Software advice
Best AI Pre-Employment Assessment Software in 2026
AI pre-employment assessment software is a web platform where employers send candidates structured tests, work simulations or interviews before a hiring decision and receive scored results, with AI that builds tests, grades open answers, flags cheating or runs the interview itself. It is not an applicant tracking system or a background check service, although it connects to the first.
Quick answer
TestGorilla ranks first because it prints plan prices, has a free plan and states EEA hosting and a no-training rule for generative models. Codility fits technical hiring, and Bryq fits small teams that want a $69 monthly entry price.
A talent acquisition lead choosing an assessment tool faces a stack of demo requests and very few price lists. Vendors now pitch AI test builders, auto-scoring, AI interviewers and proctoring, yet the pages say little about what the AI costs, where candidate answers are stored or whether they feed model training. Hiring is also a regulated activity, so bias audits and EU AI Act statements matter more here than in most software categories. This ranking weighs printed prices, stated hosting and retention, named certifications, plan gates and how clearly each vendor describes its AI. We read vendor pages in November 2026 and ran no tool, so performance claims are the vendors' own.
The 13 best AI pre-employment assessment software in 2026, ranked
There is no score. Vendors can pay for more visibility on this website, for example a higher place in a ranking or more room on a page. Commercial relationships can also affect which products we feature and how prominently. How we rank.
| # | Tool | Best for | Free plan | Entry price |
|---|---|---|---|---|
| 1 | TestGorilla | Small and mid-size employers hiring across many roles who want printed prices and a free start. | Yes | $215 per month per account, billed annually (free plan at $0) |
| 2 | Codility | Engineering hiring teams that want candidates to work with AI tools and be assessed on it. | Trial | $800 per month for 25 invites, or from $4,800 per year |
| 3 | Bryq | Small recruiting teams wanting a printed monthly price, EU hosting and cognitive plus personality testing. | Trial | $69 per month on annual billing (Pro) |
| 4 | Maki People | High-volume screening teams that want a voice interviewer and named ATS connections. | No | Quoted |
| 5 | Criteria | Employers wanting cognitive, personality and video interviewing from one vendor with named certifications. | Trial | Quoted |
| 6 | Canditech | Recruiters who want AI-built, role-specific assessments and a chatbot for pre-screening by WhatsApp. | Trial | Quoted |
| 7 | eSkill | Mid-size US employers screening up to roughly 1,000 candidates a year with a ready test library. | No | Quoted |
| 8 | Vervoe | Enterprise teams that want candidate answers graded by AI with an explanation attached. | No | Quoted |
| 9 | Mercer | Mettl | Large employers wanting proctored online tests, AI-assisted coding assessments and competency simulations. | No | Quoted |
| 10 | iMocha | Companies that assess candidates and also want skill profiles of current employees inferred by AI. | No | Quoted |
| 11 | Alva Labs | Employers wanting psychometric tests and an AI agent that runs the whole hiring flow. | No | Quoted |
| 12 | Harver | Large employers hiring in volume that want game-based and situational judgement assessments. | No | Quoted |
| 13 | SHL | Large organizations that want established personality, cognitive and situational judgement instruments. | No | Quoted |
- 1
TestGorilla leads because a buyer can read its prices, start free and find retention periods and a no-training rule for generative models in the privacy policy. EEA hosting and SOC 2 Type 2 add to that. The 350-plus test library suits generalist hiring. It loses ground for deep technical screening, where Codility is stronger, and the best AI features sit on Plus.
Strong on
- Free plan plus Core and Plus prices printed on the pricing page
- Privacy policy says personal data does not train generative AI models
- Servers in the European Economic Area, SOC 2 Type 2 stated
Weak on
- AI interviews beyond the free allowance, ID verification and API access require Plus
- Free plan caps usage at 10 credits a month
- Monthly prices are not listed, only annual rates
Skip it if: Teams that need validated psychometric instruments or a tool built specifically for senior software engineering hires. Core adds the full test library and a shared credit pool; AI interviews, ID verification and ATS access need Plus.
- 2
Codility takes second place because it prints invite-pack prices, states retention periods and names EU or US hosting, which few rivals do. Letting candidates use an AI Assistant or Claude Code, with every exchange recorded, matches how developers now work. The price puts it out of reach for light use, and it is a developer tool first, so it ranks below the generalist TestGorilla.
