In short: professional rankings use the same five-level scale, scoring rules and statistics as our academic rankings, with indicators built for training providers, credentials and firms. Each ranking applies the indicators of its track, weighted for its field. The people behind an entity count only through what they bring to it, and no individual is ever profiled. This page is the complete framework used by our assessors.
How to score
| Level | Name | Points |
|---|---|---|
| 1 | Absent | 0 |
| 2 | Emerging | 25 |
| 3 | Established | 50 |
| 4 | Advanced | 75 |
| 5 | Leading | 100 |
Indicator points are combined into pillar scores, and pillar scores into the overall score, each weighted for the ranking's field and level.
- Score the evidence, not the claim. Marketing pages count only when corroborated by registers, documentation, client or partner publications, independent reporting or data.
- Award a level only when all of its description is met. If an entry sits between two levels, award the lower one.
- Cite at least one evidence item for every level you award, by its ID in the evidence file.
- Use "Insufficient evidence" only after checking the evidence file and finding nothing relevant. It is never a neutral default, and it is never imputed above Established.
- Read the note for the ranking's track (provider, credential or firm) where one is given. It overrides the general reading.
- Judge the people behind an entity only as far as they shape what it delivers. Assess the team as a whole: never profile individuals, and never record personal details beyond professional standing.
- Evidence should be no older than 36 months.
- Score alone. Do not discuss an entry with other assessors before you submit.
- Record your confidence honestly. Low confidence widens the published band; it does not lower the score.
Confidence
| Confidence | Meaning |
|---|---|
| High | Two or more independent sources agree, or one authoritative primary source (program regulations, audited data) settles the question. |
| Medium | One good primary source, or several sources with minor gaps or inconsistencies. |
| Low | Indirect, partial or dated evidence; the level is a reasoned judgment rather than a clear reading. |
When two assessors' scores for a pillar differ by more than 15 points, a senior assessor decides. An entry with more than 25% of its indicators insufficiently evidenced is not ranked.
Product & Services
What does it actually deliver, and how good is it?
OFR.1 Depth and quality of the offering
What the entity actually delivers: the substance, technical depth and quality of its products, services or programs, judged against current professional practice in AI.
What does it actually deliver, and how good is it?
| Level | Description |
|---|---|
| 1 Absent | What is delivered cannot be established, or it is generic content with no substance beyond marketing claims. |
| 2 Emerging | A described offering of introductory depth: awareness-level content or standard services, with little evidence of method or technical substance. |
| 3 Established | A substantial offering with documented scope and methods, sound against current practice, covering the essentials of its field at a professional level. |
| 4 Advanced | An offering of clear depth and quality: advanced, hands-on and specific to real professional use, with documented methods and the outcomes it is designed to achieve. |
| 5 Leading | A reference offering in its field: advanced and comprehensive, technically rigorous, demonstrably current, and recognized by practitioners as a standard others follow. |
By track
- Training providers: organizations that train professionals and teams: Read as the curriculum: learning outcomes, sequence, hands-on work with current AI tools, and depth beyond awareness training.
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as the body of knowledge the credential certifies: the published blueprint or competency framework, its scope and its depth.
- Firms and ventures: agencies, consultancies and companies that deliver AI products or services: Read as the services or products delivered: technical depth, methods, and what is built, deployed or advised in practice.
Evidence to look for
- Course outlines, exam blueprints or service descriptions
- Product and technical documentation
- Independent reviews and professional assessments
OFR.2 Verifiable examples of work
Public, checkable examples of what the entity delivers: case studies, portfolios, demos, sample materials or published assessments.
Can what it delivers be seen and checked?
| Level | Description |
|---|---|
| 1 Absent | No examples of work are available, or the examples cannot be traced to real deliveries. |
| 2 Emerging | A few examples, mostly self-described, without enough detail to check what was delivered. |
| 3 Established | Several documented examples with enough detail (scope, approach, result) to check what was delivered, at least one corroborated by a client, partner or independent source. |
| 4 Advanced | A broad set of documented, recent examples, most of them corroborated independently, showing consistent quality across deliveries. |
| 5 Leading | An extensive, open record of work: detailed and independently corroborated examples, open materials or code, and results others can inspect and reuse. |
By track
- Training providers: organizations that train professionals and teams: Count sample lessons, open course materials, learner projects and published syllabi.
