Switchboard Connect

Every other week, I curate a few relevant pieces of AI & Automation news and insights for SMB operators & execs.

The last couple weeks have been a lot of travel (yay Main St Summit!) and supporting new teammates joining us at Switchboard but I’m now back to our regularly scheduled programming with the newsletter.

Let’s dive in.

There’s a lot of hype about AI Agents being the “next big phase” of AI being integrated into business operations.

Well, that “next big phase” isn’t coming, it’s here.

What to Know: Microsoft’s Copilot Agents are transforming workflows by automating everyday tasks directly in Microsoft 365 across different types of data and workflows.

How it works: Microsoft’s Copilot Studio lets teams build AI agents directly into Microsoft 365 workflows. These agents tackle repetitive tasks, process updates, and keep team communications synced—saving substantial time and effort.

In action: McKinsey’s demo of their autonomous agents, built with Microsoft Copilot Studio, shows a 90% reduction in time spent on project management tasks, like email coordination and updates. It’s a bold example of how AI can redefine daily operations.

Limitations: While powerful within Microsoft’s ecosystem, Copilot agents currently work best with Microsoft tools alone. This could be a challenge as most teams use a mix of software.

Learn more here → 

And for those not on a Microsoft stack, I’ve been checking out Lindy & Relay which are great, more broad options.

What it is: Claude’s Computer Use is a beta feature that allows Claude AI to directly interact with apps and data on a computer—think automating repetitive tasks or performing complex operations autonomously.

The big picture: The days of overpromised “digital transformation” are behind us. Now, operational AI is accessible to companies of all sizes, meaning more efficient teams, faster project turnaround, and ultimately, better results.

What to watch: Computer Use is early days is meant more for individuals to utilize as it isn’t meant for large teams to have automated workflows. I expect Anthropic’s competitors to release their versions of this in the next quarter or two.

Learn more here →

Recently, the Biden administration recently released a national security memo to keep the U.S. competitive against China in the AI race.

State of play: The new AI executive order mandates this national security memo, largely unclassified but with a classified annex addressing adversary threats.

Why it matters: While this may read as a bland memo meant for government agencies, it’s no coincidence that this memo came out 1 month after the CEO’s & leaders from OpenAI, Anthropic, NVIDIA, Google met with the White House to share where AI is headed.

What’s inside: The memo establishes the U.S. AI Safety Institute, guiding standards for international AI governance. Key directives include:

• Prioritizing the procurement of advanced AI systems for national security—cybersecurity, counterintelligence, military logistics

• Strengthening chip supply chain security and diversity

• Boosting the National AI Research Resource

• Tracking competitor operations in the U.S. AI sector

Framework for AI Governance: The National Security Council has also published a companion Framework for AI Governance that defines:

• Prohibited uses: AI for censorship, legal interference, or nuclear decisions without human oversight is banned.

• High-impact guidelines: Restrictions apply to AI uses that risk national security, international norms, and human rights.

🔗 Read the full memo here.

Big picture: People-heavy businesses (professional services, consultancies, etc.) are heading toward “Service-as-a-Software” where technology—not people—drives work.

The shift is about more than efficiency; it’s about breaking free from the costly cycle of legacy systems and bloated processes.

What is Service-as-a-Software?: Service-as-a-Software (SaaS) reimagines traditional services through automation and AI, reducing the need for human intervention. Instead of hiring more staff for repetitive tasks, companies use AI-driven platforms to automate and scale operations. By 2030, this approach is expected to transform service delivery into a streamlined, tech-powered model that emphasizes efficiency, scalability, and cost reduction over manual labor.

Why it matters: Companies are weighed down by “legacy debt”—outdated systems and processes that stifle agility. Enterprises need to adopt AI-led solutions to stay competitive, or they risk becoming obsolete. A massive $2 trillion in global tech debt underscores the urgency to change .

5 phases of transformation: The evolution from people-heavy to tech-led services includes:

1. Staff augmentation: Humans with minimal tech support.

2. Tech-enabled services: Using technology to aid human tasks.

3. Platform-led services: Platforms begin to automate routine tasks.

4. AI-led agentic services: AI takes on decision-making roles.

5. Service-as-a-Software: Fully autonomous, technology-driven services.

Challenge: Most people only understand jobs they can visualize (think dentists or mechanics). This has fueled a narrow view of AI “taking jobs,” but the reality is more nuanced. Like my cousin with a PhD in applied math, people thought he’d just be a high school teacher. But he’s actually training AI to help cure cancer—way beyond what people imagine.

Bottom line: Winning firms will be the ones that pay down their legacy debts by adopting AI-driven solutions, which means less focus on people doing tasks and more on technology automating work. This shift will require enterprises and service providers to partner closely to manage the transition .

🔗 Read HFS full thoughts here → 

That’s all for this week.

Seeing all this AI Automation news out there and not sure where to start? Feel free to reply with any questions you may have – happy to help!

Chat soon 👋