Strong on
- Prices printed for online plans, with invite counts stated
- AI Assistant and agentic coding interviews with interactions recorded for review
- SOC 2 Type II and ISO 27001 stated, EU hosting option on Enterprise
Weak on
- Smallest monthly pack costs $800 and covers only 25 invites
- SSO, ID verification, video proctoring and the full API sit only in Enterprise
- Online plans include only one standard ATS integration, with the rest on Enterprise
Skip it if: Companies hiring mostly non-technical staff or recruiting only a few engineers a year on a small budget. Online plans cover technical Screen and Interview; SSO, video proctoring and the full API need Enterprise.
- 3
Bryq ranks third because it shows a low entry price, hosts on AWS in the EU and holds ISO 27001 and a SOC 2 Type II examination. An AI Fluency Assessment and an MCP connection for AI assistants are newer touches. It loses to TestGorilla on free usage and on documented retention, and its fair-usage clause excludes agencies.
Strong on
- Pro plan has all assessment types and unlimited invitations at a stated price
- Hosted on AWS in the EU, with ISO 27001 and SOC 2 Type II
- AI Fluency Assessment tests how candidates work with AI tools
Weak on
- Contents of the Individual and Team plans are not published
- Recruitment agencies must contact Bryq because of fair-usage limits
- Retention periods and AI training use are not stated on the pages read
Skip it if: Staffing agencies and teams that need to bring candidate data into a deeply customized reporting stack. Pro totals $828 a year with a one-year commitment; a monthly option exists and Enterprise is quoted.
- 4
Maki People sits fourth because it names its integrations (Greenhouse, Workday, iCIMS and more), publishes an ISO 27001 claim and cites a New York City bias audit from March 2026. The voice interviewer in 45-plus languages is distinctive. No prices are shown and retention for assessment data is not stated, so it trails the three tools above on transparency.
Strong on
- Voice interviewer covers 45-plus languages around the clock
- Six named ATS and HRIS connections including Workday and SAP SuccessFactors
- ISO 27001 plus a bias audit by Holistic AI cited for New York City
Weak on
- Credit prices are not published, only a custom quote route via sales
- Retention periods for assessment and proctoring data are not stated
- Hosting region is not named on the pages we read
Skip it if: Small teams that need to compare prices before talking to a salesperson or that hire only a few people a year. Credit-based pricing tied to candidate interactions; products are Shiro, Ken, Mochi and Kumi, all by custom quote.
- 5
Criteria is fifth: the product range is broad (assessment, async and live video interviews, and a combined Hire plan) and the security page lists ISO 27001, ISO 42001 and SOC 2 Type II, which is more than most rivals state. Every price is hidden, AI proctoring and AI interviewing are upgrades, and hosting is US or Australia only.
Strong on
- ISO 27001, ISO 42001 and SOC 2 Type II all stated on the security page
- One vendor covers testing and video interviews in a combined plan
- AI Proctoring and an AI Interview Agent are available as upgrades
Weak on
- No plan price is published, only request-pricing buttons on all three plans
- AI features are paid upgrades rather than part of base plans
- Hosting is in the United States or Australia, with no EU region named
Skip it if: European employers who need EU-only data storage or buyers who want AI scoring included in the base price. All three plans show a Request Pricing button; a free trial link exists with no stated length.
- 6
Canditech lands sixth because its AI features are clearly named (assessment builder, auto-scoring agents, AI proficiency tests) and the trust center states EU hosting on AWS and ISO 27001. Tier caps are visible. Prices are not, the legal entity and country are not stated on the pages read, and the privacy policy page did not load, so data terms are unconfirmed.
Strong on
- AI Assessment Builder and AI Auto-Scoring are named on the site
- EU hosting on AWS and ISO 27001 listed on the trust center
- Plan caps (assessments and users) are stated for each tier
Weak on
- No plan prices are published on the pricing page, only usage caps per tier
- Legal entity and headquarters country are not stated on the vendor pages
- Privacy policy language on retention and training could not be read
Skip it if: Buyers who need a named legal entity and a readable data processing policy before running a pilot. Four tiers (Individual, Team, Pro, Enterprise) show usage caps but no prices; start free with no card.
