Tech Nation Global Talent Visa: what the guidelines don't tell you (but rejection proformas do)

I have been working with Tech Nation endorsement applications for six years. Up until this year, my success rate was around 93%. That was not because I had some special knowledge others did not have. It was because what was written in the Notion guide was what assessors actually applied. The guidelines were not always easy to interpret, but once you understood them the process was consistent. You could build a portfolio that met the published criteria and expect it to be assessed against them.

Things have changed. The bar has risen, and that is entirely Tech Nation's right as the endorsing body. They can set the standard at whatever level they consider appropriate.

But what I get increasingly worried about is not the bar. It is the gap that has opened up between what the Notion guide says and what assessors are actually applying. People prepare carefully using the published guidance. They build portfolios that satisfy every criterion as written. And they receive rejections citing standards, thresholds and distinctions that were never in the guidance they prepared from. That gap is what this post is about.

One thing worth understanding before we get into the specifics: mandatory criteria is assessed based on your letters of recommendation and the other mandatory documents you submit. Optional criteria is assessed based on the documents you submit specifically for each criterion. These are separate assessments with separate evidence. What goes wrong in mandatory criteria almost always comes back to the letters, which is where we will start.

The letters contradiction: the most important thing to understand

What the guidelines require from your referees

The Tech Nation Notion guide sets out what your three referees need to be. Each person needs to be an established expert in digital technology, who has known your work for at least 12 months with detailed knowledge of your achievements, skills, experience and contribution. They need to be able to provide specific examples.

That is a meaningful requirement. Twelve months of detailed knowledge is not something a stranger can provide. Your referees, by necessity, are people who have worked with you, mentored you, invested in you, or collaborated with you. They are part of your professional world.

What assessors look for in those letters

Here is where the contradiction emerges. The standard I consistently see applied in assessments is that letters from your immediate professional network are treated as personal endorsements rather than independent sector-level recognition. A letter that is specific, detailed and evidenced can still be considered insufficient if the assessor reads it as coming from someone too close to the applicant.

These two standards cannot both be satisfied at the same time. Someone who has known your work in detail for more than a year is, by definition, a professional contact. They are in your network. You cannot simultaneously have someone with 12 months of deep professional knowledge of your work AND have them be independent of your professional circle. The guidelines require the first. In practice, the second is what gets scrutinised.

The distinction that appears to matter, based on what I have seen succeed and fail, is the difference between vouching and expert assessment. A letter that reads as "I know this person and I can tell you they are excellent" tends to fail, even when it includes specific examples. A letter that reads as "in my capacity as someone with expertise across this field, I have observed this person's work and it is exceptional relative to what I typically see" tends to succeed. The framing matters as much as the content. But this distinction is never explained in the guidelines, and most applicants and most referees would have no way of knowing it.

Mandatory criteria: the undocumented thresholds

The bar is sustained, sector-wide recognition, not one or two things

The Notion guide describes the mandatory criteria as requiring recognition as a leader or emerging leader in the digital technology sector. What I consistently see applied in practice is something closer to a pattern of sustained recognition across multiple sources and over time. A single award, a single conference appearance, or a single published piece tends not to satisfy the standard even when it is a strong one. Assessors are looking for a pattern, not an instance.

The types of evidence that appear to satisfy this standard consistently are: coverage in established technology media with editorial credibility, speaking roles at major industry conferences where you were invited rather than applied, recognised industry awards granted to you in recognition of achievement rather than competition entries, and letters from people who can speak to your reputation across the broader field. Each of those on its own may not be enough. Together, they build a picture of sustained recognition.

National recognition and the wider sector question

One of the more frustrating patterns I have seen is how local ecosystem contributions are treated versus wider sector recognition. An applicant with a significant role at an EU national-level accelerator had that contribution noted as useful to the local startup ecosystem but not as evidence of recognised leadership in the wider digital technology sector.

The problem is that the guidelines do not define what "wider" means. They do not say recognition must be international. In fact, the Tech Nation guidance explicitly lists national awards as eligible evidence for some criteria. If national recognition counts for awards, it is not obvious why a significant national-level role does not constitute sector recognition. For applicants from smaller tech ecosystems, national recognition can genuinely represent more influence in their field than a peripheral role at a large organisation in a major tech hub. But the practical standard that appears to be applied is one of sector-wide visibility that goes beyond your immediate geography, even though that standard is never stated.

Media coverage and the editorial credibility question

The Notion guide says media coverage counts as evidence. What I see applied in practice is a distinction between coverage that has editorial credibility and coverage that does not. Founder interviews, sponsored content and company PR tend to be noted as demonstrating subject-matter knowledge without demonstrating significant external recognition. The undocumented standard appears to be that the journalist or outlet independently chose to cover you, not that the coverage was arranged or paid for. That distinction is not defined in the guidelines.

