What Kirnova delivers
This guide defines the terms used across the developer documentation: what Kirnova and TIGER are, what each assessment is, who can start it, and what you can put in front of a person.
Overview#
Kirnova is a decision intelligence company. It produces readings about people that help an organisation make better decisions.
TIGER is the framework that reads how a person works. A person answers an assessment, and TIGER turns the answers into five readings. TIGER has three tiers: TIGER Light, TIGER Core and TIGER Student. Refer to these by name. There is no single "TIGER test".
Kirnova builds on published research and human behaviour. It is a signals and inference layer, not a laboratory instrument. The right comparison is the CV, the unstructured interview and the manager's impression a decision-maker would otherwise rely on. Each reading is published with the range it most likely sits in, because a bare number would imply precision the assessment does not have.
Terms#
| Term | Meaning |
|---|---|
| Dimension | One of the five readings TIGER produces |
| Range | The band a reading most likely sits in. Every reading has one |
| Sitting | One person answering one assessment, start to finish |
| Respondent | The person sitting the assessment |
| Partner | An organisation delivering a Kirnova assessment in its own product or process |
| Commission | To start an assessment for another person and pay for it |
| Embedded frame | The assessment running in a frame you place in your product |
| Machine-driven API | Your servers drive the sitting and you render every screen |
| Quality note | A sentence shown when a sitting's answers did not hold together |
The five dimensions#
TIGER returns the same five readings, in the same words, for every respondent.
| Dimension | What it reads |
|---|---|
| Trust | How honestly and fairly the person deals with others when nobody is checking. This sentence is sent with the reading; render it beside the reading. It describes conduct drawn from the person's own answers, not their character or integrity |
| Influence | Whether the person's view tends to become the group's view. It does not read seniority, popularity or how much someone talks |
| Grit | How the person responds to pressure and difficulty. It does not measure how hard someone works |
| Execution | Whether intentions turn into finished work. It reads follow-through, not skill at any task |
| Resilience | How quickly the person recovers after a setback. A low reading does not mean a person is fragile |
These rules apply to all five:
- A high reading is not better and a low reading is not worse. There is no overall score and no composite.
- Every reading arrives with its range. If two people's ranges overlap, the assessment cannot distinguish them.
- If a range still covers the middle, the reading is reported as one that cannot be called either way. This is not "average", and your interface must not display it as average.
Kirnova does not send formulas, weightings or inference rules. See Reading a result for display requirements.
Light and Core#
Light can help you understand yourself. Core can be used to make decisions about you.
TIGER Light is started by the person for themselves. No company, recruiter, employer, institution or partner can start it or send a person to it. It is never available to partners, on any surface. It is not a smaller Core or a preview: it covers all five dimensions, and differs in depth of questioning, depth of reading and permitted use.
TIGER Core is commissioned by a decision-maker about another person. It can inform an employment decision, as can TIGER Student. Every third-party invitation is for Core or Student, never Light.
A candidate cannot start their own first TIGER Core.
The tiers#
| TIGER Light | TIGER Core | TIGER Student | |
|---|---|---|---|
| Who starts it | The person only | A decision-maker: employer, recruiter, organisation or other paying party | An education institution or programme |
| Purpose | Understanding yourself | A decision about the person | Careers and course guidance for adult learners |
| Can inform an employment decision | No | Yes | The trait readings can. The interests reading never can |
| Dimensions | All five | All five | All five, plus an interests reading |
| Available to partners | Never | Yes | Yes |
What a partner can put in front of a person#
| Assessment | Embedded frame | Machine-driven API | Who can commission it |
|---|---|---|---|
| TIGER Core | Yes | Yes | A decision-maker or other paying party |
| TIGER Student | Yes | Yes | An education institution or programme |
| Kirnova Calling (three lengths) | Yes, result returned to you | Yes | A partner, for career exploration only |
| Kirnova DISC | Yes, result and downloadable document returned to you | Yes | A decision-maker or other paying party |
| Kirnova HEXACO | Yes, result and downloadable document returned to you | Yes | A decision-maker or other paying party |
| Kirnova ORVIS | Yes, result and downloadable document returned to you | Yes | A partner. Never a signal for selecting between people |
| Kirnova Cognitive | Yes, result returned to you, while Kirnova has it switched on | Yes, while Kirnova has it switched on | A decision-maker or other paying party |
| TIGER Light | Never | Never | Only the person themselves |
Kirnova sets which assessments your account can run. If you request an assessment your account cannot run, or one a surface cannot administer, the request is refused. Kirnova never substitutes a different assessment. An account with no assessments configured cannot start any sitting, so agree the list with Kirnova before you build.