- Professional credentials: certifications and certificates, assessed with their issuing body: Count published sample questions, practical tasks, exam blueprints and technical reports on the assessment.
- Firms and ventures: agencies, consultancies and companies that deliver AI products or services: Count named case studies, client publications, demos and open-source work. Confidential work counts only where its existence and scope are independently corroborated.
Evidence to look for
- Case studies and portfolios
- Client and partner publications
- Sample course materials, demos or open repositories
- Published sample assessments
OFR.3 Currency and renewal
How current the offering is, and how systematically it is renewed as AI changes.
Does the offering keep pace with the field?
| Level | Description |
|---|---|
| 1 Absent | No evidence of updates in the last 36 months; content or methods are visibly out of date. |
| 2 Emerging | Occasional updates without a stated cycle; parts of the offering lag current practice. |
| 3 Established | Updated within the last 24 months, covering the main recent developments in its field. |
| 4 Advanced | A stated renewal cycle of 12 months or less, with documented recent updates that track current tools, models and regulation. |
| 5 Leading | Continuous renewal: changes in the field reach the offering within months, versions and change logs are published, and the entity contributes to the developments it teaches or applies. |
By track
- Professional credentials: certifications and certificates, assessed with their issuing body: Include recertification or continuing-education requirements, and how often the blueprint is revised.
Evidence to look for
- Dated versions and change logs
- Announcements of updates
- Archived pages compared over time
OFR.4 Quality assurance
The checks that make what is delivered mean what it claims: assessment of learners, validity of credentials, and quality, security and data-protection controls in delivery.
Does what it delivers mean what it claims?
| Level | Description |
|---|---|
| 1 Absent | No quality assurance can be identified. |
| 2 Emerging | Informal quality checks, or assessment that only confirms attendance or completion. |
| 3 Established | Documented quality assurance: assessed outcomes or acceptance criteria, stated standards, and a process for complaints and corrections. |
| 4 Advanced | Quality assurance to recognized standards, applied consistently, with independent elements such as external review, proctoring or audit. |
| 5 Leading | Independently certified or audited quality assurance (for example ISO/IEC 17024 for certification bodies, ISO/IEC 42001 or 27001 for firms), with published results and evidence that it is acted on. |
By track
- Training providers: organizations that train professionals and teams: Read as assessment of learners: assessed practical work rather than attendance, and certificates that state what was assessed.
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as assessment validity: proctored or supervised exams or practical assessment, documented pass standards, psychometric review and identity checks. Level 3 requires assessment beyond course completion.
- Firms and ventures: agencies, consultancies and companies that deliver AI products or services: Read as delivery quality: methodology, testing and evaluation of AI systems, security and data protection, and responsible-AI controls in projects.
Evidence to look for
- Assessment and certification policies
- Quality, security and AI-management certifications
- Audit and external review reports
People & Expertise
Are the people behind it qualified to deliver it?
PPL.1 Leadership expertise
The relevant skills, professional experience, education and record in the field of the people who lead the entity, assessed as a whole.
Is the entity led by people with real expertise in what it delivers?
| Level | Description |
|---|---|
| 1 Absent | Leadership is not identifiable, or its relevant expertise cannot be established. |
| 2 Emerging | Leadership is identified, with general business experience but little verifiable expertise in AI or in what the entity delivers. |
| 3 Established | Leadership combines relevant education and several years of verifiable professional experience in the entity's field, including work with AI. |
| 4 Advanced | Leadership with deep, verifiable expertise in the field: senior experience, advanced education or equivalent professional standing, and a public record of relevant work. |
| 5 Leading | Leadership recognized as authorities in the field: a sustained record of contributions, senior roles and recognition by peers, directly relevant to what the entity delivers. |
By track
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as the leadership of the issuing body and of the credential program.
Evidence to look for
- Leadership pages and official filings
- Professional records corroborated by employers or institutions
- Publications, talks and recognized contributions
PPL.2 Depth of the team
The qualifications and relevant experience of the people who deliver the work, and their share of it.
