HDI Certification for Partner Auditors: Tutorial and Test

Why does HDI certify auditors?

Every HDI audit should be as good as the next one, whichever partner runs it. Phase 2 audits are therefore run only by certified auditors of the partner. Certification protects the client, who gets the same audit quality from every partner, and it protects the partner's reputation.

HDI frees qualified people from routine work. It is not about replacing people with AI. An audit shows where capable people spend their time making up for weak systems, so that automation can give that time back. Handled well, a high score shows how much rests on the best people. It does not accuse them.

How do Phase 1 and Phase 2 work?

Phase 1 is a free self-assessment of about 15 minutes that the client completes online. It gives the HDI score, the three layer scores, an indicative process heatmap and the operational leak in hours. It gives no recommendations and no money figures, because a self-assessment can show what is happening and where, but not why. Phase 2 is the audit: it explains why the dependency exists and what to change first.

What are the three layers?

HDI splits dependency on people into three layers, because it works through three different mechanisms. Each layer has its own block of behavioural questions and its own score.

• Coordination: how work travels. Approvals on top of the required ones, hand-offs that wait for a reminder, status meetings, and what happens when the person who pushes the work is away.
• Control: how problems surface. Whether a system or a person detects them, how incidents are escalated, how long closure takes, and how much re-checking happens beyond the required reviews.
• Cognitive: where knowledge lives. Single-expert steps, undocumented procedures, how long a new hire takes to work alone, and how much reporting people assemble by hand.

Each layer calls for a different fix: workflow automation for coordination, monitoring and alerts for control, and knowledge capture and cross-training for cognitive. Required GMP steps, such as QA review of an executed batch record or QP certification, do not raise the score. The questions ask about the work on top of those steps, and about whether the steps wait for one specific person.

What do the score and the bands mean?

The overall HDI score runs from 0 to 100, and a higher score means more dependency on people. The score does not grade the team: strong teams often score high because capable individuals hold the operation together. Each score falls into one of five bands:

• Systemised (0–14). Systems run the processes; people supervise and handle exceptions.
• Managed (15–34). Processes are documented and have working backups. Dependency exists, and the organisation keeps it under control.
• Moderate (35–54). Key people and manual hand-offs carry a noticeable share of the work. Growth will increase it.
• High Risk (55–74). Specific individuals sit in critical paths. Absences cause delays, and growth needs proportional hiring.
• Critical (75–100). One unavailable person can stall a regulated process. This needs action now.

Each screening result also carries a confidence percentage. It reflects how complete and consistent the client's answers are, and the Phase 2 workshop resolves low-confidence results.

What are the levels of dependency in the instrument?

The Phase 2 instrument asks behavioural questions in three blocks, one for each layer. Each question describes a concrete situation, such as a key person being away or a deviation being opened, and you listen to what the interviewee says happens. Ask a manager how automated the operation is and you get optimism. Ask what happens in a specific situation and you get behaviour.

Each answer is matched to a behavioural anchor: a fixed description of what a given level of dependency looks like in practice. The anchors run from a process that runs on systems to a process that stops without one specific person. They add up to the three layer scores and the overall 0–100 score. You score the behaviour that is described, not the interviewee's opinion of their own process.

Take a key person who is away. The answers run from a documented backup that takes over, through a colleague who partly covers and work that slows to urgent items only, to work that stops. The further along that list, the higher the dependency.

The instrument itself is proprietary. You see its questions and anchors in the Phase 2 tool after you enter your partner key, not on this site.

How does a workshop run?

The core of Phase 2 is a workshop of about half a day with the client's key operational people. No financial data is needed, and the client is not asked to send controlled documents.

1. Before. The client takes the Phase 1 screening through your partner link.
2. Set up the session. In the Phase 2 tool, enter the client's work email, the client company, the interview date, the interviewee's name, role and department, and the industry. The email links the interview to the client's screening, and the tool loads it. You see only the screenings of clients who came through your own link.
3. Interview. Ask each question and listen for behaviour. Match what you hear to an anchor and score it. Record warning signs as flags, and short quotes that explain the score.
4. Check. If a document is needed to check an answer, look at it with the client in the session.
5. Summary. The tool shows the layer scores and the total, the comparison with the client's screening, your flags and quotes, and a plan for measuring HDI again after the audit.
6. Results session and report. You take leadership through the findings in a live session. The written report follows.

What is the perception gap?

The perception gap is the measured difference between how leadership believes the operation runs and how it runs day to day. Phase 1 is usually completed from management's view of the process. In the workshop, the people who do the work describe the same behaviours, and you score them against the same anchors.

Because both phases use the same 0–100 anchors, the gap is a direct comparison per layer. A screening score of 45 against a workshop score of 70 is a 25-point gap in that layer. The gap can go either way: leadership can underestimate coordination and cognitive load, and it can overestimate control problems that the frontline has already solved.

When an interviewee describes something different from the screening, record the difference as a finding and flag it. In the results session, take leadership through the gap layer by layer. The discussion moves from whether to automate to where the picture of the operation differs from how it runs, and what that puts at risk.

What does the audit report contain?

The written report follows the results session. It contains the findings with their causes, losses quantified on assumptions agreed with the client, the processes to automate first, and an implementation plan. The sample report shows the structure on a fictional company: summary, verified score and perception gap, findings, what the dependency costs, priorities, and a plan with the figures to track.

Money enters the report only on assumptions the client agrees. The monthly cost is the hours per month multiplied by a loaded hourly rate, and the rate is the client's own figure, not yours. The free screening reports hours and FTE only.

The report describes processes and roles. It keeps the evidence and quotes behind each score, but it does not attribute statements to named people, and it contains no patient data. A real report is delivered to the client by the partner who ran the audit.

What are the rules for client data?

Treat the client's data as the Data & security page describes it:

• The screening collects the client's work email, company name if given, industry, team size band and answers. It does not ask for patient data, clinical records, batch numbers, product names, financial data or the names of employees. Tell clients not to enter them.
• In the workshop, names are recorded only if the client wants follow-up questions to go to a specific person.
• If a document is needed to check an answer, look at it with the client in the session. Do not ask the client to send controlled documents.
• HDI is an advisory assessment. It does not create, store or approve GxP records, and nobody uses it for batch, release or quality decisions, so it needs no computer system validation.
• Changes the client makes after the audit, such as a new workflow or a new system, go through the client's own change control and validation as usual.

Who does what in the partner model?

HDI supplies the method, the tools and the certification. The partner runs the screening with its clients, sells and runs the Phase 2 audit, and keeps the client and all the implementation after the audit. HDI does not sell to pharma directly, does not run audits and does not build automation, so it never competes with a partner.

Phase 2 audits are run only by certified auditors of the partner. A partner needs at least one certified auditor in its team to run a Phase 2 audit.

How does the test work?

The test has 20 questions with four answers each, and exactly one answer is right. The pass mark is 80%, which is 16 correct answers. Questions and answers appear in a random order.

• To take the test you need your partner key, your name and your work email.
• You can take the test at most 3 times in 24 hours with the same work email.
• If you do not pass, you see the topics to review, not the right answers.
• If you pass, you get a certificate that is valid for 12 months. It carries a link that anyone can use to check it.
• Phase 2 audits are run only by certified auditors of the partner.
• The test covers this page: Phase 1 and Phase 2, the three layers, the score and the bands, the levels of dependency, the workshop, the perception gap, the audit report, client data, and the partner's role.

Take the certification test →