Kirnova Calling is licensed for career counselling, planning and exploration. It must not be used as an input to an employment decision, and this restriction applies to you.
The assessments#
TIGER Core#
Used for a decision someone else has to make about a person, most often hiring. See Permitted use for the wider list.
- Questions: 48 scored questions and two attention checks. Some sittings include a short extra exercise of eight ranking groups. Typically about 15 minutes.
- Returns: all five dimensions, each with a whole number, its range, a one-word confidence reading and a sentence describing the reading. It also names the dimension that stands out most within the person's own profile and two worth a closer look. This ordering is within one person, never a comparison between people. Where they apply, two short written readings are included: how the person's answers shifted between general conditions and situations where others rely on the outcome, and openness to new ways of working. A quality note appears when answers did not hold together.
TIGER Student#
Used for careers and course guidance for adult learners in colleges, universities and adult programmes.
- Questions: 80 in total: a 48-question block on how the person works, built to the same shape as Core's, a 30-question interests block, and two attention checks.
- Returns: the five dimensions as Core returns them, plus an interests reading: six career types, the person's six-letter pattern and the closest occupations. Career types that tie are named together.
- Scope: 18 and over. Nothing in your product may imply that a secondary school can use Kirnova.
- Display obligation: the interests reading comes from a licensed public occupational database. The attribution and non-endorsement wording travels with the result, and you must display it wherever you show that reading.
Kirnova Calling#
Used for career exploration only.
- Questions: one of three published lengths: 180, 60 or 30 scored questions, each with two attention checks. The 60-question form is the usual one and takes about 10 minutes. The 180-question form uses a three-point scale; the others use a five-point scale.
- Returns: six career types, the person's six-letter pattern and the closest occupations from a public database of more than 900 occupations. Tied types are named as tied.
- Delivery: runs in the embedded frame and on the machine-driven API. The person sees their reading in the frame, and the same reading is returned to you by signed message and on request. No downloadable document is produced.
- Licence: the restriction on employment decisions above applies.
Kirnova ORVIS#
Used to show what kinds of work appeal to a person.
- Questions: 24 scored questions and two attention checks, about 10 minutes.
- Returns: an order of eight kinds of work, with tied themes named as tied and no number for any theme. The result carries a sentence saying it is not a signal for selecting between people.
- Delivery: runs in the embedded frame and on the machine-driven API. The reading is returned to you by signed message and on request, with a downloadable document.
Kirnova Cognitive#
Used to show how well a person solves new puzzles against the clock.
- Questions: 36 picture puzzles under one 40-minute clock.
- Returns: a score out of 100 with its range and a word (Low, Moderate, High or Very high). It never returns a comparison with other people. While the hardest puzzles are still being tuned the result says so.
- Delivery: runs in the embedded frame and on the machine-driven API, only while Kirnova has it switched on. The reading is returned to you by signed message and on request. No downloadable document is produced.
Earlier behavioural assessment#
This assessment is retired. No new sitting can be started on any surface. Results already held remain readable, with their scores exactly as stored. Its separate personality-trait block was removed from every channel on 2026-10-20.
Interests#
TIGER Student's interests reading and Kirnova Calling both describe what a person is drawn to. Two rules apply:
- Access. A hiring buyer can hold and read an interests result.
- Interpretation. Interests are never a hiring signal. They are not evidence of suitability for a role, and they must be visibly separated in your interface from anything used to inform a decision.
If you display an interests reading to anyone making a decision about the person, carry the statement that it is not a measure of suitability.
