Read, not retyped
Job descriptions, performance reviews and the situations your people describe are read into a brief for the build, and the psychologist sets what the assessment has to measure.
Built by Miraai, delivered on Podium365
Technology that scales the science, with psychologists in charge.
A custom assessment used to take a team of psychologists months. Ours take weeks, because the slow work of drafting, screening and piloting is automated and a psychologist approves every stage. The finished assessment then runs on Podium365, run by Allied Assessments at scale for over a decade, with an anti-cheating design built on our own published research.
Weeksfrom your job description to a live, documented assessment
Every stageapproved by a psychologist by name, and the approval kept
ISO 27001Allied Assessments’ certified information security
20+languages on the platform
Scaling the science
AI does the slow work, and a psychologist decides every stage.
Most providers use automation to cut what they do. We use it to do more of the science than a team could do by hand, so that a judgment test built for one role gets the same drafting, screening, piloting and trial a flagship product would, and a psychologist signs off each step before the next one starts.
Thousands, not dozens
The system drafts far more scenes, responses and items than the assessment needs, then screens them for target, reading level and wording that could disadvantage a group.
Set by experts
For judgment scenes, an AI rater panel informs the scoring key while the test is built, and your experts rate the responses too. Their agreement sets the final key, and the panel’s job is to inform it, never to decide it. No AI model scores, ranks or rejects a candidate.
Before anyone sees it
Pseudo Factor Analysis reads the structure of the items, and Qarin digital twins answer them, so items that fail or are unfair come out before a candidate meets one.
People, every time
The assessment is tried on real people before it counts for anybody, and the documentation that records all of it is handed over with the assessment.
This is what the automation is for. A 2026 meta-analysis in the Journal of Applied Psychology priced a customized judgment test at about 24 times an off-the-shelf cognitive test, which is why most organizations have only ever had the off-the-shelf kind. The SJT Studio is why a test built for one role costs what an off-the-shelf test does. Lievens et al., 2026, doi:10.1037/apl0001420
- Fast tests, expert judgmentHow the nine-step pipeline builds a video judgment test with a psychologist deciding every step, on the Open Science Framework
- The AI panel against the expertsHow the panel’s ratings compare with a panel of human experts’
- The Pilot Before the PilotThe Qarin digital twins, with the code to run them on GitHub
The platform
Set the assessment up to match the job.
Most assessment platforms make you take the whole product on their terms and in their look. Podium365 is modular, so you pick the scales, map them to your competency model, brand the experience as your own and run it in the languages your candidates speak.
Your brand, end to end
Candidates see your logo, your colors and your messaging from the invitation to the last screen, so to them it is your assessment.
Only the scales you need
Run the whole set or a single test, or combine personality, judgment and reasoning in one sitting, and keep the time it takes where your process needs it.
Mapped to your competencies
Scores and reports align to your own framework, so what comes back reads in your organization’s language, and role profiles and pass marks are yours to set.
More than twenty languages
The candidate experience, the instructions and the reports run in more than twenty languages, so one program can cover every market you hire in.
Reports the way you want them
Selection and development reports for individuals and group reports across a date range, as PDF or Word, with norms and role profiles applied at the point of scoring.
Hosted, on any device
Nothing to install and nothing for IT to maintain. Candidates use a phone, tablet or laptop, and administrators manage everything from a browser.
Anti-cheating
We watch how a test is taken, not what the answers contain.
Language models now pass standard reasoning tests with ease, which is why ours is built to resist them, and tools that try to recognize AI-written answers break with every new model and misfire on people writing in a second language. So the integrity design on Podium365 does not look at the answers at all. It looks at the session, which no model release can change, and it is built on two studies of ours that are public.
The reasoning test adds a design of its own. A question shows only while the candidate holds it on the screen, and the clock keeps running when they let go, so working it out is faster than looking it up.
Clear expectations first
Before the test starts, candidates are told plainly what counts as their own work and agree to it. Setting the expectation is the cheapest safeguard there is, and it changes behavior.
Identity, all the way through
Facial validation confirms that the person who completes the assessment is the person who was invited, and that the same person is there from the first item to the last.
How the test was taken
The session itself is the evidence. Leaving the test window, timing that does not fit the items and other signals are combined by a statistical model into a confidence score for each session, whatever model or tool a candidate might have used. It reads how the test was taken, never what the answers contain.
A person decides
A low confidence score raises a flag for a person to review, never a verdict. It never fails a candidate on its own, and the threshold is set so that people who worked alone are almost never flagged.
106,000+reasoning test sessions in the field study, with and without monitoring
372participants in the experiment that tested the confidence score against sessions with and without help
AUC .82in telling helped sessions from honest ones, and every flag goes to a person
Source: Behavioral, Model-Agnostic Integrity in Online Ability Testing, 2026, a field study of more than 106,000 sessions and an experiment with 372 participants. Read it on the Open Science Framework
Security
Certified to ISO 27001.
Assessment data is some of the most sensitive information an employer holds about a person. Allied Assessments’ information security management system, which runs Podium365, is certified to ISO/IEC 27001.
Access control
Single sign-on for your administrators, role-based permissions and an audit trail of who did what and when, so candidate data stays with the people who need it.
Data protection
Data protection practices are GDPR compliant, and every API call runs over HTTPS with TLS 1.2 and short-lived, token-based authentication.
A certificate you can ask to see
The ISO/IEC 27001 certificate is issued by an accredited certification body, with surveillance audits between certifications, and your security team can ask to see it.
Integration
Connected to your talent systems.
Your recruiters should never have to leave their ATS to send an assessment or read a result. The Podium365 integration API is documented, versioned and in production, with integrations live in Workday, Greenhouse, SAP SuccessFactors and iCIMS, and the same API and webhooks reach any system that can make an authenticated call.
Your ATS or HRIS
A candidate reaches the assessment stage in your process.
- Create the participant with your own record ID
- Assign the assessment, language and deadline
- Sign in once with SSO
Podium365
The candidate completes the assessment on any device, in your brand.
- Status available on demand
- A webhook fires on completion
- A webhook fires when results are scored
Back into your process
Results land where your team already works.
- Raw, standardized and percentile scores
- Scored against your norms and role profiles
- Reports as PDF or Word
For your engineers
A REST API with JWT bearer authentication, TLS 1.2 and webhook callbacks for completion and scoring, and your own integration IDs stored on every participant so records reconcile without manual matching. The full documentation is on podium365.com.
For your recruiters
Nothing changes about where they work. Assessments are sent from the ATS, progress shows in the ATS, scores come back to the candidate record and reports open with one click.
Built by Miraai. Delivered on Podium365
The judgment scenes and the SJT Studio that produces them, the hold-to-view reasoning test and the integrity model are ours.
Next step
Tell us what you already use.
Tell us which ATS you run, which languages your candidates need and what your security team will ask, and we will come back with the answers rather than a sales deck.