The AI conversation in credentialing has become curiously binary. Either a platform is “AI-powered” and AI is woven into every decision whether you want it there or not, or the platform is “traditional” and offers nothing modern. Both framings are wrong, and both make life harder for the people who actually have to make procurement decisions in regulated environments.
“Real credentialing organisations do not want to be told whether to use AI. They want to choose.”
Real credentialing organisations do not want to be told whether to use AI. They want to choose. They want to use AI where it adds value, leave it out where it does not, and have that choice available on every assessment, configured by people who understand the stakes. That is the design principle behind Assess for Learning, and it is one of the things we are most consistent about. The platform is a fully fledged assessment platform that can be used without AI at all if that is the operating environment you want.
Why “AI Optional” Is Not a Compromise
There is an assumption in some quarters that a platform which does not force AI on users is somehow behind the curve. The opposite is true. Building a credentialing platform that gives you genuine choice over AI involvement is harder than building one that hard-codes AI into the workflow, because every feature has to work both with and without AI assistance, and the governance model has to accommodate both modes.
Assess for Learning is built on that harder principle, and the practical result is a platform that respects the operational reality of credentialing. Some assessments are high stakes and benefit from human-only grading. Some are high volume and benefit from AI assistance to keep cycles manageable. Some sit in regulated environments where AI use is restricted by policy. Some sit in markets where AI use is expected. The platform handles all of these, on the same infrastructure, without forcing a single approach.
The Range of Configurations You Actually Get
When you set up an assessment, you choose the grading model from a set of configurations that cover the realistic spectrum of credentialing needs.
Seven grading configurations
- Single grader — one human grades each submission, with no AI involvement at all
- Sequential graders — a second human picks up if the first cannot resolve the submission
- Double grading — two humans grade independently for calibration or moderation
- Single or double grading with optional copilot — the human can call on AI assistance during grading if they want it, but is not required to
- AI as an additional grader inside a multi-grader workflow, alongside humans
- AI plus self grade — AI provides an initial grading and the candidate reviews and responds
- AI only — for low-stakes contexts where AI grading is appropriate
Note that several of these configurations involve no AI at all in the grading decision, and several leave AI assistance entirely to the discretion of the human grader on a submission-by-submission basis. This is not a marketing claim. It is how the platform is built.
Why This Matters at the Procurement Level
For C-suite and operational leadership evaluating assessment platforms, the AI question is often the hardest part of the procurement. Boards want assurance that AI is not making consequential decisions without human oversight. Standards bodies want evidence that AI involvement is governed and traceable. Some markets, particularly in regulated education and professional certification, simply do not permit AI grading at all for certain assessment types. Buyers need a platform that handles all of this without becoming a different platform every time the rules change.
“The platform is a fully fledged assessment platform that can be used without AI at all if that is the operating environment you want.”
Assess for Learning solves this by making AI involvement a configuration choice, not a fundamental property of the system. The same platform can run a fully manual high-stakes professional certification, a high-volume CPD assessment with AI copilot support, and a low-stakes formative assessment graded entirely by AI. The governance evidence is consistent. The reporting is consistent. The user experience for graders is consistent. What changes is the role AI plays in any given assessment.
That has real procurement consequences:
- a single platform decision covers the full range of your assessment portfolio, rather than requiring different tools for different stakes
- regulatory changes can be absorbed by reconfiguring assessments rather than replacing infrastructure
- internal policy debates about AI use become productive conversations about which configuration to use, not blocking objections to the platform itself
- subject matter experts who are sceptical of AI can grade entirely without it, while those who find AI useful can opt in
- the organisation can move along the AI adoption curve at its own pace, on its own terms
Governance Built Into the Configuration
The AI optional principle is reinforced by the platform’s governance model, which sits at three levels and is selected when you configure each assessment. The governance level controls many things, including the psychometric thresholds applied, the standards alignment of the precision report, and the AI involvement permitted. A high-governance configuration locks down AI involvement appropriately. A lower-governance configuration opens up more flexibility for lower-stakes work. The governance choice and the AI choice are made together, by the people who understand the assessment.
This is the right design. It puts the people running the credentialing programme in control of the trade-offs, rather than baking those trade-offs into the platform. Every assessment becomes a conscious decision about how much AI to use, governed by the standards the organisation is held to, with full traceability of the choice and the reasoning. The companion piece on the grading copilot walks through how this works in the live grading flow.
From All-In or All-Out to Genuine Choice
“If you have been told that going modern means going all-in on AI, you have been told the wrong story.”
Most platforms offer two settings: AI on, or AI off. Assess for Learning offers a spectrum, configurable per assessment, governed by your standards model, transparent to your auditors. That is what AI optional actually means in a credentialing context, and it is the only design that survives contact with the real complexity of running an assessment programme in a regulated environment.
If you have been told that going modern means going all-in on AI, you have been told the wrong story. The right platform gives you the choice, every time, on every assessment, with the governance to back it up.
Ready to choose how AI fits into your credentialing programme, on your terms?
Talk to us about how Assess for Learning gives you genuine choice over AI involvement, configured per assessment.