The Future of Professional Standards in an AI World

Three posts this week have approached the same underlying question from different angles.

The first looked at universities making bold public commitments to graduate AI-fluent professionals and asked who is actually verifying those claims. The second looked at organisations investing heavily in AI training, measuring completion rates, and calling it competency assessment and asked whether a HR platform tracking module completion has anything meaningful to say about professional capability. A third, from my colleague Robert Kay came at it from a different direction entirely asking whether, in training a generation to direct AI rather than do the work, we risk producing professionals who lack the depth to know when the AI gets it wrong.

All three circled the same gap. Today I want to give it a name and suggest where the solution has to come from — the emerging challenge of AI professional standards.

Across every profession where AI is fast becoming operationalised as practice, the role is changing faster than the profession's ability to govern it. The institutions charged with defining and assuring professional competency the bodies whose credential is supposed to mean something are being asked to certify capability for work that didn't exist in its current form two years ago.

That is not a criticism. It is a description of a problem that belongs to the profession, not the firms deploying the technology.

Three professions. The same gap.

In law, Allen & Overy has rolled out a series of AI agents handling antitrust filing analyses, cybersecurity incident responses, fund formation workflows, and loan agreement reviews each capable of conducting research, breaking down complex issues, and producing complete work product with minimal lawyer oversight.¹ The junior lawyer's traditional function learning the profession through volume, repetition, and supervised exposure to complex work is being compressed. The question of what a qualified solicitor or barrister now needs to demonstrate to justify that title is genuinely open. The Solicitors Regulation Authority has approved the first fully AI-powered law firm in the UK.² No equivalent framework exists for assessing whether the humans overseeing those systems are competent to do so.

In financial services, FINRA's 2026 Regulatory Oversight Report identifies AI agents operating in financial markets as carrying risks including autonomy without human validation, scope creep beyond intended authority, and multi-step reasoning that makes outcomes difficult to trace or explain.³ The regulator has named the risk. What it has not resolved and cannot resolve alone is what competency looks like for the financial adviser, the compliance officer, or the fund manager now responsible for overseeing those systems on behalf of clients. The fiduciary duty has not changed. The work environment in which it must be discharged has changed completely.

In accounting, KPMG's global AI workforce lead has stated publicly that the firm wants its junior consultants to become managers of agents not preparers or checkers, but supervisors of AI systems handling the analytical and administrative work that once defined entry-level practice.⁴ Graduate intake has fallen sharply as agentic AI absorbs the work that traditionally formed the basis of professional learning. The role being credentialled is no longer the role being performed.

The structural problem

Professional bodies exist precisely for this moment. They are the institutions with the domain knowledge to define what competency looks like in practice not generically, but in the specific contexts where their members work and where the public relies on their judgment. They are the institutions with the authority to make a credential mean something that employers, regulators, and clients can rely on.

But that authority rests on a foundation that is under quiet pressure and not only for the reasons most people are discussing.

The pyramid model has been the informal learning infrastructure of most regulated professions for generations. Large cohorts of juniors, learning through supervised exposure to volume work, developing the judgment that competency frameworks subsequently tried to describe. But the pyramid is also the revenue model.

Membership fees, examination fees, CPD requirements the income that sustains a professional body flows disproportionately from the base. If agentic AI compresses entry-level hiring across law, accounting, and financial services simultaneously, the bodies face a structural financial problem at exactly the moment they are being asked to invest in new competency frameworks. Fewer juniors means fewer members in formation. Fewer members in formation means less income to fund the standard-setting work the profession most needs right now.

Courts are already catching fabricated AI citations in legal filings at a rate of four or five new documented cases per day, up from 120 total cases between April 2023 and May 2025 to 660 by December 2025.⁵ These are not cases of AI being used irresponsibly by bad actors. They are cases of qualified professionals failing to exercise the kind of critical oversight their designation is supposed to guarantee. The credential said they could. The work said otherwise.

That gap between what the credential asserts and what the practice requires is widening in every regulated profession simultaneously. And it is widening faster than any standard-setting cycle has moved in living memory.

The window

The firms are moving. The technology is moving. Universities are making public commitments to graduate AI-fluent cohorts on timelines that run well ahead of any accreditation review cycle.

The professional body that defines what AI competency looks like within its domain rigorously, in a way that reflects the actual demands of contemporary practice will hold something genuinely valuable. A standard the market comes to trust. A credential that assumes legitimacy with regulators, employers, and the professionals who hold it.

But legitimacy is not permanent. It is renewed every time the market decides a credential still means something worth relying on. The body that moves slowly cedes that decision to others. Firm-issued credentials, platform certifications, and regulatory frameworks are already filling the space. As they do, employers begin to preference what those alternatives signal. Candidates follow. Membership at the base thins. Revenue contracts. The capacity to invest in the standard-setting work that would restore relevance diminishes at exactly the moment it is most needed.

The body that waits for consensus will find the standard has been set without it. By the firms. By the platforms. By whoever moved first. And by the time that becomes undeniable, it may be too late to do anything about it.

What we are working on

This week we explored the topic of AI skills from several different perspectives: universities, employers, and professional bodies. But we never forget the person in the middle of it all. The learner, wanting to invest in their future and build a resilient career.

At Ripltec we are building the infrastructure to help professional bodies and the professionals they represent operationalise this kind of standard whatever form it takes, whatever the domain requires.

We are not here to tell professional bodies what competency should look like in an AI-shaped practice environment. That is their domain and their authority. We are building the tools to help them define it, assess it at scale, and issue verifiable credentials that hold up to scrutiny. Not the ones that measure participation at a CPD lunch.

We are looking for professional bodies who want to work through this problem, not as recipients of a product roadmap, but as partners in building something their profession can actually rely on.

The image we chose for this article is a framed credential on a wall.

Not a badge in an inbox. Not a completion certificate from a platform. Something earned. Something that carries a standard behind it. Something worth displaying.

That is what a genuine AI competency credential could be. We are building the infrastructure to help professional bodies create exactly that, on their terms, within their domain, under their authority.

If you lead a professional body and anything in this article resonates, I would love to have that conversation.

Mark Runnalls is CEO of Ripltec , which builds AI fluency assessment infrastructure for professional credentialling and workforce development.

Part 3 of 3.

¹ Allen & Overy Shearman, April 2025, cited in Integrated Cognition, January 2026 ² Garfield AI, cited in Lupl, November 2025 ³ FINRA 2026 Regulatory Oversight Report ⁴ Niale Cleobury, Global AI Workforce Lead, KPMG, Business Insider, November 2025 ⁵ Artificial Lawyer, January 2026