Frequently Asked Questions

Emails classified by VerifiTrust as "Phishing" are highly likely to be malicious (scams), and should not be acted on or replied to, but instead simply deleted (and reported if your ISP supports reporting).

Emails classed as "Junk" are usually looking to sell you something, or to get you sign-up to something. These emails typically emanate from large distribution lists, and have embedded hyperlinks (e.g. "click here to join"). Although you may at some point have consented to being contacted, frequently Junk emails are unsolicited and unwanted, using email lists compiled and / or obtained illicitly.

Emails classed as “Genuine” are emails which VerifiTrust has evaluated as posing a very low threat, and which have been assessed as being neither marketing emails (Junk) nor phishing emails (Phishing). Personal or work emails should usually fall into the Genuine category.

Phishing emails often masquerade as being sent from legitimate organisations. Phishing emails almost always seek action from the recipient (either via the “reply to” function, or more commonly using embedded hyperlinks or response buttons). Phishing emails may ask you to (or lead you to websites which ask you to) divulge your personal data, logins, credit card or bank account details, or to download malicious software, or to make a payment. Phishing emails are sometimes used in tandem with telephone based scams involving live callers.

The primary objective of VerifiTrust is to correctly identify malicious Phishing emails with near 100% accuracy. Our current accuracy rate for the correct identification of phishing emails is 99.0%.

Identifying Junk emails, which are annoying but less harmful than Phishing emails, is the secondary objective. Our current accuracy rate for the correct identification of junk emails is 96.0%.

Finally, 98.0% of emails which are genuine are correctly classified as Genuine.

Just reply to the verdict email (you don’t need to put any text in the reply). This will trigger a re-analysis of the email, including a final assessment by a person. It is extremely helpful for users to query verdicts if they believe our system has got the initial verdict wrong. This helps to train the system’s machine learning capability.

The confidence level is provided to give users a measure of how confident the VerifiTrust system is about a particular verdict for a particular email. There are many factors, both positive and negative, that affect the confidence level, and it is felt useful for users to see an overall summary of confidence.
When the confidence level is low or medium, we advise users to examine the system response very carefully during this training phase.

There is a slight and deliberate built-in “cautionary” bias in VerifiTrust. What the system seeks to avoid, above all, is incorrectly classifying Phishing emails as either Genuine or Junk.

This slight cautionary bias means that a small proportion of Junk emails are wrongly initially classed as Phishing, and similarly some Genuine emails are classed as Junk (or very occasionally Phishing). This small bias is designed to minimise the risk of harmful false negatives, namely classifying real phishing emails as either genuine or junk.

In this training phase, the VerifiTrust system is storing forwarded emails, which allows us to analyse and re-analyse a growing pool of suspect and genuine emails as we improve the system.

Once the training phase is completed, we will publish a clear retention policy.

We cannot accept any liability for mis-evaluated emails. This is made clear in our Terms & Conditions. VerifiTrust is a tool which should help in the fight against fraudulent emails, but we cannot guarantee 100% reliability, as scammers methods and tricks evolve continuously.

An important component of the various analyses performed by VerifiTrust is semantic analysis. For the present, this analysis can only be reliably conducted on emails in English. We hope to progressively extend the language capabilities of VerifiTrust in time.

When emails are diverted to spam folders, they are frequently automatically stripped of many key attributes, including live hyperlinks, formatting, blocks of text etc. These truncated emails make it much harder for VerifiTrust’s system to evaluate them.

Users can often move emails held in spam folders to another folder (for example your inbox) and original features of the email will be restored. These emails can then be forwarded (without being opened if you are concerned about their content) to VerifiTrust for a full analysis.

Verifitrust has been founded by Merul Patel and Jack Colbourne, and is based in the UK. We're passionate about helping to prevent phishing scams, having personally seen close friends and elderly family members fall victim to fraudulent emails and phone calls.

Merul has worked in technology and cybersecurity for over 20 years, including developing mobile payment systems and digital wallets with several patents. Originally a scientist, Merul has a PhD in Physics from Cambridge University.

Jack spent most of his career investing in digital infrastructure, primarily fibre networks, mobile telecommunications and data centres. He holds an MBA from INSEAD.