HCD.7 - Arrange Providers, Access, Tools, and AI Support for Human Capability Development

Type: Human Capability Development practitioner support-arrangement pattern Status: Stable

HCD.7:0 - Use This When

Use this pattern when a capability-development programme, Method, or practice set needs contributions from people, sources, tools, AI, an environment, or an organization, and the practitioner must establish what will actually be available. Begin with the concrete functions and their timing: explanation, task preparation, safe access, feedback, checking, independent assessment, specialist input, and fallback. Do not begin with a provider label or a total-hours budget.

The first useful result is an enabled development arrangement: each required contribution has an identified supplier, competent and authorized scope, access conditions, capacity at the required time, data and privacy boundary, human checks, fallback, and evidence of availability. If one of these values is missing, return the supported design and the exact unavailable contribution, timing, access condition, or provider result. A plausible design is not an enabled arrangement.

The practical gain is feasibility before participation. The practitioner can see whether feedback arrives before the next attempt, whether peak assessment demand fits the provider window, whether a peer contribution prepares rather than replaces competent assessment, and whether AI support leaves enough human time to check consequential output.

Do not use HCD.7 to create a missing provider capability, infer that a provider is competent from a job title, prove learning from access, or decide that a person is suitable for a class-level arrangement. Use HCD.8 when an exact unavailable contribution needs an obtain-or-build comparison and the preparation or operating test that the chosen route requires. Personal fit remains with the applicable demand, target, profile, programme, and participation decisions.

HCD.7:0.1 - Working Terms

Name in this pattern	What it denotes
required development contribution	One result or action that the selected programme, Method, or practice needs from a learner, teacher, assessor, specialist, peer, source, tool, AI system, environment, or organization.
provider function	The bounded contribution to be supplied, such as explaining, preparing a case, giving actionable feedback, checking a critical error, assessing independently, maintaining access, or supplying a specialist result. A role title does not define the function.
provider	The identified person, group, organization, service, tool, source, or other System relied upon for a stated contribution under stated conditions. Different contributions may have different providers.
access condition	A permission, credential, data boundary, physical or digital route, timing, environment, interface, or participation condition needed to obtain a contribution.
capacity window	The amount and timing of provider attention or system availability that can serve the required work, including concurrency and peak demand. A total can fit while a critical window fails.
human-check contribution	The human judgement, verification, explanation, authorization, or accountability that remains necessary when AI or another tool assists. It has its own competence and workload.
fallback	A separately feasible way to preserve the required result or return safely when the preferred provider, tool, data, or timing becomes unavailable.
enabled development arrangement	The contribution allocation plus established providers, access, capacity, constraints, checks, fallback, and evidence sufficient to begin the bounded development use. It does not establish that the use occurred or succeeded.

Intended help in a programme, help permitted by a practice task, help actually available, and help received in an attempt are different facts. HCD.7 establishes actual availability; HCD.9 and HCD.11 preserve what was received and what the resulting attempt can support.

HCD.7:1 - Problem Frame

A programme can describe excellent practice and still be impossible to run. The teacher may have enough annual hours but not enough time during final assessments. The relevant project data may be inaccessible. A specialist may be available only after the decision deadline. An AI tool may be permitted for public material but prohibited for the learner’s project. A peer group may generate useful questions while lacking the competence or independence required for assessment.

Support is not one undifferentiated resource. Explanation, coaching, task preparation, feedback, technical access, domain checking, independent assessment, and specialist judgement have different competence, timing, privacy, and authority conditions. Replacing one with another can change both the learning activity and the evidence it yields.

The practitioner therefore has to allocate exact contributions, test actual availability and peak capacity, and preserve human checking and fallback rather than treating provider names, aggregate hours, or tool access as proof of an operating arrangement.

HCD.7:2 - Problem

The common shortcut is provider-by-label: a teacher, mentor, community, apprenticeship, platform, or AI tutor is assumed to supply every support function. The label hides task competence, assessment independence, access, supervision, workload, and the conditions under which the contribution is legitimate.

A second shortcut divides a total by the number of learners. The resulting “hours per learner” can be useful for costing but says nothing about individual contact, concurrency, feedback latency, or a final-week peak. Shared meetings also do not supply additional individual checking merely because more people attend.

A third shortcut treats automated assistance as free capacity. Searching, drafting, or calculating can reduce one burden and create another: prompt preparation, source checking, error challenge, data protection, and fallback. The arrangement needs a contribution account that remains true under actual use.

HCD.7:3 - Forces

Aggregate capacity is easy to budget, while timing, concurrency, feedback deadlines, and peak load determine actual service.
A provider title is easy to name, while competence and authority are contribution-specific.
Peer participation can broaden feedback and reflection, while independent or high-consequence judgement may require another qualified assessor.
AI can accelerate search, generation, calculation, or critique, while human checking and data governance consume real capacity.
Supported performance can improve the immediate result, while it does not establish unaided capability or retained performance after support changes.
One platform or community can host many functions, while access, facilitation, leadership, task expertise, and coordination remain distinct contributions.
A standard arrangement supports class planning, while personal needs and participation conditions become known only for an identified learner.
Redundancy improves resilience, while a nominal fallback that lacks competence, access, or time is not feasible.
HCD can expose the required support, while employment, privacy, safety, legal, clinical, target-domain, and organizational owners retain their decisions.

