Next Generation Verification

Our verification engine has just had its biggest upgrade yet.

Meet Gocertify Intelligence: the new AI layer at the core of our verification. Built on the latest AI and LLMs, it understands every uploaded document rather than just its text. That one change affects how many genuine shoppers get through, and how many new audiences a brand can reach down the line.

‍

Why we built it

Gocertify's mission is to give brands the tools to run the most effective closed-group strategy possible.

Verification is at the very core of this, and it’s why 500+ brands across 100 countries trust Gocertify to manage their offer programmes.

We're always looking for ways to drive more incremental revenue for our brands, and we saw a real opportunity here: make the thing our entire business is built on better. Faster, more accurate, and built for the scale our brands are asking for.

‍

What actually changed

Document verification used to work via OCR technology that read the text on a document and checked it against a fixed set of rules. That worked, but it had limits. The more groups, markets, and document types we added, the more complex those rule sets became, and rigid rules struggle with anything that doesn't follow a standard format.

Gocertify Intelligence changes how that works, in two steps.

First, AI and LLMs read the document properly. When someone uploads a document, the engine now understands what it is, where it has come from, what organisation it relates to, and whether it is valid. For example, A UCAS letter is recognised as a UCAS letter, with the right name and the right expiry date pulled out as structured information. The AI then extracts the data in that document in a structured format.

Second, that structured information is checked against a policy. A defined set of acceptance criteria for that specific group, applied consistently every time. The policy decides the outcome; the AI's job is to understand the document well enough for that decision to be accurate. And when something doesn't pass, the engine knows exactly why, not just that it failed.

What this means for brands

‍

As we were building this new engine, we had five core pillars that we built around to improve conversion rates and drive more incremental revenue.

1. Built to scale. A more capable engine means new groups, new document types, and new markets no longer require a rule set built from scratch. Verification shouldn't be the reason a brand can't expand into the audiences that matter, and increasingly, it isn't.
2. More accurate, fewer false rejections. Genuine shoppers who'd previously been knocked back by rigid, text-matching rules, because the document types do not follow a set structure, are now approved faster, without manual review, because the system understands the document rather than just scanning it.
   [X]% improvement in verification accuracy; X% reduction in shoppers incorrectly rejected.]
3. Stronger fraud resistance. Understanding a document, not just reading the text on it, closes a gap that text-matching could not, particularly against AI-generated fake documents, a fast-growing threat. Gocertify Intelligence extracts and structures the data in a format matched against set policies. These policies are the equivalent of what a human would check for when verifying a document.
   [X% reduction in fraudulent verifications accepted.]
4. Faster verification, better conversion. Every second between a shopper proving they're eligible and getting their code is a chance to lose them.
   [Verification time reduced from X to X; conversion increase of X%.]
5. Better shopper experience. Our 24/7 support team is trained in handling queries for closed user group documents. Gocertifty intelligence reduces the number of manual reviews needed, and gives our support team exact information as to why the verification was rejected.
   [Reduced support response ratre from X to X mins]

More eligible shoppers complete their journey, and have a better experience while they're at it, and that’s why we’ve seen conversion rates increase across the board.
[Show graph with conversion rates]

‍

Why it’s so powerful 

Gocertify processes 4 million+ verifications a year across a 2 million-strong shopper network for 500+ brands. Every document that comes through teaches the engine something, like what a real payslip looks like in a country we've just expanded into, what a fraudulent document tends to get wrong, what a genuine but unusual document looks like.

That's a depth of real-world document data that's near impossible for any single brand or any provider to replicate. The more shoppers we verify, the better Gocertify Intelligence will get. 

Why we built it this way

Most verification approaches fall into one of two camps. Some rely entirely on institutional databases - checking a shopper's details against university registers or employer directories. That works well where those databases are extensive, but they have limitations. They rely on every person that falls into that group to be signed up; it's slow to extend to new countries, and most providers fall back to manual document review wherever a database connection doesn't exist.

Others rely on membership cards or apps a shopper signs up for in advance, proving they were eligible once, rather than confirming they still are. That's a different model entirely, and it doesn't replace document-based verification; the two run side by side without conflict.

Gocertify's approach starts with the document itself, understood properly by AI, with the ability to layer database checks on top wherever they're genuinely useful. That means brands aren't choosing between speed and accuracy. They get fast, accurate verification today, with the flexibility to add deeper checks for specific markets or groups as needed. In practice, that means verification in under 30 seconds, compared to processes that can take days elsewhere.

Where this is going

Gocertify Intelligence is live now, working underneath every verification happening on the platform today. What we’re most excited about is where this can take our brands in the future, new groups and new regions, so that they have a verification tool that doesn’t get in the way of their growth objectives.