Attackers rarely break in.
Someone lets them in.
We are an independent security company working on the part of the problem technology alone has never solved: the few seconds before a person clicks, pays or replies to a message that only looks real.
Human risk should be measured, not assumed
Most security spending goes to the perimeter, and most successful attacks walk straight past it. The firewall was never asked the question that mattered, which is whether the person reading the message at nine in the morning will recognise it for what it is.
That question usually gets answered with an assumption. An organization runs a training module once a year, records who completed it, and treats the record as evidence. It is not evidence. It measures attendance, and it says nothing about how a particular team will respond to a particular attack on a particular Tuesday.
We think the human layer deserves what every other control already gets: deliberate testing, honest reporting, and improvement where the test finds a gap. That is the whole design of what we build.
Scams stopped looking like scams
The grammar is clean, the logos are right, the sender looks familiar, and the request is urgent enough that checking feels rude. Fake payment confirmations, job offers with an upfront fee, a message that appears to come from your bank: all engineered to bypass careful judgement rather than defeat a firewall.
We kept watching people we know lose money to attacks that were obvious in hindsight. Not because they were careless, but because nobody had a fast, honest way to ask is this real? without feeling foolish. That question is what CyberWatch AI answers.
Inside companies we saw the same thing at scale. Budgets went on software while the attack walked in through an inbox, and nobody could say how their own staff would respond until the day it mattered.
A free tool, a console, and the research behind both
CyberWatch AI Scan
A free checker for a suspicious link, message or screenshot. It reads the content itself rather than matching a list of known scams, so it works in any country and any language, including attacks that are days old.
The security console
Realistic phishing simulations, targeted training for the gaps those simulations expose, and reporting that tells management where the organization actually stands.
Published research
Several hundred articles on the attacks we see land on real people, written to be useful to someone who is not a security professional.
Test, train, then measure the difference
Test your people safely
Realistic phishing simulations show who opens, who clicks and who reports, without putting anyone at risk.
Train the gaps it finds
Short lessons go to the people and topics the tests point to, rather than a yearly slideshow for everyone.
Measure where you stand
One security score and plain reports show management what changed and what to do next.
Every member can also check anything suspicious before acting on it and report it to their security team in one step. The same checker is free for anyone to use.
Five positions we can be held to
These decide what we build, and what we refuse to build.
Measure, do not assume
An organization that has not tested how its people respond does not know how they will respond. Completion rates and policies are records of activity, not evidence of resilience. If we say your risk has fallen, there is a measurement behind it you can inspect.
Clear beats clever
A verdict a non-specialist cannot act on has failed, however good the analysis behind it. Advice that needs a security background to follow excludes most of the people who need it.
Say what you do not know
A confident wrong answer is more dangerous than an honest uncertain one, because it ends the thinking. Our results state what they could not determine.
Blame the process, not the person
Someone who clicks is usually responding sensibly to a well-built deception. Organizations that punish mistakes teach staff to hide them, which is how a containable incident becomes a serious one.
Earn the access we ask for
We see suspicious messages people have not yet shown anyone. That access is given on trust, so what we keep is deliberately small: a risk score and a short summary, never the content.
Built by people who answer the alerts
CyberWatch AI is led by its two co-founders, Gideon Egyabeng and Divine Egyabeng, brothers who work in security operations. The detection logic comes from attacks they respond to in their day jobs rather than from a bought threat feed.
What we promise about your data
- We do not store what you check. Links, messages and screenshots are analysed and discarded. We keep the risk score and a short summary, never the content, unless you choose to report it to your security team so they can act on it.
- Private checks stay private. In an organization, administrators never see a check you did not report.
- We do not sell your data. There is no advertising business here, and there never will be.
- We tell you when we are unsure. A confident wrong answer is worse than an honest uncertain one, so the verdict says what it could not determine.
- It supports your judgement, it does not replace it. Especially before sending money, treat a low-risk result as one input, not permission.
Ways to build this with us
We are independent and self-funded, and we intend to stay deliberate about who we build with. If any of the following describes you, we would like to hear from you.
Investment
We are open to conversations with investors who understand that human risk is a long game and that the product has to stay honest to be worth anything. Tell us how you think about that, not just the cheque.
Partnerships and resellers
If you already advise organizations on security, compliance or IT, the console can sit alongside what you do. We are interested in partners who will stand behind the results it produces.
Research and data
We publish on the attacks we see land on real people. If you run a CERT, a fraud team or a research group and want to compare notes or co-publish, get in touch.
Working here
There is no open vacancy list today. We would still rather hear from someone who has read what we publish and disagrees with part of it than miss them because nothing was advertised.
Write to info@cyberwatchai.com and say which of these it is. A short note about what you have in mind is more useful than a deck.
See it for your organization
A short demo walks you through the console. Or check something suspicious right now, no account needed.