JANSON Research & Strategy · Research Report · 2026

The Currency of Credentials

A provider-side study of quality, relevance, assessment, AI practice, and value in credential programs — and of the gap between what credentials claim and what they can demonstrate.
24provider programs
6findings
12exhibits
90day roadmap
In brief

Quality intent is outpacing quality infrastructure.

For a credential to retain credibility, its quality system must evolve alongside the work itself. That means aligning evidence to claimed competencies, protecting assessment integrity as artificial intelligence changes the environment, creating a signal employers can trust, and ensuring learners receive meaningful value in return.

The providers surveyed appear broadly aligned with that standard. They describe competency frameworks, formal standard-setting, human oversight, employer input, flexible delivery, and ambitions for richer assessment. What remains underdeveloped is the institutional machinery behind those practices: the systems needed to maintain standards, demonstrate quality, govern decisions, and measure outcomes consistently.

71%of formal-assessment programs use multiple-choice or selected-response-only exams
65%report no systematic tracking of what happens to credential holders
40%of clear AI users report a formal AI governance framework
Exhibit 1

Intent outpaces infrastructure across five operating domains

Across seemingly unrelated dimensions of credential quality, the same structural pattern repeats. Viewed together, these are not five independent quality problems, but manifestations of the same underlying capacity gap: the distance between understanding what quality requires and building the systems needed to sustain it.

Domain
What providers report strongly
Where the infrastructure lags
Currency
Competencies reflect current work
88%
Systematic labor-market monitoring
52%
Competence
Formal cut-score methods in use
80%
Assessment beyond selected response
29%
AI practice
Human confirms final decisions
100%
Formal AI governance framework
40%
Value
Employers recognize the credential
91%
Systematic holder-outcome tracking
35%
Equity
Flexible delivery widens access
83%
Participation / completion monitoring
46%
Provider track, n = 24, with item-specific denominators. AI practice items are limited to the 10 clear AI users. Figures are rounded; exact bases appear in the report appendix.
The findings

Six findings, read in sequence

Finding 01 · Currency

Confidence in current relevance is high; verification mechanisms are less consistent

87.5% of respondents agree that program competencies reflect the work people in the role actually perform today. Formal maintenance practices are less consistent: only half report using a systematic process to monitor labor-market change, and four in ten programs were last substantially updated three or more years ago.

Each measure addresses a different aspect of relevance. Across the section a consistent pattern emerges: confidence in relevance is strong, while the mechanisms used to verify, document, and maintain it are less consistently developed.

Exhibit 2 · Relevance versus the systems that maintain it
Competencies reflect current role
87.5%
Systematic labor-market monitoring
52.4%
Content addresses AI-enabled tasks
50.0%
Content not reviewed for AI-influenced tasks
73.7%
Agreement or strong agreement on a four-point forced-choice scale. Denominators are item-specific (n = 16–24). The final item is negatively worded — agreement indicates a review gap.

“The single biggest barrier is securing sufficient subject-matter expert time to continuously identify, evaluate, and translate changes in practice into validated exam content.”

Credential provider — professional certification program
Finding 02 · Assessment

Selected-response testing remains the dominant evidence of competence

Applied competence is central to the credential claim, while the evidence base remains dominated by selected-response assessment. That does not make multiple-choice assessment weak: well-designed items can assess application, analysis, evaluation, and judgment, and 80% of programs with a formal assessment use a formal method to establish cut scores.

The issue is construct-to-evidence alignment. The strength of the credential signal ultimately depends on how closely the evidence matches the claim.

Exhibit 5 · Applied assessment is gaining ground, while richer methods remain distributed
MCQ / selected response only
70.6%
Performance tasks / simulations
35.3%
Portfolio / work sample
23.5%
Workplace verification
23.5%
Oral / viva
17.6%
Project-based
11.8%
Base: 17 provider programs reporting a formal assessment. Multi-select item, so shares sum to more than 100%. Accreditation cross-tabulation: 85.7% of accredited and 60.0% of unaccredited formal-assessment programs report MCQ-only. Groups are small (seven and ten programs), so the comparison is descriptive rather than causal.
Finding 03 · Assessment infrastructure

Assessment aspirations are advancing faster than implementation capacity

Providers can describe the assessment they would prefer. What they cannot always do is deliver it reliably and at scale. Every desired method meets a specific operational constraint: scenario items need trained scorers, portfolios need rubrics and moderation, observation is labor-intensive, video raises technology and privacy demands, and adaptive testing requires item-bank depth and psychometric cost.

