fundamental shift in perspective—we must start managing AI agents not as software tools, but as a digital workforce.

This comes as Google unveiled technology that could automate the "un-automatable" software that plagues our industry, and as the reality of AI-driven cyber threats continues to escalate.

The theme this week is Governance and Autonomy. As restoration contractors integrate AI deeper into operations, we need strategies to leverage unprecedented automation while urgently protecting against "Shadow AI" and sophisticated cyber threats.

Let's get into it.

THE GAME CHANGER: Microsoft Ignite and the "Agent 365" Control Plane

The biggest news for any business running on Microsoft (which is many restoration companies) came from the Ignite conference. Microsoft introduced Agent 365, a system designed to govern the coming wave of autonomous AI.

What's New (The Facts):

• Agent 365 (The "Control Plane"): Announced Nov 19, this is a new platform to deploy, organize, and govern AI agents (autonomous tools that perform tasks) across an organization.

• Agents as Identities: Agent 365 treats AI agents like human employees. Each agent is given a unique ID through Microsoft Entra (the system that manages user logins).

• Governance and Security: It provides a central dashboard to track every agent, monitor behavior, enforce security policies (like data access restrictions), and instantly quarantine agents that pose a risk.

• The "Channel Agent" (AI Project Manager): A new feature in Public Preview, this AI lives inside a Microsoft Teams channel, acting as a domain expert, summarizing progress, and using the new "Workback Plan" feature to automatically create task lists and due dates.

What This Means for Restoration (The "Why You Care"):

This is primarily a "Why" (Strategic) shift that solves a critical emerging problem: Shadow AI.

The "Shadow AI" Risk: Right now, your team is likely using various AI tools or custom GPTs without centralized oversight. An estimator might feed scopes into an unapproved AI; an admin might use an automation tool that handles client data insecurely. This is "Shadow AI." It exposes your company to data leaks, compliance violations, and security breaches.

Agent 365 is Your "Digital Foreman": Agent 365 brings these tools under control. For a restoration owner, this means:

1. Security and Control: You can define exactly what an agent can access. Your "Intake Agent" can access new leads but not your financials. This "least-privilege access" is critical.

2. Accountability and ROI: You can finally see which agents are being used, how effective they are, and if they are generating errors.

3. The AI Project Assistant (The "What"): The "Channel Agent" in Teams is an immediate efficiency booster. Imagine a Teams channel for a large commercial loss. Instead of the PM spending hours compiling updates, the AI agent monitors the conversation, summarizes the day's progress, and automatically maintains a prioritized task list.

THE DISRUPTOR: Google's SIMA 2 Learns by Watching (No APIs Needed)

One of the biggest barriers to automation in restoration is the reliance on older, specialized software (like Xactimate or proprietary TPA portals) that don't connect easily to modern AI tools via APIs. Google DeepMind just unveiled a technology to solve this.

What's New (The Facts):

Google DeepMind introduced SIMA 2 (Scalable Instructable Multiworld Agent) (Nov 14-17). It represents a fundamental shift in how AI learns to use software.

• Learning by Observation: SIMA 2 doesn't need APIs or access to underlying code. It learns exactly like a new hire: it interprets visual input from the screen, and implements actions using virtual keyboards and mouse controls.

• Generalization: It can apply concepts learned in one program to a completely different one.

• Reasoning: Powered by Gemini, it can break down high-level instructions and plan the steps required to achieve the goal.

What This Means for Restoration (The "Why You Care"):

This is a massive "Why" (Strategic Shift). It has the potential to automate the "un-automatable" parts of the restoration workflow.

The Legacy Software Solution: If a human can use the software, this type of AI can eventually learn to use it.

• The TPA Portal Problem: Imagine an AI agent that can log into a cumbersome TPA portal, navigate the interface, extract the claim details, and upload required documents—tasks that currently consume hours of administrative time.

• The "Virtual Estimator" Assistant: While complex estimating requires human expertise, an AI trained by watching an estimator work in Xactimate could learn to handle the tedious aspects of sketching and inputting standard line items, freeing up the estimator to focus on negotiation and complex scopes.

