AI Scams Are Getting Better. Customer Support Has to Be Faster.
The U.S. Postal Inspection Service is warning that scammers are using AI to make old cons look new and believable. Voice clones. Polished screenshots. Profiles that feel real enough to pass a quick glance. With National Consumer Protection Week running March 1-7, 2026, this is a good moment to tighten your frontline playbook.
Learn more from the Postal Inspection Service's guidance here: National Consumer Protection Week campaign.
What changed: the con looks real
Scammers are using AI to generate photos, clone voices, and spin up convincing profiles. This supercharges romance and investment scams, fake tech support, cryptocurrency pitches, and even "emergency" messages that seem to come from friends or family.
Fast red flags your team can spot
- Thin or odd social footprint: New account, few friends, comments that feel scripted, or profile images that don't match the name or details.
- Mismatched domains: Slight misspellings, non-secure login pages, or emails that don't match the site (example: john.doe@USPSUS.com instead of USPS.com).
- Overly polished "evidence": PDFs, screenshots, or invoices that are too clean and easy to fabricate.
- Moving off-platform: Pressure to switch to WhatsApp, Telegram, or a private email to avoid platform protections.
- Audio/video inconsistencies: Off lip-sync, odd lighting, jerky motion, or a voice that doesn't quite match the person you know.
Frontline playbook for support reps
- Slow it down: Any "act now" request is a red flag. Calm the customer, confirm details, and pause the transaction if risk is high.
- Verify identity in-channel: Use on-file info and secure account prompts. Never accept "we switched numbers" as proof.
- Hold the line on payments: Refuse gift cards, crypto, cash, or wire as "verification." Those are standard scam asks.
- Refuse off-platform support: Keep communication inside approved channels with audit trails.
- Use call-back and passphrase checks: If someone claims to be a customer or internal staff, call back using verified numbers and require a pre-set passphrase.
- Escalate early: If voice clones, deepfakes, or mismatched domains are in play, tag the case and route to fraud/security.
- Document artifacts: Save headers, URLs, wallet addresses, and screenshots. They're useful for internal analysis and reports.
Quick phrases you can use
- "For your safety, we only process requests inside this support channel and after identity verification."
- "We don't accept gift cards, crypto, or wires. If someone is asking for that, it's almost always a scam."
- "Let me call you back using the verified number on file before we proceed."
If the issue mentions USPS or shipping
- Advise customers to check tracking and fees only on USPS.com and official USPS apps. Ignore texts or emails that demand urgent payment.
- Point customers to the Postal Inspection Service's prevention tips: USPS NCPW guidance.
Team safeguards to put in place this week
- Update macros and scripts: Add AI scam red flags and the phrases above.
- Tighten verification: Require 2FA and account prompts before sensitive changes.
- Block the usual exits: Disallow off-platform support and untraceable payments across policy, tooling, and training.
- Create an "AI scam" case type: Standard fields, routing, and evidence collection.
- Share a one-pager with customers: Red flags, how to verify, and where to report suspicious messages.
Where to report
- Report scams to the FTC: reportfraud.ftc.gov.
Keep your support team sharp
AI makes scams look convincing. Your process makes them harmless. Train your reps, script the responses, and remove guesswork.
National Consumer Protection Week is a useful deadline. Share this with your team, update your playbooks, and make "verify first" your default.
Your membership also unlocks:
