The Artificial Intelligence standard recognises people who can take a model from problem framing to production and stay accountable for what it does once it is there. It is assessed on shipped work, not on recall of architectures.
Each area is evidenced from your own work. Assessors look for what you decided and why, not for a rehearsed answer.
The same six-level framework, expressed in the terms of this discipline. Experienced applicants can enter directly at a higher level through the portfolio route.
| Level | Post-nominal | What it certifies | Professional Experience | Annual CPD |
|---|---|---|---|---|
| 1BIPS Foundation | BIPS-F | Understands model families, data requirements and the limits of automated decisions. | 0–1 yr | 10 h |
| 2BIPS Practitioner | BIPS-P | Trains and evaluates models on defined problems, with a senior reviewer on the work. | 1–3 yrs | 15 h |
| 3BIPS Professional | BIPS-Pro | Owns an AI feature end to end and is accountable for its behaviour in production. | 3–6 yrs | 20 h |
| 4BIPS Specialist | BIPS-S | Leads a specialism — safety, MLOps, applied research — with peer-reviewed evidence. | 6–10 yrs | 25 h |
| 5BIPS Expert | BIPS-E | Sets AI direction across an organisation and is cited beyond one employer. | 10+ yrs | 30 h |
| 6BIPS Fellow | FBIPS | Shapes the profession itself through standards, public work or sustained contribution. | By election | 30 h |
You submit real work against the six capability areas.
An assessor checks provenance and your role in it.
A recorded conversation about decisions and trade-offs.
Your credential is issued and listed publicly.
Builds and runs the models that carry live decisions.
Decides what gets automated, and what deliberately does not.
Holds the line on bias, explainability and regulation.
Sets direction and carries organisational accountability.
What employers now want, and why proof beats a qualification.
Working effectively with AI is now a distinct, certifiable capability.
The 2026 revision weights testing and oversight far more heavily.
The Human-AI Collaboration standard recognises people who get more out of automated systems than the systems get out of them — directing, verifying and knowing exactly when to take the decision back.
The Data & Analytics standard certifies people who can take a question from a stakeholder, get to a defensible answer, and communicate it well enough to change what happens next.
The Software Engineering standard certifies people who write code others can safely change years later.