Similarly, the phrase recognised major technology media or established industry publications appears in assessments without ever being defined in the guidance. There is no list of qualifying outlets. Applicants have no way of knowing whether a specific publication meets the threshold before they submit.

OC1: the innovation criteria

The guide says innovation. Assessors apply breakthrough novelty.

OC1 in the Notion guide requires a proven track record for innovation as a founder or senior executive of a product-led digital technology company. That is the published standard. It does not use the words "breakthrough" or "novelty." Those words appear in how assessments are applied in practice. The working standard appears to be not just that you built something original, but that it represents a genuine advance that did not previously exist in this form.

An important point on this, specifically for founders. If you have built a product that is live in the market, with real paying users, that is a commercially validated product. Users paying for something is, in one reading, the market's verification that the product works and addresses a real need. Where does the Tech Nation guide say that is not sufficient to demonstrate innovation for a founder? For an employee, independent verification makes obvious sense: you need to prove your individual contribution to something your employer owns, and a patent or external recognition does that. But a founder's live product with real demand is the evidence. The guidance does not state that this requires additional independent validation on top of the product existing and being used. That requirement appears in practice but not in writing.

Commercial success alone does not demonstrate innovation, and the OC1 versus OC3 distinction is real. Revenue and user growth belong in OC3. But a founder arguing innovation through a genuinely novel product that is live and in demand is arguing something the guidelines appear to support. The gap is the undisclosed standard that this also needs to be validated by someone outside the company.

Third-party validation is required, but not stated

The Notion guide does not say OC1 requires independent third-party validation of your innovation claims. In practice, self-described innovation without external confirmation tends to be insufficient. What appears to satisfy this standard is an independent investor citing the innovation specifically, a grant from a recognised public body, a patent granted by an official body, or editorial media coverage that describes the innovation rather than simply reporting what you say about it. Self-authored descriptions of what you built, however detailed, do not appear to satisfy the standard on their own.

OC2: advancing the field

Advisory roles need contribution detail, not just a listing

Being named as an advisory board member or listed in a programme is not sufficient on its own. What appears to be required is evidence of what you specifically contributed in that role: what you advised on, how much time you committed, and what changed in the field or organisation as a result. A name in a list demonstrates you were associated with something. It does not demonstrate you advanced the field.

Social media evidence is not accepted

Screenshots of your own social media posts, including LinkedIn, are not accepted as OC2 evidence. This is stated relatively clearly in the guidance but is still a common mistake. Even when the content itself is substantive, the format is the issue.

Self-published content without editorial process

Articles published on LinkedIn, Medium, or your own platform tend to be treated as self-published rather than as contributions to the advancement of the field. Even content published by a third-party outlet can face the same issue if the editorial process is not clear. The distinction being applied is between content that a publication chose to commission or accept through an editorial process, and content that was placed, sponsored, or self-submitted without independent editorial judgement.

What happened after matters as much as what you did

Publishing an article, giving a talk, or running a workshop used to be sufficient for OC2. The standard now appears to include impact after the activity: was the article cited or acted upon? Did the talk change how people in the field approach something? Did the mentoring result in measurable outcomes for the mentees? What you did is only part of the evidence. What happened because of it is increasingly what assessors appear to assess.

OC4: research contribution

Sustained track record is the standard, but it is never defined

OC4 requires research contribution to the field. The Notion guide does not define how many papers, over what period, with what level of citation or reach. In practice, a small number of recent papers tends to be considered insufficient as a sustained track record. First-authored papers at major international conferences have been rejected as too recent and with limited adoption and reach. The number of papers, the period over which they span, and their adoption by other researchers all appear to be assessed, without any of those thresholds being disclosed.

GitHub engagement as a research metric

Assessors appear to use GitHub stars and engagement as a proxy for whether research has been adopted by the field. This is not mentioned in the Notion guide at all. Applicants submitting strong academic work have no way of knowing that their repository's public engagement may be scrutinised as part of the assessment. If your research has an associated codebase, its public engagement is worth paying attention to before you apply.

The timing question

The guidance updated to say activities done too close to the application date will not be considered. It does not define "too close." The safe working assumption, based on what I have seen rejected, is that the most recent activity in your portfolio should be at least eight months before your submission date. Activities from six or seven months before submission have been rejected as too recent. Any activities you are building now cannot be submitted for at least eight months, and the overall portfolio needs to show a pattern over time rather than a cluster of things you did specifically to apply.

The inconsistency problem

None of the above would be as significant if it were consistently applied. In my experience, it is not. I have worked with applicants who were rejected with profiles substantially stronger than others I know to have been endorsed in the same period. The gap between what the guidelines say and what assessors apply varies, and what satisfies the standard for one assessor does not always satisfy it for another.

This does not mean the route is impossible. It means that preparing only against the published guidelines is not enough. The portfolio needs to go beyond the technical minimum because you do not know which undocumented standards will be applied to your specific application. And a first rejection, with a clear proforma, is often more useful than a successful application would have been, because it tells you precisely where the gaps are.