- 7
eSkill is seventh. Plan sizes are clear (500 and 1,000 candidates a year), a SOC 2 Type II trust center exists and the Linea tools turn a job title into a test. Prices stay behind a sales call, storage is US only, and the AI claims come with no model details, so it ranks below tools that print more.
Strong on
- Candidate caps are stated per plan, so volume fit is easy to judge
- SOC 2 Type II listed on a public trust center
- Linea Build creates a test from a job title using the question library
Weak on
- No prices are published and an annual subscription is required
- One-way interviews and custom questions begin on the Scale plan
- Storage is in the United States with no EU option stated
Skip it if: European employers with data residency rules and teams that expect more than 1,000 candidates a year on the entry tier. Growth, Scale and Custom plans all say Talk to sales; annual subscription required per the FAQ.
- 8
Vervoe ranks eighth. Explainable AI grading and generative AI detection are the right features for answer-based assessment, and data can be stored in Australia, the United States or Europe. But the pricing page assigns nothing to a plan, certifications appear only as footer badges, and the vendor does not state what happens to answers for training.
Strong on
- Explainable AI grading and generative AI detection named on the pricing page
- Personal data stored in Australia, the United States or Europe
- API and white-label options exist for enterprise customers
Weak on
- No plan tiers or prices are published, and enterprise features are not tied to a plan
- ISO 27001 appears only as a footer badge, not in the policy text
- Whether customer answers train models is not stated
Skip it if: Small teams that cannot afford a sales cycle and buyers who need certifications confirmed in a document. No plans or prices published; the pricing page offers a quote request and a demo booking.
- 9
Mercer | Mettl is ninth. The scope is wide: AI-assisted proctoring, coding assessments, speech evaluation and a virtual assessment center, with human review of AI output stated in the privacy notice. Against that, no price or tier is shown, retention periods are not given and the brand and operating entity chain is hard to follow.
Strong on
- AI-assisted proctoring and an AI-powered secured exam browser are both named on the homepage
- A virtual assessment center covers competency-based simulations delivered fully online
- Privacy notice says all AI output is reviewed by a human
Weak on
- No prices or plan tiers are published, so cost cannot be estimated from the site
- Retention periods are not given in the privacy notice
- Hosting country is not stated, only AWS cloud hosting
Skip it if: Buyers who want to estimate cost from a web page or who need written retention periods before signing. No tiers or prices are published on the pages opened; pricing is by sales contact.
- 10
iMocha is tenth. The libraries are large and the AI side is broad (Skills Inference, AI Interviews, an agents registry), and the policy gives two retention periods. Hosting is Azure in Virginia with access from India, and no plan or price is shown. The mix of workforce analytics and hiring makes it better for skills programs than pure screening.
Strong on
- Policy states 12-month general retention and 3-month video proctoring retention
- Skills Inference and an AI-Readiness Index extend beyond hiring
- Library covers coding, cognitive, domain and communication skills
Weak on
- No prices or plans are published, and no pricing page is linked from the homepage
- Primary data center is in Virginia, with employee access from India
- Headquarters country and contracting entity are not stated
Skip it if: European buyers who want EU-only data handling and teams that need public pricing to build a business case. No pricing page is linked; the calls to action are Talk to an expert and Book a demo.
- 11
Alva Labs sits eleventh. The per-hire model with no seat fee is an unusual and fair structure, and the Agentic Hiring System drafts criteria, scores transcripts and leaves approval to people. Yet no price, hosting region or retention rule is stated, the privacy policy page returned an error, and the vendor's outcome figures were not verified.
Strong on
- Per-hire pricing with no platform fee and no per-seat licence charged
- AI suggested interview scores wait for hiring team approval
- Psychometric tests are described as independently certified by DNV, scope not named
Weak on
- No prices, hosting region or retention periods are stated on the pages opened
- Privacy policy page returned a 404 when we opened it
- Coding tests and structured interview kits are add-ons on the psychometrics plan
Skip it if: Companies that want to keep their current ATS and buyers needing documented data handling before a pilot. Priced per hire by annual volume, billed in arrears, with no platform fee or per-seat licence.