Who actually does the work, and are they qualified to do it?
| Level | Description |
|---|---|
| 1 Absent | The people who deliver are not disclosed. |
| 2 Emerging | Some delivery staff are named, with limited or unverifiable qualifications. |
| 3 Established | Most delivery staff are identified, with relevant qualifications and professional experience in AI or the entity's field. |
| 4 Advanced | A strong team: nearly all delivery staff have verified, relevant qualifications and practical experience, with depth across the areas the entity covers. |
| 5 Leading | An exceptional team: recognized practitioners or researchers deliver the core of the work, with documented selection, training and quality standards for everyone who delivers. |
By track
- Training providers: organizations that train professionals and teams: Read as the instructors and mentors who teach, and the share of teaching they deliver.
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as the subject-matter experts and psychometric specialists who design, review and maintain the assessment.
- Firms and ventures: agencies, consultancies and companies that deliver AI products or services: Read as the consultants, engineers and researchers who deliver client work.
Evidence to look for
- Team, instructor and examiner pages
- Subject-matter expert panels
- Professional records and publications
PPL.3 Contribution to the field
Public contributions by the entity and its people that advance the field: research, open-source work, standards, teaching and talks.
Does the entity add to the field, not only draw from it?
| Level | Description |
|---|---|
| 1 Absent | No identifiable contribution to the field. |
| 2 Emerging | Occasional blog posts, webinars or promotional content, without substantive contribution. |
| 3 Established | Regular substantive contributions: technical writing, talks at recognized events, open materials or open-source code. |
| 4 Advanced | Recognized contributions: peer-reviewed or widely cited work, maintained open-source projects with real use, or participation in standards and professional bodies. |
| 5 Leading | Leadership in the field: influential research, widely adopted open work, or a formative role in standards, frameworks or professional practice. |
Evidence to look for
- Publications and citations (OpenAlex)
- Open-source repositories and their use
- Standards bodies and professional associations
- Programs of recognized events
Track Record
What has it delivered, for whom, and with what results?
REC.1 Time and continuity in operation
How long the entity has operated in its field, and the continuity and scale of its activity.
Has it done this long enough, and steadily enough, to judge?
| Level | Description |
|---|---|
| 1 Absent | Less than a year of verifiable operation in the field, or activity that cannot be established. |
| 2 Emerging | One to two years of verifiable operation, or a longer history with major interruptions or changes of business. |
| 3 Established | At least three years of continuous, verifiable operation in the field, at a steady scale. |
| 4 Advanced | Five years or more of continuous operation, with documented growth or stable scale and a consistent focus. |
| 5 Leading | A long, continuous record in the field (typically eight years or more, or since the field emerged), with sustained scale and a consistent reputation. |
By track
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as the history of the credential and its issuing body: first award, revisions, and holders over time.
Evidence to look for
- Company register filings and incorporation dates
- Archived versions of the official website
- Annual reports
REC.2 Clients and outcomes
Verifiable clients, learners or holders, and the documented outcomes of the entity's work.
Who has it served, and what came of it?
| Level | Description |
|---|---|
| 1 Absent | No verifiable clients, learners or outcomes. |
| 2 Emerging | Clients or learner numbers are claimed but not verifiable; outcomes are anecdotal. |
| 3 Established | Named clients or verifiable learner and holder numbers, with some documented outcomes. |
| 4 Advanced | A substantial, verifiable base of clients or learners, with outcomes documented with method (for example completion and placement data, or measured client results), several of them corroborated independently. |
| 5 Leading | An extensive, verifiable record of outcomes over time, independently corroborated, showing consistent results and recognition of the work or credential by employers and industry. |
By track
- Training providers: organizations that train professionals and teams: Read as learners trained, completion, and what learners or client organizations achieved afterwards.
- Professional credentials: certifications and certificates, assessed with their issuing body: Read as the number of holders, recognition by employers and demand for the credential in job postings, and evidence that holders perform as certified.
- Firms and ventures: agencies, consultancies and companies that deliver AI products or services: Read as named clients, delivered projects in production, and measured results for clients.
Evidence to look for
- Client and partner publications
- Outcome reports with their method
- Credential registries and holder counts
- Job postings that require the credential
REC.3 Claims versus evidence
The consistency between what the entity claims about itself and what can be shown.