HCD.7:4 - Solution

Translate the selected programme, Method, and practice tasks into required contributions; assign each contribution to an identified provider or resource; and test competence, access, timing, capacity, constraints, human checking, and fallback before calling the arrangement enabled.

HCD.7:4.1 - Bind the Development Use

Name:

the human holder or defined learner class, participation basis, programme or practice recipient, and receiving decision;
the later-Work target, selected HCD.5 Method, HCD.6 practice set, required work products, protected errors, and intended evidence uses;
the calendar, attempt and feedback timing, learner workload, provider windows, peak periods, and decision deadlines;
project, source, data, environment, tool, AI, privacy, confidentiality, accessibility, permission, safety, and authority conditions; and
every direct owner whose competence, result, or authorization the arrangement consumes.

For a class-level design, state the audience assumptions and capacity range. Do not infer that a particular person fits the arrangement. When that person becomes known, recheck their task, access, participation, support, and accommodation conditions rather than inheriting the class average.

HCD.7:4.2 - Allocate Contributions, Not Labels

Construct one contribution allocation:

Allocation field	Required content
contribution and recipient	The explanation, demonstration, case, access, feedback, check, assessment, specialist result, coordination, or fallback needed; who or what uses it; and which task or result it enables.
provider and scope	Identified supplier, contribution-specific competence, authorization, independence where needed, and explicit non-responsibilities.
timing and service condition	Required date or interval, concurrency, response or feedback deadline, duration, handoff, and dependency on other contributions.
access and constraints	Data, source, project, environment, interface, accessibility, privacy, confidentiality, safety, legal, employment, or organizational conditions.
capacity	Preparation, delivery, checking, correction, coordination, fallback, and peak-load effort; competing uses and unavailable periods.
evidence and fallback	Evidence that the provider and access are available for this use; failure signal; feasible fallback or exact missing result and retry condition.

One person or tool may fill several rows, but each row keeps its own competence, workload, and constraints. Several providers may jointly supply one result, but their interfaces and the person responsible for the final judgement remain explicit. A catalogue, contract, platform account, or calendar slot is evidence input; none establishes every row by itself.

HCD.7:4.3 - Test Competence, Access, Timing, and Capacity

For each contribution:

Check competence and scope. Obtain evidence suitable for the task and consequence. Familiarity with content, teaching ability, domain judgement, facilitation, and independent assessment are not interchangeable.
Check authority and independence. State who may provide feedback, access confidential data, change a project, judge a protected criterion, authorize a retry, or make the later decision. A helpful participant is not automatically the authorized assessor.
Check actual access. Verify source edition or data, credentials, project permission, environment, tool or model availability, interface, and accessibility during the required window.
Check total and peak capacity. Include preparation, learner contact, review, correction, coordination, disputes, fallback, and competing work. Test the busiest window and the promised feedback interval, not only the total.
Check handoffs and fallback. Make the next provider, due time, result form, and failure signal visible. A fallback is feasible only when its competence, access, capacity, and authority also fit.
Return exact gaps. If a contribution is unavailable, state the blocked practice or evidence use, current supported design, safe interim boundary, and retry condition. Use HCD.8 when obtaining or building that contribution must be compared; a build decision is not a prerequisite to that comparison.

An enabled arrangement can be small: one learner, one task, a source, a teacher feedback window, an assessor, and a fallback. It need not establish a complete institutional service when the bounded use needs less.

HCD.7:4.4 - Arrange AI and Tool Support without Attribution Drift

For each AI or tool use, state:

the exact permitted contribution, such as search, example generation, calculation, drafting, critique, translation, or accessibility support;
the data and confidentiality boundary, model or tool condition when it changes risk, and prohibited uses;
the human action that remains, including selection, source verification, assumption and calculation checks, error challenge, explanation, judgement, or authorization;
time for prompting, checking, correction, escalation, and recording support actually received;
a fallback for outage, unsafe output, incompatible data, or unacceptable uncertainty; and
which practice or evidence uses require a changed-support or unassisted attempt.

AI access can improve equipped performance without establishing human learning. A no-AI observation answers only the bounded contribution and conditions it was designed to expose. HCD.10 selects a justified support change; HCD.11–HCD.13 interpret the resulting evidence.

HCD.7:4.5 - Keep Peer, Teacher, Assessor, and Specialist Work Distinct

Peers can expose alternative reasoning, rehearse feedback, notice a question, and help a learner prepare a revision. Those contributions can reduce avoidable teacher burden. They do not replace a teacher or assessor contribution whose result needs task competence, calibration, independence, authority, or a protected-error judgement.

A teacher may explain and coach while another assessor judges the independent result. A specialist may supply a bounded domain fact without teaching the whole task. An organization may enable project access without becoming the capability holder. State each contribution and handoff rather than turning all participants into one generic support role.