The practical implication is not to replace multiple-choice items indiscriminately, but to use the least burdensome evidence that validly supports the claim.

“Would like to use more scenario-type questions that allow flexibility in the examinee’s approach to solve the problem, but don’t have anyone available to grade such an exam.”

Credential provider — professional certification program
Finding 04 · AI governance

Human supervision is leading the transition to formal AI governance

Every clear AI user answering the item confirms that final credential decisions are reviewed by a human. Formal governance is less mature: half report defined protocols, four in ten report a governance framework, and half report that informal individual staff use continues.

Providers are applying human review as AI use expands, even where formal policies remain incomplete. Governance is developing alongside practice rather than preceding it. The safeguards exist; what is missing is the written record that makes them auditable.

Exhibit 7 · Safeguards in practice versus governance on paper
Human confirms final decision
100%
Qualified SME reviews AI content
88.9%
Tools evaluated for accuracy and bias
87.5%
Defined protocols for AI use
50.0%
Formal AI governance framework
40.0%
Base: the 10 providers identified as clear AI users. Denominators vary by item. Figures describe reported practice, not independent audit.
Finding 05 · Credential value

Providers report strong market value without equivalent downstream outcome evidence

Confidence in external value is strikingly high. The outcome-evidence infrastructure is substantially less developed. The distinction is between market-signal evidence and longitudinal holder evidence: a provider may know employers want the credential without knowing whether holders advance, stay in the field, use the skills, or earn more after earning it.

Exhibit 9 · Confidence in credential value is high; evidence of what happens after is not
Helps employers close skills gaps
95.0%
Employers recognize and value it
90.9%
Appears in postings / promotion criteria
81.0%
Systematically tracks holder outcomes
34.8%
Provider track, item-specific denominators (n = 20–23). The tracking measure is the positive complement of the 65.2% who agree they do not systematically track credential-holder outcomes.
Finding 06 · Equity

Flexible access is more common than systematic equity measurement

Flexible scheduling, asynchronous options, remote delivery, and additional languages remove genuine barriers. The limitation is evidentiary: without monitoring, organizations cannot determine whether access improvements change who enters, persists, completes, or benefits afterward.

The goal is not a universal equity dashboard. It is a proportionate measurement system — identify the barrier the program is trying to reduce, collect the minimum data needed to test whether it is changing, and use privacy-safe thresholds.

Exhibit 10 · A 37-point gap between expanding access and measuring equitable participation
Flexible delivery increases equitable access
82.6%
Monitors participation and completion by group
45.8%
Provider track, n = 23 and 24 respectively. The gap does not imply access interventions are ineffective — only that most programs lack the evidence to know.
The signature finding
71%
of formal-assessment programs use multiple-choice or selected-response-only exams.
41%list selected response with no other competency-assessment method at all
73%agree their program relies solely on a single assessment format
Synthesis

The field has a clear view of quality. Infrastructure remains the constraint.

The qualitative data place operational capacity at the center of the implementation challenge. 70% of substantive barrier responses mention staffing, time, funding, or related resources; 14 of 20 answers about content currency name capacity or resource constraints directly.

The implication is one of ownership. A construct-to-evidence gap can originate within assessment design, while persistent inability to close it reflects a wider organizational challenge. Budget, governance, and accountability therefore become part of the quality system. The literature offers limited direct evidence on operating capacity as a cross-cutting constraint, so the capacity explanation should be treated as an empirically grounded hypothesis rather than an established result.

Recommendations

Build quality infrastructure that can be maintained, measured, and defended

Six operating priorities. They are not universal standards and do not replace accreditation requirements, psychometric standards, or sector-specific governance.