- 12
Harver ranks twelfth. Its assessment range is strong for volume roles (pymetrics games, situational judgement, spoken language) and the security page claims SOC 2 and ISO 27001. But the AI is described only in general terms, an AI PREVAIL product has no description, the SOC 2 type is unclear and no price or hosting region is stated.
Strong on
- Harver Assess covers personality, cognitive, situational judgement and language tests
- Retention stated: eight weeks after applications end, one year with consent
- pymetrics game-based assessments and a Volume Hiring product
Weak on
- No prices or plan tiers are published, only a demo request route
- AI features are described only in general terms
- Security page mixes SOC 2 Type 1 and Type 2 references
Skip it if: Teams that need to see specific AI scoring features and prices before booking a demo. No pricing page exists; the only calls to action are Request a Demo and an ROI calculator.
- 13
SHL closes the list. The catalog is deep (OPQ personality, Verify cognitive, a Global Skills Assessment, simulations) and an EU AI Act guidance page exists. The pages we read describe no AI scoring feature for candidates, only an AI support chat, and give no prices, hosting region or retention period. It is a testing catalog more than an AI product.
Strong on
- Large catalog including OPQ, Verify and situational judgement tests
- Global Skills Assessment measures 96 behavioral skills in 15 minutes
- EU AI Act guidance page and Data Privacy Framework certification
Weak on
- No prices are published, and the only purchase route is an online store link
- No candidate-facing AI scoring feature is described on the pages read
- Hosting region and retention periods are not stated
Skip it if: Buyers looking mainly for AI scoring, AI interviewing or transparent self-serve pricing. No prices on the catalog or homepage; the only purchase route is an online store link.
How these AI pre-employment assessment software were judged
How good is the AI in pre-employment assessments?
The AI shows up in four jobs: building a test from a job description, scoring open answers, running an interview, and detecting cheating. Canditech names an assessment builder and auto-scoring agents, Maki People lists a voice interviewer with confidence levels on hire recommendations, and Codility records every AI Assistant exchange for human review. Harver and SHL describe their AI in general terms, so a buyer learns little about what it scores. Ask each vendor for validity evidence on the role you hire for, then run a pilot with current employees and compare the scores with their manager ratings before candidates see anything.
What does AI assessment software cost?
Printed prices are the exception. As of November 2026, TestGorilla lists a free plan, Core from $215 a month and Plus from $520 a month per account, both billed annually. Bryq shows Pro at $69 a month on annual billing, and Codility starts at $800 a month for 25 invites or $4,800 a year. eSkill prints candidate caps of 500 and 1,000 a year but no price, and Alva Labs charges per hire by annual volume. Criteria, Vervoe, SHL, Harver, Maki People, Mercer Mettl and iMocha all require a sales conversation.
Where do candidate answers and recordings get stored?
Candidate data is personal data, often with video or webcam stills, so location and retention carry legal weight. TestGorilla states EEA servers and keeps camera stills and audio or video answers for 6 months. Bryq and Canditech name AWS hosting in the EU, Codility offers EU or US hosting and keeps proctoring recordings 30 days, and iMocha names Azure in Virginia with video proctoring data deleted after 3 months. eSkill states US storage. Harver, SHL and Alva Labs do not name a hosting region on the pages we read, so request it in writing.
How hard is it to switch assessment vendors?
The sticky part is not the candidate list but the history behind your hiring decisions. Scores tied to a vendor's own scales do not transfer, and custom tests built in one tool must be rebuilt in another. Contracts add friction: eSkill requires an annual subscription, Bryq Pro asks for a one-year commitment on its listed price, and TestGorilla bills annually by default with a monthly toggle. Programmatic access helps exports, and TestGorilla Plus, Codility Enterprise and Vervoe's enterprise offer an API. Before signing, ask how candidate results can be exported and how long they remain available after cancellation.
Which assessment tool fits your hiring volume and roles?
Match the tool to the roles. A small company hiring across many functions fits TestGorilla or Bryq. Engineering teams hiring developers fit Codility, with iMocha as a broader skills option. Large employers running thousands of applications a year look at Harver, SHL or Criteria, and teams that want a voice interviewer for screening look at Maki People. Alva Labs suits employers who want per-hire pricing and an AI agent running the process. eSkill suits mid-size teams whose volume stays under its stated candidate caps. Check whether you will be scoring candidates in New York City or the EU, where bias audit and AI Act rules apply.