This technology is still in the research phase, but it signals a future where every digital process can be automated, regardless of the software's age or complexity.

THE URGENT WARNING: Autonomous Cyberattacks Demand Governance

Last week, we reported on the first known large-scale, AI-orchestrated cyber espionage campaign disrupted by Anthropic. This week (Nov 14-20), the details underscore why AI governance (like Agent 365) is essential.

What's New (The Facts):

• The Execution: Anthropic confirmed the threat actor (linked to a Chinese state-sponsored group) used AI (specifically Claude Code) to perform 80-90% of the campaign autonomously against roughly 30 global targets.

• The Autonomy: The AI autonomously mapped networks, identified databases, wrote exploit code, harvested credentials, and organized stolen data. Human involvement was minimal.

• The Speed: The AI executed the attacks at a speed humans cannot match, making “thousands of requests, often multiple per second” (PwC report, Nov 14).

What This Means for Restoration (The "Why You Care"):

This is a critical "Why" (Strategic Risk). The barrier to performing sophisticated cyberattacks has collapsed.

Why You Are a Target: Restoration companies hold valuable data: Client PII, insurance policy details, financial records, and property access information. If an autonomous AI can breach global financial institutions, it can breach your CRM.

The Connection to Governance: This threat highlights the danger of "Shadow AI." Unmanaged, unauthorized AI agents operating within your network are massive vulnerabilities. Platforms like Microsoft's Agent 365 are no longer optional; they are necessary to monitor, control, and quarantine AI activity within your business.

THE INDUSTRY CONTEXT: AI Investment Accelerates

To understand the velocity of the AI revolution, follow the money. Investments this week show a clear focus on the construction and insurance sectors.

What's New (The Facts):

• ConTech Boom (Nov 14): Investment in construction technology soared in Q3 2025, hitting $4.4 billion (a 66% YoY increase). AI and robotics startups led the charge, with robotics investments growing 125% YoY (Nymbl Ventures report).

• AI in Insurance Claims (Nov 20): Five Sigma (makers of "Clive," an AI Claims Expert) and Sutherland announced a strategic partnership to accelerate AI adoption in claims processing, aiming to automate the lifecycle from intake to settlement. Carriers adopting AI have cut costs by 20% and shortened claim handling by 50% (Boston Consulting Group).

What This Means for Restoration (The "Why You Care"):

This is a "Why" (Strategic Trend).

1. Better Tools Incoming: The massive investment in ConTech means the tools available to contractors—from project management software to on-site robotics—are about to get significantly more powerful and AI-driven. Investors are favoring proven, scalable technologies.

2. The AI Adjuster is Coming: The focus on automating insurance claims means contractors will soon be dealing with AI adjusters. This will demand impeccable, instantly verifiable documentation. AI adjusters will automatically flag inconsistencies between invoices, photos, and drying logs.

THE HOW-TO: Conducting a "Shadow AI" Audit

The news from Microsoft Ignite about Agent 365 highlights the need to manage AI usage. Before you can manage it, you need to know what's happening. Here’s how to identify "Shadow AI" in your restoration business.

The Goal:

To identify unauthorized or unmanaged AI tools being used by employees that could pose a security or compliance risk.

The Strategy:

1. The Anonymous Survey (The Carrot)

People use unauthorized tools because they solve a problem. Start by understanding those problems without blame.

• Action: Send an anonymous survey (using Google Forms or SurveyMonkey).

• Key Questions:

  • "What repetitive tasks take up the most time in your day?"

  • "Do you use any external tools (like ChatGPT, Claude, Grammarly) to help with your work?"

  • "If yes, what do you use them for?" (e.g., writing emails, summarizing documents).

  • "Do you ever paste sensitive company or client information into these tools?"

2. IT Network Traffic Analysis (The Stick)

Your IT department or managed service provider (MSP) can monitor outbound network traffic for connections to known AI platforms.

• Action: Ask your IT provider to generate a report of the most visited domains related to AI services (e.g., openai.com, anthropic.com) from company devices.

