Helio Health: A Funnel That Kills Ideas and a Near-Miss

What the standard requires. Clause 8 requires the organization to establish, implement, maintain and continually improve processes for innovation, covering opportunity identification, concept creation, concept validation, solution development and solution deployment, together with portfolio management, intellectual property management and the management of uncertainty. These processes must be defined with criteria and controls and must produce documented information where needed.

What the standard does not require. Stage-gates, committees, forms or a fixed number of phases. Helio's fear is misplaced: the standard constrains outcomes, not the shape of the process. A lightweight design can conform.

Design.

• Single entry, fast triage: published criteria (strategic fit, customer value, feasibility, risk class) and a triage decision within 10 working days. This alone addresses the 1,190-to-11 collapse caused by an unscheduled committee.
• Two tracks: 'experiment' (small budget, four-week timebox, kill-or-scale decision) and 'development' (full validation and regulatory assessment). Most ideas belong on the first track.
• Mandatory validation gate for anything touching clinical or patient-facing function, with a named accountable reviewer. This is not bureaucracy; the near-miss is the argument for it.
• Decision rights and service levels for each decision point, plus a defined closure notice stating the reason and what would change the answer.
• Portfolio view: quarterly review of balance across horizons and risk, with explicit kill criteria.

Audit test. Take one of the 11 pilots: can Helio show the criteria applied, the decision taken, who took it and the supporting evidence? Today, no. That is the gap, not the absence of a stage-gate.

Why culture is a conformity issue. ISO 56001 addresses innovation culture within leadership and support requirements, and the innovation management principles include culture and collaboration. An auditor cannot certify 'culture', but tests it through interviews, decision records, recognition practice, and whether people at the edge of the organization can get a decision.

Symptoms mapped to requirements.

• One-line rejections with no criteria: absence of defined opportunity-evaluation criteria and feedback (Clause 8).
• A committee that meets 'when it can': no defined timing or decision rights (Clause 8).
• 'Isn't worth the risk': leadership and awareness gap (Clauses 5 and 7.3), and a warning that risk-avoidance is being rewarded.
• A near-miss shipped without clinical review: missing validation gate (Clause 8) and ineffective risk management (Clause 6).

Actions.

1. Publish decision criteria and a service level; report adherence monthly.
2. Mandatory written closure feedback on every submitted idea, including the reason and what would change the answer. The cost is small; the signal is large.
3. Recognise well-run experiments that failed, not only launches: make the learning record the celebrated artefact.
4. Train first-line managers on how they respond to ideas; the survey suggests the problem sits closer to the team than to the committee.
5. Add three culture questions to the engagement survey and report the trend into management review.

Evidence for the auditor. Decision logs showing turnaround times, closure notices, recognition records, survey trends, and minutes where leadership acted on a negative signal. Logs showing the loop closing matter more than a values poster.

Principle (Clause 7.5). Documented information is required to the extent necessary for the effectiveness of the system and must be controlled: available, protected, versioned, retained, disposed. More is not better; sufficiency and control are what is audited.

Intellectual property.

• An IP register: asset, owner, protection route chosen (patent, trade secret, publication, open), decision date and rationale.
• Evidence the decision precedes public disclosure. The conference paper is the control failure to fix: a pre-publication check owned by a named role, embedded in the deployment stage.
• Third-party IP and freedom-to-operate considerations recorded for product-bound innovations.
• Employee and contractor IP assignment evidence, plus awareness material for staff.

Concept validation.

• Criteria agreed before the test, the results, and the go/no-go decision with rationale. One page per experiment is enough.
• The clinical and regulatory assessment record for patient-facing changes.

Avoiding over-documentation.

• Record decisions and their basis, not meetings.
• Do not build a second documentation system: use the issue tracker for experiments, the PLM for development, and a single index linking them.
• Set and state retention periods; auditors ask how you know what to keep, not where form 12 lives.
• Apply a one-page rule: if a record would not change a future decision, do not require it.

Do not forget Clause 10. The near-miss should already have produced a corrective action; the absence of a record is itself a finding.