Competency framework and certification standard

The standard we assess against.

Seven competency domains, four assessment components, published weights and published thresholds. This is the document an assessor scores against, and the one a candidate is entitled to read in full before sitting anything.

Version
2.0
Issuing body
Bulletproof Automations
Owner
Bulletproof Automations
Issued
Platform basis
n8n 2.31.6
Next review
Status
Published
Copyright
© 2026 Bulletproof Automations. All rights reserved.

Framework section 1

Purpose and scope

This framework defines the credentials issued under the Bulletproof Automation Builder Pathway: what each attests, what evidence is required, how it is assessed, and where the pass line sits.

It exists so that candidates know what is being measured, assessors apply one standard, and employers — including Bulletproof Automations itself — can interpret what a holder can actually do.

The subject of this pathway is automation engineering

n8n is the platform on which competence is currently demonstrated, not the subject being certified. This distinction is deliberate and protects the credential’s value against platform change.

Framework section 2

The pathway

The three stages of the Bulletproof Automation Builder Pathway
StageCredentialTypeWhat it establishes
1 Bulletproof Automation Foundations Certificate Certificate of completion Completed foundational training and demonstrated basic workflow-building proficiency
2 Bulletproof Certified Automation Builder (BCAB) Professional certification Can independently design, integrate, troubleshoot, test, and document moderately complex automation systems
3 Bulletproof Certified Automation Engineer (BCAE) Professional certification Can architect, deploy, secure, operate, monitor, and govern production automation systems

Scroll the table sideways to see every column.

Stage 1 is active. Stage 2 launches with the first Intermediate cohort. Stage 3 is defined but not issued.

Training and certification are separate products

Completing a cohort does not confer a certification. It qualifies the learner to attempt the assessment.

Framework section 3

Stage 1 — Bulletproof Automation Foundations Certificate

This is a certificate of completion, not a professional certification

It makes no independent competency claim beyond readiness to progress. This is stated on the certificate itself.

What it attests

The holder completed the Foundations training programme and demonstrated the ability to:

  • navigate n8n and describe workflow structure in plain language;
  • build a workflow from trigger to output;
  • inspect items, fields, and JSON, and read node input and output;
  • use expressions to move data between nodes;
  • connect an application using a safely stored credential;
  • make a basic API request and read the response;
  • apply IF and Switch logic with a fallback;
  • implement basic error handling;
  • add a controlled AI step;
  • test multiple paths, including one handled failure;
  • export, document, and explain a workflow.

Award standard

Three requirements, all mandatory:

  1. Participation — meets the published attendance minimum.
  2. Portfolio evidence — submits the required weekly artifacts (workflow export, screenshots, notes) for each class.
  3. Capstone — completes the capstone to the published specification and explains it satisfactorily in the walkthrough.

There is no separate examination and no direct-entry route. A certificate of completion cannot be awarded for training that was not completed.

Certificate wording

[Name] completed the Bulletproof Automation Foundations programme and demonstrated the foundational skills required to progress to intermediate automation training. This certificate records completion of training. It is not a professional certification.

The Foundations cohort

Framework section 4

Stage 2 — Bulletproof Certified Automation Builder

The holder can independently take a moderately complex automation requirement and design, build, integrate, harden, document, and maintain a solution — and can diagnose a malfunctioning system they did not build.

The seven competency domains

  1. 01

    Data Engineering

    Transform nested, multi-item, inconsistent, and incomplete workflow data

  2. 02

    Integrations & APIs

    Build integrations from API documentation rather than step-by-step instructions

  3. 03

    Workflow Architecture

    Decompose automation systems into maintainable workflows and reusable components

  4. 04

    Logic & State

    Design decision paths, fallbacks, duplicate prevention, and state-aware processing

  5. 05

    Reliability

    Design retries, error paths, validation, recovery, and observable failure behaviour

  6. 06

    Troubleshooting

    Diagnose workflows the candidate did not originally build

  7. 07

    AI Automation

    Use AI selectively inside controlled workflows, with validation and deterministic safeguards

Documentation and communication are assessed within every domain, not as a separate subject. A candidate who cannot explain a design has not demonstrated the domain.

Framework sections 4.3 to 4.7

Assessment components and coverage

The four assessment components, their weights and how each is delivered
PartComponentWeightDelivery
1Knowledge & Scenario Judgment20%Timed, written. Judgment scenarios, not definitions.
2Practical Build Challenge35%Business requirement, unseen. Candidate chooses the approach.
3Troubleshooting Challenge30%Live, timeboxed, supervised. Workflow the candidate did not build.
4Architecture Walkthrough15%Live, 5–10 minutes plus follow-up questions.