Delivery surfaces#
| Embedded frame | Machine-driven API | |
|---|---|---|
| Who draws the questions | Kirnova, in a frame you place | You |
| What you build | The launch, the surrounding page and what happens after | The whole sitting: questions, controls, progress, navigation and error states |
| Control over appearance | Limited, see Appearance | Complete |
| Effort | Lower | Higher |
| Assessments | TIGER Core, TIGER Student, Kirnova Calling, Kirnova DISC, Kirnova HEXACO | TIGER Core, TIGER Student |
On both surfaces:
- The person is identified before the sitting starts.
- Kirnova scores answers on its servers. Your system submits answers, never scores. A result that cannot be re-derived from its answers is refused.
- The result is sent to you when it is stored, and delivery is retried if the first attempt fails.
In the embedded frame, a respondent is never shown a Kirnova sign-in page and cannot navigate out of your product. Before the first question, each respondent reads a six-point notice and selects "I understand and acknowledge". The notice carries no Kirnova name. It is not consent and does not replace the notice you give the person. The full text is in the API guide.
For mechanics, see the API guide and the integration playbook.
Permitted use#
TIGER informs; people decide. A result can inform any employment decision: hiring, promotion, retention, pay, performance, development, management, disciplinary action and termination. A person must make the final decision, and the organisation remains responsible for applying its own judgement and the law that applies to it.
A TIGER Light result must not inform a decision about the person. An interests reading is never evidence of suitability.
What Kirnova does not do#
- There is no pass mark and no automatic rejection, and no setting creates one.
- There is no team or organisation-level reading and no comparison of two people. Do not build screens that assume one exists.
- There are no profile names, archetypes or type labels. The current framework returns five readings, a standout dimension and two to look at more closely.
- Kirnova does not send percentiles, percentages or accuracy figures. Confidence for each reading is one word: High, Moderate or Low.
- Scoring uses only the answers. It does not use a name, photo, age, school or CV.
- No AI model reads answers, and none is involved in scoring.
- There is no anonymous respondent, on any surface.
Appearance#
You can set the following in the embedded frame:
| Setting | Reaches |
|---|---|
| Light, dark, or follow the person's device | All assessments you can run in the frame |
| Primary brand colour | All assessments you can run in the frame. Kirnova adjusts lightness where needed to keep it readable and keeps the hue |
| Logo, welcome and completion wording, removing the Kirnova name | Nothing you can start today |
| Second colour, typeface, button shape | Nothing |
| Styling on the downloadable document | Nothing. The document uses Kirnova's presentation |
The sitting displays the assessment name, such as "TIGER Core". If your launch requires a pixel-matched appearance, use the machine-driven API and build the screens yourself. A custom web address can be configured on an account, but no link Kirnova generates uses it today.
Who can see results#
Your integration reads results with its own credentials. A result from a sitting your system started comes back to your system, and access does not expire on a timer.
People signed in to Kirnova read results through their account. The account that holds the relationship with the candidate can open the reading, as can team hiring managers and administrators (on a recruiter team, sourcers and administrators), for people your organisation invited. A viewer can see that a person and a reading exist and when it was taken, but cannot open the reading. A person invited to a sitting on Kirnova is asked for agreement before the first question, for your company only. Without agreement, only the existence of a reading and its date are visible.
The respondent#
Every respondent becomes an identified person on Kirnova.
- Identity. In the embedded frame, an email address and a name are required. On the machine-driven API, an email address, a phone number or your own identifier is enough. Email is preferred because it is how the person reaches their result later. A name alone is never an identity.
- Own access. The person can sign in to Kirnova and see their own result. Their access never lapses. When a buyer's access ends, the result is locked, not deleted.
- Answers. Raw answers are retained with the sitting so it can be reproduced under the version it was administered under. Your notice to candidates should say so.
- Transparency. The person is told who asked and can see who holds their result.
For what the person sees, read The respondent journey.
Next steps#
| Question | Guide |
|---|---|
| How do I connect? | API guide |
| In what order do I build and test? | Integration playbook |
| What does the person see? | The respondent journey |
| How do I display a result? | Reading a result |
| What is and is not available? | Capabilities and limits |
| What can I build with it? | Use cases |