Does what it says match what can be verified?
| Level | Description |
|---|---|
| 1 Absent | Material claims are contradicted by evidence, for example invented clients, partnerships, recognitions or results. |
| 2 Emerging | Several material claims cannot be verified, or are overstated against the evidence. |
| 3 Established | The main claims can be verified; minor claims are unverified but plausible. |
| 4 Advanced | Nearly all claims are verifiable and accurately stated, including figures and partnerships. |
| 5 Leading | Full consistency: claims are precise, sourced and independently corroborated, and the entity publishes the evidence behind them. |
Evidence to look for
- Claims on the official website compared with registers and client publications
- Partner and certification directories
- Independent reporting
Standing & Conduct
Is it accountable, recognized and responsible?
STD.1 Legal and organizational standing
Transparent legal existence, ownership and governance, beyond the minimum the verification gate requires.
Is it clear who the entity is, who owns it and who is accountable?
| Level | Description |
|---|---|
| 1 Absent | Legal existence or accountable leadership cannot be confirmed. Such entities do not pass the verification gate. |
| 2 Emerging | Registered and with identified leadership, but ownership, address or terms of service are unclear. |
| 3 Established | Registered, with identified leadership, a real address, published terms and a clear point of contact for complaints. |
| 4 Advanced | Transparent ownership and governance: published company details, leadership responsibilities, policies, and filings in good standing. |
| 5 Leading | Full transparency and sound governance: published ownership, independent oversight where relevant (for example a board or advisory council with external members), and audited or published accounts. |
Evidence to look for
- Company register records and filings
- Terms of service and imprint pages
- Governance and ownership disclosures
STD.2 Professional recognition
Recognition by credible professional bodies, accreditors, platforms and peers, weighted by the standing of each.
Do credible bodies in the field recognize it?
| Level | Description |
|---|---|
| 1 Absent | No recognition, or only from bodies of unclear standing (paid listings, awards without a method). |
| 2 Emerging | Memberships or listings that require little scrutiny. |
| 3 Established | At least one recognition that requires meeting stated criteria: a vetted partner status, an accreditation or a credible award with a published method. |
| 4 Advanced | Several criteria-based recognitions from credible bodies, including at least one independent accreditation or high-tier vetted partnership. |
| 5 Leading | Recognition at the highest level of the field: accreditation against international standards, top-tier partnerships with leading platforms, and standing as a reference among peers. |
By track
- Professional credentials: certifications and certificates, assessed with their issuing body: Recognition of the credential itself counts: accreditation of the certification program (for example under ISO/IEC 17024), endorsement by professional bodies, and recognition by employers or regulators.
Evidence to look for
- Accreditor and certification-body registers (for example ANAB or UKAS)
- Partner directories of AI platforms
- Professional bodies, and award records with their methods
STD.3 Responsible practice and conduct
Published commitments on responsible AI, data protection and fair dealing, how they are applied, and any reported violations or problems.
Does it work responsibly, and what is on the record?
| Level | Description |
|---|---|
| 1 Absent | Serious reported problems (regulatory findings, sanctions, fraud findings or well-founded complaints) that remain unresolved. Disqualifying problems exclude an entity at the verification gate. |
| 2 Emerging | No responsible-practice commitments published, or reported problems that were only partly addressed. |
| 3 Established | Published policies on data protection and responsible use of AI, fair terms, and no unresolved reports of serious problems. |
| 4 Advanced | Responsible practice applied and visible in delivery (risk assessment, transparency to clients or learners, data-protection controls), with a clean record or problems resolved openly. |
| 5 Leading | A reference for responsible practice: independently audited controls, public reporting on its own conduct, and a clean record over time. |
Evidence to look for
- Responsible-AI, privacy and security policies
- Regulatory, court and consumer-protection records
- Independent reporting
External Reference
What do established, permissibly usable sources say?
EXT.1 External reference signal
A normalized 0 to 100 value from established sources that may lawfully be used: open-licensed data, data used with permission or under licence, or limited factual references consistent with the publisher's terms. Computed by the data team, not scored by assessors.
What do established, permissibly usable sources say?
Evidence to look for
- CWTS Leiden Ranking Open Edition (CC0)
- Licensed or permitted data, listed in the edition notes
Assessor toolkit
The same rubric drives these tools, so what assessors see is exactly what is published here.
- Scoring form: Offline form for assessors. Open the evidence file, score each indicator, export your scoring file.
- Evidence file template: Structure of the evidence file prepared for each entry.
- Evidence log (spreadsheet): Spreadsheet version of the evidence items, convertible into the evidence file.
- External signal template: How the data team records the capped external reference signal.