“The value of an idea lies in the using of it.”

—Thomas Edison, Inventor

A new alliance between Google, Amazon, Microsoft, Meta, OpenAI, and several major retailers signals something important: fraud has reached a scale that no single company can manage alone.

The companies signed the Industry Accord Against Online Scams & Fraud, committing to share threat intelligence, coordinate investigations, and deploy new detection systems across platforms.

That coordination matters because fraud has exploded. Americans lost more than $16B to scams in 2024, and 71% of U.S. companies report an increase in AI-powered fraud attacks increase over the past year.

Generative AI allows scammers to produce convincing messages, mimic internal communication styles, and impersonate vendors or executives across email, messaging apps, and websites.

Business email compromise remains the most common entry point, affecting 62% of organizations. Fake websites, text scams, and social impersonation continue to expand.

But corporate defenses have not evolved at the same pace.

Nearly 48% of organizations still rely on manual verification methods such as callbacks or email confirmation when validating information.

The result is a widening gap between the scale of attacks and the speed of detection.

Fraud schemes often succeed because of gaps in vendor data management.

Many organizations report confidence in the accuracy of vendor banking information, yet only 32% validate that data continuously or in real time.

Validation usually occurs at specific checkpoints—during onboarding, before payment, or after an account change request. Long stretches of the process remain unchecked.

Fraudsters exploit those windows through altered invoices, supplier impersonation, and payment-detail changes that appear legitimate.

1️⃣ Verify Every Payment Change
Never update vendor banking details from an email alone, and always call the vendor using a phone number already on file to verify changes.

2️⃣ Treat Urgent or Vague Requests With Suspicion
Any pressing requests involving money should get a pause and a secondary verification. Ask specific questions tied to the deal: invoice number, VIN number, previous payment history, etc. Scammers rarely have those details.

3️⃣ Lock Down Employee Email
Require multi-factor authentication for all dealership email accounts. Most fraud starts with a compromised inbox.

4️⃣ Separate Financial Responsibilities
Divide duties so one person doesn’t control the entire payment process. Entry, approval, and release should involve different people.

5️⃣ Use Technology to Flag Suspicious Activity
Email filtering, account verification tools, and CRM alerts can help detect fraud patterns faster than manual review.

6️⃣ Run Fraud Drills
Test your team quarterly with a simulated phishing or vendor-payment scam. Practice builds awareness.

Claude Cowork is Anthropic’s attempt to move AI beyond a chat window and into actual day-to-day work. Instead of asking an assistant how to do something and then completing the steps yourself, Cowork can carry out the task directly once it has access to the relevant files or folders.

Inside the Claude desktop app, Cowork works alongside Chat and Code modes. You describe the task, grant permissions and access where necessary, and Claude generates a step-by-step plan. You can approve, adjust, or cancel the plan before anything happens.

The tool is designed for common workplace tasks such as organizing messy folders, analyzing spreadsheets, compiling research from multiple documents, and generating formatted reports from raw data. It supports standard office formats like DOCX, XLSX, PPTX, and PDF, which makes it practical for everyday work.

One feature that stands out is automation. Cowork can schedule tasks to run on a daily, weekly, or monthly basis, which opens the door for things like automatically generating weekly reports, summarizing project files, or keeping download folders organized without manual cleanup.

Claude Cowork is currently available in research preview for paid Claude plans on Windows and macOS. Even in its early form, it reflects a larger shift happening across AI: assistants are starting to move from answering questions to actively helping complete work.

With scammers increasingly targeting businesses, this prompt asks AI to pressure-test the way your dealership actually operates.

AI adoption in dealerships has moved past the curiosity phase. The real conversation now is about implementation, data quality, and whether these tools actually improve operations.

In this mini-webinar, our very own Kyle Mountsier joins Greg Uland and AJ McGowan (from Reynolds and Reynolds) to unpack what they’re seeing inside real dealership environments. The discussion gets into where AI is delivering measurable value today, where it’s still mostly noise, and why data integrity and system integration are becoming the real make-or-break factors.

If you’re evaluating AI vendors or trying to make your current stack actually work together, this is a masterclass worth the watching.

😵‍💫 Local residents of Southhaven, Mississippi say the noise generated by Elon Musk's make-shift power plant is keeping them up day and night.

🏀 57% of fans admit to using AI to build their March Madness™️ bracket this year.

🐶 A tech professional in Australia used ChatGPT and AlphaFold to develop a successful cancer treatment for his pet pup, Rosie.

🚨 A new study finds that 8 out of 10 chatbots will actively assist users in planning violent attacks.

Thanks for reading, Friend! Remember: AI is only as good as the way you implement it.