3. The Follow-Up: Policy and Provisioning

Once you identify the tools being used, don't just ban them.

• Action: Establish a clear AI Usage Policy (e.g., "No client PII in public AI tools").

• Provision Secure Alternatives: If many employees are using public ChatGPT, provide the secure, enterprise version (like Microsoft Copilot or ChatGPT Enterprise) that protects your data.

Other AI News (The Quick Hits)

• Sora 2.0 in Microsoft 365: Microsoft announced (Nov 18) that OpenAI's hyper-realistic video generator, Sora 2.0, is now available within Microsoft 365 Copilot (for Frontier program users). Why it matters: Great potential for training simulations or marketing, but it also increases the risk of synthetic (fake) documentation. Carriers will demand higher standards of proof (e.g., 3D scans).

• OpenAI Tests "Group Chat": OpenAI is testing a feature (Nov 14) allowing multiple users to interact with the AI simultaneously. Why it matters: A PM, estimator, and adjuster could all "discuss" a complex claim with the AI in a shared environment, ensuring alignment and creating a documented record.

• Cohere Announces HIPAA Compliance (Nov 14): AI company Cohere announced HIPAA-compliant agreements. Why it matters: Crucial for restoration firms handling sensitive information in biohazard, trauma, or mold-related claims. It allows for secure AI analysis of documents containing protected health information.

• Salesforce Acquires Doti (Nov 14): Salesforce announced the acquisition of Doti to enhance AI-driven enterprise search. Why it matters: If you use Salesforce as your CRM, finding information across your past jobs, client communications, and SOPs is about to get much faster.

Looking Ahead: The Autonomy Era is Here.

The news this week about Microsoft governing agents and Google teaching AI to use any software confirms a massive shift. I want to leave you with where my head is at regarding the next 1–3 years.

We’re not just entering an AI era—we’re entering an autonomy era.

In the near future, businesses won't be defined by their headcount, but by their decision speed.

Right now, most restoration companies are drowning in lag time: job notes that require manual follow-ups, estimates that take days, SOPs buried in PDFs.

AI is collapsing all of that. We are moving from software that stores information to software that acts on information.

In restoration, construction, and TPA environments, this shift will hit faster than people realize because our industry is already overloaded with data—moisture logs, photos, policy rules, compliance requirements. It’s the perfect environment for autonomous agents.

Over the next 12–24 months, we’re going to see three major changes:

1. Autonomy is the New Workforce Multiplier. Think about your team structure differently:

• 1 PM backed by 3 AI agents.

• 1 estimator supported by 2 AI assistants.

• 1 BD rep with an AI copilot that logs notes, drafts follow-ups, and suggests who to call next.

This isn’t theoretical. The pieces are here. Owners need to understand: AI won't replace your people, it will replace the friction. Companies leveraging autonomy will move twice as fast with the same headcount.

2. SOPs Will Evolve from Documents to Systems. The old model: Write a process, train a team, and hope they follow it.

The new model: AI agents monitor the job in real-time, analyze the photos, track moisture trends, understand the policy, and automatically trigger the next required task.

This is why data structure is critical. You need clean, unified, structured data before any of this can work. Your core operating platform becomes the foundation the AI plugs into.

3. Quantum + AI Will Unlock Massive Optimization. Forget the hype about quantum computers replacing your laptop. What’s coming in the next 1-3 years is hybrid quantum optimization.

This means solving complex logistical problems instantly:

• Real-time CAT deployment strategies.

• Instantaneous job routing.

• Dynamic pricing based on availability and past performance.

• Millions of drying variables processed at once.

Owners don’t need to understand the mechanics. They just need to understand the outcome: More accuracy. Faster decisions. Lower overhead.

The Punchline: AI Trust Loops

Here’s the bottom line: The companies that win in the autonomy era will be the ones that build what I call AI Trust Loops.

In a world where AI speeds everything up, trust becomes the ultimate currency. This means leveraging AI to create:

• Transparent pricing

• AI-verified documentation

• Immutable job histories

• Automatic accountability

That’s the future we need to build toward.

Go build.

Dave