How to choose an AI pre-employment assessment tool in three steps
- Define the roles and the volume first. Write down the five roles you hire most and how many candidates apply to each in a year. That number decides whether a self-serve library, a per-hire model or a sales-negotiated contract fits. eSkill caps plans at 500 and 1,000 candidates a year, Codility sells invite packs and Alva Labs prices per hire, so a volume estimate turns vague quotes into comparable figures. Include seasonal peaks, because invite packs and credit models behave differently when applications arrive in bursts rather than evenly through the year.
- Test the assessment on people you already employ. Ask three vendors for a pilot and have current employees in one role take the assessment, ideally a strong performer, an average one and a recent hire. Compare scores with what their managers say. If the tool cannot separate them, candidates will not be sorted well either. Ask each vendor what validity evidence exists for your job family and what adverse impact analysis it has run. Maki People cites a bias audit and Bryq references the EU AI Act, so request the documents rather than the badge.
- Get hosting, retention and AI-training answers in writing. Send every shortlisted vendor the same three questions: where are candidate answers stored, how long are recordings and scores kept, and is candidate data used to train any model. TestGorilla and Codility state retention periods on their policy pages, while several others state none. Ask who the data controller is for your account, because Harver and Mercer Mettl name different group entities in different documents. Keep the replies with your contract file so a later audit or candidate request can be answered quickly.
Frequently asked questions
What is the best AI pre-employment assessment software in 2026?
TestGorilla ranks first here because it prints plan prices, offers a free plan, states EEA hosting and says personal data is not used to train generative models. Codility is the pick for developer hiring, and Bryq suits small teams that want a stated $69 monthly entry price.
Is AI pre-employment testing legal in the EU and New York City?
Rules apply to both. The EU AI Act treats employment screening AI as high risk, and New York City Local Law 144 requires a bias audit for automated employment decision tools. Maki People cites an audit by Holistic AI and Bryq references both frameworks. Confirm what each vendor will provide for your own use, and take legal advice before relying on a badge.
Which AI assessment tools publish their prices?
Only a few. TestGorilla prints free, Core and Plus pricing, Bryq prints Pro at $69 a month, and Codility prints online plans from $800 a month or $4,800 a year. Alva Labs states a per-hire model without figures, and eSkill states candidate caps without prices. Criteria, Vervoe, SHL, Harver, Maki People, Mercer Mettl, iMocha and Canditech show no figures.
Can candidates cheat with ChatGPT on these assessments?
They can try, which is why vendors sell integrity controls. Vervoe lists generative AI detection, Mettl names AI-assisted proctoring, TestGorilla Plus adds ID verification and Codility records behavioral signals. Canditech and Codility go further and let candidates use AI tools inside the test, scoring how they work with them. Decide which approach matches the job before choosing.
Do I need an applicant tracking system to use an assessment tool?
No, but integration saves manual work. Bryq and Canditech describe ATS integrations on their pages, Maki People names Greenhouse, iCIMS, Workday and others, and TestGorilla offers ATS and API connections on Plus. Alva Labs sells a full hiring system that replaces an ATS. Check which integrations your plan tier includes, because several vendors reserve them for higher tiers.
How long do vendors keep candidate data?
It varies widely. TestGorilla keeps scores and signals for 2 years and camera stills and recordings for 6 months, Codility keeps proctoring recordings 30 days and iMocha deletes video proctoring data after 3 months. Harver states up to eight weeks after an application ends. Many others, including Alva Labs and Bryq, state no periods on the pages we read.
Are these tools suitable for high-volume hourly hiring?
Some are built for it. Harver sells a Volume Hiring product, Maki People runs voice screening in 45-plus languages around the clock and Criteria bundles assessment with video interviews. TestGorilla and Bryq fit lower volumes with a recruiter reviewing results. Ask each vendor how pricing scales when applications reach the thousands, since per-invite and per-hire models behave differently.
How should a small company start with AI assessments?
Start with a free or low-cost plan and one role. TestGorilla has a free plan with 10 credits a month, and Bryq starts free with no card. Run the assessment next to your normal interview for a few hires, compare outcomes, and only then pay. Avoid annual contracts until you have seen how candidates react to the test length.