Scroll the table sideways to see every column.

Domain-to-component coverage

Every domain must be evidenced by at least two components. This prevents gaps, and prevents any single domain being judged on one artifact.

Which assessment component evidences each competency domain
Domain P1 P2 P3 P4
1. Data Engineering Assessed Assessed Assessed Not assessed
2. Integrations & APIs Assessed Assessed Assessed Assessed
3. Workflow Architecture Assessed Assessed Not assessed Assessed
4. Logic & State Assessed Assessed Assessed Assessed
5. Reliability Assessed Assessed Assessed Assessed
6. Troubleshooting Assessed Not assessed Assessed Assessed
7. AI Automation Assessed Assessed Assessed Assessed

Scroll the table sideways to see every column.

Part 2 — Practical Build Challenge

The candidate receives a business requirement they have not previously built. No node-level instructions are given. The assessment is whether they can translate a business requirement into an automation design.

Reference scenario shape, rotated each attempt: a service company receives work requests containing customer name, email, service type, priority, location, and description. The system must validate required information, normalise input, look up external information through an API, route the request, prevent duplicate processing, store the result, handle API failure, and produce a final notification.

Part 3 — Troubleshooting Challenge

The candidate receives a workflow they did not build, with the instruction: this workflow has been producing inconsistent results. Investigate it and produce a corrected version plus a short diagnostic report.

The number of defects is not disclosed. The task is to determine whether the system is behaving correctly — not to find a stated quantity of bugs.

A rotating subset of the defect bank is seeded per attempt:

  • a field renamed upstream but still referenced downstream
  • wrong IF operator (contains where equals is required, or the reverse)
  • missing fallback route on a Switch
  • duplicate insert with no idempotency guard
  • retry configured on a permanent 4xx error
  • a branch that never reconnects to the output path
  • malformed AI structured output passed downstream unvalidated
  • missing-data path allowed to continue rather than stop
  • API result trusted without validation
  • silent partial failure: fewer records leave than entered

All troubleshooting workflows run on pinned or sandboxed data. No candidate spends assessment time configuring credentials.

Part 4 — Architecture Walkthrough

Five to ten minutes, plus assessor follow-up. The candidate covers: the problem solved, the design, major data structures, where decisions occur, what can fail, what happens when it fails, how duplicate processing is prevented, what they would monitor, and why they chose this design over alternatives.

This component also serves as authorship verification

A candidate who cannot answer follow-up questions about their own submission has not demonstrated the competency, regardless of the artifact’s quality. This is stated to candidates in advance.

Framework section 4.8

The passing standard

Thresholds for the first assessment cycle are provisional and will be confirmed after calibration.

  • 75%overall
  • 70%minimum, Practical Build
  • 70%minimum, Troubleshooting
  • 50%no domain below

Framework section 4.9

Critical safety failures

Certain outcomes trigger Remediation Required regardless of overall score.

  • Exposing credentials, API keys, or tokens in any submitted artifact
  • Sending sensitive or unnecessary personal data to an AI provider
  • Deleting or modifying production-like data without appropriate safeguards
  • Designing a workflow that silently discards failed records
  • Leaving a critical action entirely dependent on unvalidated AI output

These are professional standards, not scoring criteria. They are published to candidates before assessment.

Framework sections 4.10 and 4.11

Outcomes and retakes

Certified

Met all requirements. Credential issued.

Assessment Not Yet Passed

Did not reach the standard. Written feedback issued. May reattempt.

Remediation Required

One or more specific competency gaps, or a critical safety failure. Named gaps must be addressed and evidenced before a further attempt.

The term “failed candidate” is not used. The assessment measures whether the standard has been reached yet.

Retake policy

  • First reattempt: minimum 7-day wait.
  • Second reattempt: minimum 30-day wait.
  • Third and subsequent: at assessor discretion, following a remediation plan.
  • Every attempt uses a different practical scenario and a different seeded troubleshooting workflow.

Framework section 4.12

Entry routes

Path A · Training route

Foundations Certificate → Intermediate Cohort → BCAB Assessment → Certified

The taught route. Holding the Foundations Certificate waives the readiness assessment.

