AI safety training for every team that uses AI

Your team uses AI every day. Prove they use it safely.

Modern, story-driven drills built from real AI incidents - not another canned deck. Mapped to each role, re-ranked with the week's incidents, adaptive to every learner. Certificates anyone can verify - examiner included. Backed by learning and behavioral science →

Get on the list Try the free demo
12interactive modules
100%real incidents
Flatper company
25 minper module
Try 60 seconds of the course - no signup
Casey - course assistantOnline now. Modules, cases, pricing. Real person on request.
Early access

Doors open Monday, September 14.

Get on the list - launch-day access, the free Module 1, and the first Brief, straight to your inbox. Founding teams get in first.

Free. Double opt-in - confirm from your inbox. No spam, ever.
-days
-hrs
-min
-sec
How it works

Live in your firm this week.

No sales call, no IT project.

01

Pick your plan

Tiered, flat per company. Never per seat.

02

Send one link

Your team trains on any phone or laptop.

03

Export the evidence

Scores and decision paths per employee, ready for the exam. LMS-ready exports for the training system you already run.

The Monday Incident Brief

One real AI incident, every Monday.

What happened, what it means for your policy, in a 3-minute read. Same watch that feeds the course.

Free. Double opt-in - confirm from your inbox. Unsubscribe anytime.
This week in the Brief
  • Deepfake payroll reroute - a 9-second voice clone, $183K gone
  • Copilot oversharing - what your permissions already expose
  • Examiner asks for AI training logs - what "evidence" means
The course

A CFO wires $25M to a deepfake. Samsung engineers hand ChatGPT secret code. Two lawyers file six cases that never existed.

Every module starts with a real, documented incident - then puts your team inside the moment the decision got made. Story-driven and cinematic, built on learning science. Module 1 is free to play today. All 12 unlock September 14 - then new modules every month, a new incident in the Brief every Monday.

Read the real cases: the $25M deepfake CFO - Samsung's ChatGPT leak - the fake case-law filing.

MODULE 1

The Leak

Samsung engineers paste secret code into ChatGPT.

Free demo →
MODULE 2

Safe Prompting

A dealership bot agrees to sell a Tahoe for $1.

Unlocks Sep 14 →
MODULE 3

The Hallucination

Two lawyers file six cases that never existed.

Unlocks Sep 14 →
MODULE 4

The Bias

AI screening discriminates. The employer owns it.

Unlocks Sep 14 →
MODULE 5

The Deepfake

A CFO on a video call authorizes $25M. Everyone was fake.

Unlocks Sep 14 →
MODULE 6

Shadow AI

The tools your team uses that IT never approved.

Unlocks Sep 14 →
MODULE 7

The Con

AI-crafted phishing and BEC. $2.9B a year.

Unlocks Sep 14 →
MODULE 8

The Agent

AI agents with real permissions, at machine speed.

Unlocks Sep 14 →
MODULE 9

The Record

AI notes and summaries are records. SEC/FINRA agree.

Unlocks Sep 14 →
MODULE 10

The Vendor

Your vendor's silent AI feature is your exposure.

Unlocks Sep 14 →
MODULE 11

The Disclosure

AI washing: the SEC's $400K word lesson.

Unlocks Sep 14 →
MODULE 12

The Incident

Capstone: run the playbook against the clock.

Unlocks Sep 14 →
A live decision point inside a module: the learner chooses whether to wire the funds or verify first
A decision point inside the course - the learner makes the call.
A hands-on lab screen from the coursework
A hands-on lab - practice, not slides.
A verifiable completion certificate with score and unique CERT ID
The certificate every learner earns - score, date, unique CERT ID.
Every module ends in a verifiable certificate.Branded seal, score, date, and a unique CERT ID per employee - backed by the admin evidence file. Auditors get proof, not promises.
For the CCO

The report your examiner asks for.

Per-employee completion, per-module scores, decision history, timestamps, certificate IDs - plus a team certificate when everyone finishes. This is the audit-ready export every customer gets.

View a sample admin report
Why it works - backed by science

A short drill at the right moment beats the annual slide deck.

The annual deck gets clicked through and forgotten within weeks. Ten minutes inside a real incident builds the reflex instead. The learning science behind every choice →

01

Decision points, not slides

Every module puts the learner inside the incident at the moment of choice - wire it or verify, paste it or don't. Produced like a modern series, not a compliance PDF. People remember decisions they made, not slides they skimmed.

