AI

How it works

• Tools such as Copilot and chatbots are trained on large amounts of text or images, then generate a likely answer to what you ask.
• Inside office tools they draft emails, summarize meetings and answer questions about your files.
• Security tools use it to spot phishing and odd behaviour. Attackers use it to write convincing lures and clone voices.
• A short policy settles which tools are approved, what client data may go in, and who checks the output.

What can go wrong (example scenario)

A three-person HR consultancy pasted a client's employee list into a free chatbot to “clean it up”. That data left the firm's control, in breach of their contract and of PIPEDA's safeguards duty.

How WILDCARD handles it

Cloud covers workspace administration: which AI features and apps are switched on in your Microsoft 365 or Google tenant, and who can use them. An AI use policy is a vCIO output. Email Security and Workstation Defender already use AI to protect your side.

• A dental office enables Copilot for the office manager and sets a rule that patient data stays inside the tenant.
• A law firm blocks consumer chatbots on firm devices and approves one enterprise tool.
• A retailer uses AI to draft product descriptions and has a human check every price.

Example scenario: A small accounting firm's owner received a voicemail that sounded exactly like her business partner, asking for an urgent transfer. It was an AI-generated clone from a podcast interview. She called him back on his known number before doing anything. The transfer never happened.

A question owners ask

Should my small business be using AI?

Probably, for drafting and summarizing, with a one-page rule about what data goes in. Start with the tools already inside your Microsoft or Google subscription.