The Intermediate cohort

Path B · Experience route

Working practitioner → Readiness Assessment → BCAB Assessment → Certified

Path B candidates are not required to purchase training. A readiness assessment establishes that an attempt is realistic; it is advisory, not a gate, and its fee is credited against the assessment fee.

Book a readiness review

Path B is essential to the credential’s independence and is to be maintained even when it is commercially inconvenient.

Framework section 5

Stage 3 — Bulletproof Certified Automation Engineer

Defined, not yet issued

Defined now for pathway clarity; not to be issued until its standard and assessment are written.

Reserved scope: architecture, databases, advanced integrations, security, secrets management, environments, deployment, version control, CI/CD, queue-based processing, scaling, concurrency, observability, recovery, operational ownership, AI governance, human approvals, permissions, production documentation, and system lifecycle management.

At this level the platform is the implementation detail, not the subject.

Framework section 6

Assessment integrity

Assessor record
Every component is scored per domain with a one-line written justification. The record is retained and forms the basis of any appeal.
Second review
Mandatory for any result within 5 percentage points of a threshold, any Remediation Required outcome, and any candidate taught by the assessing assessor.
Conflict of interest — teaching
Where the assessor also taught the candidate, second review applies to all outcomes, not only borderline ones.
Conflict of interest — hiring
See use as a hiring standard.
Calibration
Where more than one assessor is active, both independently score the same two submissions each cycle and reconcile before results are issued.
Authorship
Part 4 verifies that submitted work is the candidate’s own. Parts 1 and 2 are not assumed to be unaided; the live components carry the integrity weight, and this is stated to candidates openly rather than policed covertly.
Appeals
Within 14 days, reviewed by an assessor not involved in the original decision.
Revocation
For misrepresentation of the credential, or submitted work later shown not to be the candidate’s own.

Framework section 7

Use as a hiring standard

Bulletproof Automations uses BCAB as a minimum competency standard for automation roles. This is published, because a credential with a real consumer is worth more than one without.

  1. Separation of decisions. Certification outcomes are determined solely against this framework. Hiring interest, or its absence, is not a factor in any assessment decision and is not discussed with the assessor before an outcome is issued.
  2. No implied employment. Certification is not an offer, promise, or indication of employment. No marketing material for the pathway may suggest that certification leads to work at Bulletproof Automations.
  3. Documented basis. Because outcomes may affect employment prospects, the per-domain score record, second review, and appeals route apply without exception.
The builder pool and hiring standard

Framework section 8

Credential issuance and validity

Credential ID
Unique, non-sequential, issued per credential
Verification
Public page: holder name, credential, issue date, platform version, current status. Verify a credential.
Platform version
Recorded on the credential — for example, “assessed against n8n 2.31.6”
Validity — Foundations Certificate
Does not expire. It records completion of training on a date.
Validity — BCAB
24 months
Recertification
Submit one current workflow demonstrating continued practice, plus a short update assessment covering platform change since issue
Permitted description
Supplied with each credential. Holders may not describe it as an n8n certification.

Framework sections 9 and 10

Provisional standards and calibration

The thresholds above are set before any candidate has been assessed, and are therefore provisional. We say so rather than imply a rigour we have not yet earned.

Calibration protocol

  1. Assess cohort one against the provisional thresholds.
  2. Record the full score distribution per component and per domain.
  3. Review item-level performance in Part 1, score spread in Parts 2 and 3, and any component where nearly all candidates cluster at one end — a sign the instrument, not the candidates, needs work.
  4. Confirm or revise thresholds and publish them as Version 3.0 with the reasoning.
  5. Candidates assessed under provisional thresholds are not disadvantaged by any subsequent increase; results stand as issued.

Standard-setting after piloting is normal practice and is a credibility asset, not a weakness.

Outcome tracking

Recorded from cohort one, with holder consent:

  • numbers assessed, certified, not yet passed, and remediation required — per cycle, published annually;
  • portfolio links, where holders permit publication;
  • self-reported outcomes at 6 and 18 months (paid automation work, role change, projects shipped).

This record is the primary long-term asset of the programme. It is built from the outset rather than reconstructed later.

Framework section 11

Independence statement

Bulletproof Automations is an independent training and certification provider. This framework and the credentials issued under it are not affiliated with, endorsed by, or accredited by n8n GmbH or any other vendor. “n8n” is referenced descriptively as the platform on which competence is currently assessed. Vendor certifications are held separately and additionally.