AI helps author learning plans and resources — metered, capped per tenant and logged, with learner personal data stripped from prompts and every output validated before use.
Provider performance and risks at a glance.
Active Learners
1,248
At Risk
34
Reviews Overdue
12
Achievement Rate
68.2%
Alex Carter
Software Developer L4
Priya Shah
Business Administrator L3
Sam Jenkins
Data Analyst L4
Illustrative dashboard with sample data.
Definition
AI apprenticeship management software uses artificial intelligence to assist providers — drafting curriculum, learning plans and resources, summarising reviews and surfacing risk — within governed guardrails. In Journey, AI assists people and never makes funding, gateway or compliance decisions; prompts exclude personal data and outputs are validated before use.
Journey's AI is rate- and budget-limited per tenant, learner personal data is removed from prompts, and every model output is validated before use. If a model is unavailable, the platform degrades gracefully rather than failing. Journey is independent software and is not DfE or Ofsted approved. It does not guarantee funding or inspection outcomes.
AI is metered and budget-governed with per-tenant call, daily and monthly token limits, plus role-based exceptions for senior staff.
Learner email, date of birth, National Insurance number and full address are never sent to a model.
Every AI output is schema-validated; a failure degrades to a deterministic fallback rather than an error.
AI usage is recorded per tenant and purpose, with only compact metadata stored — never the raw learner data.
AI in Journey is there to help staff author learning plans and resources, not to make funding or compliance decisions. Crucially, learner personal data never reaches a model: email, date of birth, National Insurance number and full address are stripped from prompts, and tenant-scoped helpers assert that any record referenced in a prompt belongs to your active tenant before anything is sent.
Every output the model returns is schema-validated before it is used. If validation fails, or the model is unavailable, the platform degrades to a deterministic fallback rather than producing an error or an unchecked result — so an AI outage never breaks your workflow.
AI is rate- and budget-limited per tenant, with call, daily-token and monthly-token limits, plus role-based exceptions for senior staff who need higher throughput. Usage is recorded per tenant and purpose so spend stays visible and governed. Only compact metadata is logged — identifiers and counts — never the raw learner data, so the usage ledger itself is not a back door to personal information.
Apprenticeship providers hold special-category personal data and funding-sensitive records. Bolting a general AI assistant onto that data without controls is a real risk. Journey's approach — PII-free prompts, validated outputs, per-tenant budgets and a metadata-only log — is designed so that AI accelerates authoring while the regulated data stays inside the boundaries it belongs in.
Funding-safe and audit-ready by design
Every record is scoped to your organisation on each read and write — one provider can never see another's data.
Every state change is recorded against the real person who made it, so the history stands up at audit.
Funding is computed server-side against the right year for each start date — 2024/25 and 2025/26, and built to adopt the 2026/27 pack now published by the DWP.
AI is metered per tenant, prompts are scrubbed of learner PII and outputs are validated, falling back gracefully.
Built and operated by a UK company to externally-audited information-security standards.
No lock-in: export your data whenever you like, and your employer relationships always remain yours.
Journey is independent software and is not DfE or Ofsted approved. It does not guarantee funding or inspection outcomes.
Book a walkthrough, or start free under 50 active learners.