02

Muscle memory

Sprint drills repeat the move - verify out-of-band, check the source, pause the paste - under time pressure, until the safe action is the automatic one.

03

Skeptic and control scores

Learners are scored on both halves of the job: catching what AI gets wrong, and controlling what it touches. Weak spots surface per person, per module.

04

Built to the evidence standard examiners ask for

Completion, scores, decision history, timestamps, certificate IDs - exported per employee (CSV and PDF today; LMS-ready SCORM/xAPI on the roadmap). "Prove your people were trained" gets a file, not a promise.

05

Fresh every week

New incidents become new modules every month - and the Monday Brief covers the attack that made the news that week. Your team trains on the attack that made the news last Monday, not a deck written two years ago. And every learner makes it smarter.

06

One sitting, done

One incident, one skill, one certificate - about 25 minutes start to finish. Short enough to finish between meetings; memorable enough to matter in one.

07

Your role, your track

A marketing analyst and a finance lead face different AI risks. The curriculum maps to each role - marketing, finance, engineering, HR, leadership. Different modules, not one-size-fits-all.

Sid Yenamandra, founder of AI Safety 101
The founder

Built by an operator.

Sid Yenamandra has spent his career in cybersecurity and compliance - founding Entreda (acquired by Smarsh), and a decade in the SEC/FINRA exam cycle. A parallel entrepreneur, he is also Founding Managing Partner of Surge Ventures and Founder & CEO of SurgeONE.ai, which automates Regulatory Ops for financial firms - compliance, cybersecurity, and data governance workflows.

AI Safety 101 turns real incidents into 10-minute drills with verifiable, examiner-ready proof - new modules every month, grounded in peer-reviewed learning science.
An engineer first - Berkeley EE/CS, Silicon Valley - his career runs from designing chips to governing the AI workflows firms now depend on. More about the founder →

Pricing

Simple pricing. Flat per company.

Short drills your team actually finishes - not an annual deck nobody remembers.

Never per seat - AI risk comes from the one person you didn't train. Join the list - founding teams go live first on September 14. Role-mapped tracks for marketing, finance, engineering, HR, and leadership. Enterprise by conversation.

Every completion earns a verifiable per-employee certificate - seal, score, unique CERT ID - plus an evidence file built to the standard examiners ask for.

Every mechanic backed by peer-reviewed learning and behavioral science →

Flat per company. Never per seat.
Every firm is different - headcount, roles, modules, branding. We fit the plan to your firm and quote it in one short conversation.
Talk to us
Founding teams go first.
Join the early-access list - founding firms go live first on September 14.
Join early access

Prefer invoice? Talk to us - ACH or check.

Add-on services

Around the platform, when you want hands on.

Delivered with the course, scoped per engagement - never priced off a rate card. Talk to us and we shape it to your team.

Tabletop simulation

Your leadership inside a live AI incident: the deepfaked CFO, the leaked repo, the regulator's letter. Facilitated virtual session, half-day, up to 15 people.

Executive training

90 minutes that changes how the top floor talks about AI - what it can touch, what it must never, and who is accountable. Leadership-only session.

AI red-team assessment

Your AI systems probed the way an attacker would - prompt injection, data exfiltration, agent misuse. Delivered with a specialist partner, scoped per client.

Custom module build

Your policies, your tooling, your incident types - built into a scored module your team trains on, with the same evidence file.

Talk to us about add-ons
Questions

What careful buyers ask

Is this SOC 2 certified?

No course can be - SOC 2 has no AI training module. You get what examiners ask for: documented evidence your people trained and scored on real scenarios.

Are the cases real?

All of them, with sources shown inside each module. Samsung, the $1 Tahoe, Mata v. Avianca, the $25M deepfake call.

Does this satisfy our regulatory training requirement?

No vendor can satisfy it for you - the duty sits with your firm. What this course produces is the documented evidence your policy requires: completion records, scores, decision history, and certificates anyone can verify at /verify. That file is what an examiner asks for.

How does it stay current?

New modules every month, built from the latest documented AI incidents - the library you buy today keeps growing, and every new module lands in your plan. Plus a new incident in the Brief every Monday.

The next AI incident in your industry is already on someone's timeline.

Make sure your team trained on the last one - with training they'll actually finish.

Get on the list

Tell us about your team

We come back within one business day with times to talk.

Request times
One business day. A real person, not a sequence.

Watch a real module free

Work email and company - the full module unlocks instantly. One follow-up, no spam.

Unlock the module
One follow-up, no spam.