01

Build a formal process for maintaining credential currency

Named owner, defined external signals, review triggers, a maximum interval, and a documented path from detected change to revised competencies.

02

Align every competency claim with sufficient evidence

Map every competency claim to the evidence used to support it. Add richer evidence only where the construct requires it.

03

Formalize AI governance around established safeguards

Inventory current use, then convert existing human-review habits into approved uses, prohibited uses, validation criteria, and named accountability.

04

Measure the outcomes most closely tied to the claim

Follow holders six to twelve months after completion and collect a focused set of outcomes tied to the credential’s stated purpose.

05

Connect access strategies to measurable equity outcomes

Identify the barriers being reduced, then measure whether the intervention changes participation, completion, or pass rates.

06

Make quality infrastructure a recurring investment

Treat maintenance, psychometrics, SME review, outcome tracking, and governance as recurring operating functions rather than one-time projects.

Ninety-day roadmap

Sequence quality improvements around evidence generation

Not a transformation program. A disciplined baseline, a small number of design decisions, and one or two systems that begin producing evidence within ninety days.

Baseline diagnostic
Find out what you actually have
  • Inventory credential claims, current update cycle, and assessment methods
  • List every AI use case, including informal individual staff use
  • Identify what outcome, access, and equity data already exists
  • Name the three highest-risk gaps
Output: a current-state evidence map and a prioritized gap register.
Design minimum viable systems
Decide before you build
  • Define currency triggers, owners, and a maximum review interval
  • Complete the construct-to-evidence audit
  • Draft the AI use protocol and approve the accountable owner
  • Select three to five outcome measures tied to the value claim
Output: approved designs for currency, evidence, AI, outcomes, and access measurement.
Pilot and institutionalize
Produce the first evidence
  • Pilot one richer assessment element on the weakest claim
  • Launch the six-to-twelve month holder follow-up
  • Run the first documented currency review and log the decision
  • Establish a dashboard cadence and a decision forum
Output: a first evidence cycle, governance artifacts, and baseline metrics.
What happens next

The missing perspective is what credentials actually change for learners

This study was designed as a two-sided provider-and-learner investigation. Learner response volume was too small to support trustworthy aggregate claims, so those responses were removed from the primary findings rather than stretched into percentages.

That limitation points directly at the most important next study: a learner-first investigation focused not on satisfaction but on real-world outcomes. What does earning a credential actually change in a person’s work and life? Selected individual cases will continue to be published as JANSON Research & Strategy Field Notes.

Research notes

About the study

An exploratory provider-side analysis. Findings describe this sample and identify patterns worth testing in larger research. Percentages should not be read as population estimates.

24 provider-track respondents, collected 15 July – 24 August 2026 from 29 total submissions, of which 2 were ineligible. The sample is concentrated in professional associations and third-party professional certification. Recruitment used purposive outreach and chain-referral sampling — not probability sampling, and not nationally representative.
An online survey with routing, multi-select items, four-point forced-choice agreement items, reverse-coded statements, and open-response prompts. No neutral midpoint was offered. Strongly disagree = 1 through strongly agree = 4; reverse-worded items scored as 5 minus the raw value. “Not applicable” and “not sure” responses are unscored and retained separately.
Item-specific denominators differ because of survey routing, “not applicable” responses, uncertainty, and missing data. AI governance findings apply only to the 10 clear AI users. Assessment findings use the 17 programs reporting a formal assessment. Every exhibit states its own base.
Three learner-track respondents and one Phase 2 follow-up. Not numerous enough to support trustworthy aggregate findings, and therefore excluded from the primary synthesis. Retained only as potential case material for Field Notes, clearly labelled as illustrative rather than generalizable.
Participation was voluntary and informed. Contact data and response data were held in structurally separate systems with no shared identifier, timestamp, or link. Quotations are attributed only by provider category, with minor typographic corrections that do not alter meaning.
Read the full report

Forty-six pages, twelve exhibits, and a ninety-day roadmap.

The complete report carries the full evidence base, the construct-to-evidence argument, implementation risks, measures and decision points, and the technical appendix.

The Currency of Credentials. Provider-side findings, 2026.
Download the report