HCD.7:4.6 - Return the Arrangement or Exact Missing Contribution

An HCD.7 result contains:

Result position	Required content
arrangement boundary	Human holder or learner class, programme or practice, later-Work target, participation, calendar, evidence uses, protected conditions, and receiving decision.
contribution allocation	Every required learner, reference, tool, AI, peer, teacher, assessor, specialist, environment, organization, and fallback contribution.
provider basis	Supplier, competence, scope, authority, independence, access, data boundary, timing, capacity, peak load, and evidence of availability for each contribution.
human checking and attribution	Consequential AI or tool outputs, required human checks, checking capacity, and the distinction among joint, supported, and independent performance.
coordination and fallback	Handoffs, deadlines, failure signals, dispute path, and feasible fallback conditions.
disposition	`enabled development arrangement`, `supported design with exact missing contribution`, or `class arrangement awaiting personal fit`.
continuation	Practice start condition, personal-fit input, HCD.8 obtain-or-build question, provider return, schedule or capacity change, or smallest condition that reopens the arrangement.

Call the arrangement enabled only when every contribution needed for the bounded start and evidence use has a supported provider, access route, capacity window, and fallback or safe stop. A missing later optional contribution need not block an earlier independent result; state its reach rather than declaring the whole design impossible.

HCD.7:4.7 - Recognition and Assurance

Recognition. Use HCD.7 when a plan says “a teacher will support,” “peers will review,” “the platform provides feedback,” or “AI will help,” but the concrete contribution, recipient, timing, capacity, human check, or fallback is still unclear.

Assurance. Stronger reliance needs contribution-specific competence and authority, actual access during the required window, total and peak capacity, privacy and safety conditions, feasible handoffs and fallback, and evidence appropriate to the protected result. Service operation, provider competence, learning, transfer, and retention remain separate conclusions.

HCD.7:4.8 - What Changes in Practice

A practitioner stops asking, “Do we have a teacher and an AI tool?” and asks, “Which exact contribution must reach which learner or task, who can supply it under these data and authority conditions, does capacity fit the critical window, what human checking remains, and what happens when it fails?” Missing support becomes a named result, not a surprise during practice.

HCD.7:5 - Archetypal Grounding

HCD.7:5.1 - Twelve Participants and the 240-Hour Teacher Account

The FoundationsProgramme class design has twelve participants, thirty-two ordinary weeks, and two final weeks. An ordinary week allocates 1.5 teacher hours to group review, 3 hours to checking twelve short results, 1 hour to case and source preparation, and 1 hour to individual difficulties and assessment calibration: 6.5 hours. Each final week requires 16 hours, including individual reading, conversations, subgroup review, calibration, repeat checks, and a small delay or disagreement reserve.

The total is 32 × 6.5 + 2 × 16 = 240 teacher hours. Dividing by twelve gives a costing allocation of 20 hours per participant; it does not promise twenty individual-contact hours. The arrangement is enabled only if the two sixteen-hour peaks are reserved and ordinary feedback can arrive before the next two-week block.

With twenty-four participants, the original checking and conversation rows no longer fit. A single larger group meeting supplies no additional individual review. The supported return is to double the relevant checking capacity, divide the group with competent providers, reduce or redesign the promised contribution, or delay launch. HCD.7 does not call the larger class enabled merely because 240 total hours still appear in a budget.

HCD.7:5.2 - Supply Both Sides of Peer Learning and the Required Assessment

Participants learn to judge whether a diagnostic recommendation follows from its evidence and to revise their own work. A peer compares two permitted observation-to-conclusion fragments, identifies the relevant condition and writes a grounded comment. The author explains why to use or reject the comment, then makes the warranted revision. HCD.24 develops this bilateral exchange. The reviewer’s judgement and the author’s response are learning contributions even when no teacher time is saved.

Allocate accessible cases, criterion preparation, the peer’s reading and judgement, the author’s response and correction, and a qualified dispute window before the next block. If peers agree on an unsupported cause, a prepared teacher resolves the narrow subject question and returns its ground to both sides; each then performs the affected action again. Record the supplied help. A learner’s comment does not become an independent protected-error judgement merely because it is useful.

The teacher or designated assessor still applies the target-domain criteria to the consequential assessment result. Peer preparation can reduce search and coordination burden, while the amount saved remains to be observed. Keeping the learning purpose and the required assessor contribution distinct prevents one from silently replacing the other.

For a final attempt, the arrangement separately allocates reading, an individual conversation, possible correction, and a calibrated repeat check. If the named teacher is unavailable, a prepared colleague with compatible competence and access can be the fallback; otherwise the final week moves. A calendar placeholder or unprepared substitute is not a fallback.

HCD.7:5.3 - AI Search and Critique with Human Checking

AI may search public sources, propose examples, calculate, draft, or critique within the permitted data boundary. The learner first states the question and acceptable-result condition, then checks decisive sources, assumptions, configuration, and calculations and explains why the selected Method applies. A human assessor retains the final protected-error judgement.

The capacity account includes preparation of permitted prompts, checking consequential output, correcting failures, and a path for closed project data that cannot be sent to the tool. A narrow changed-support task can make one human contribution visible, but ordinary AI-supported work remains legitimate when the later-Work configuration includes it. Neither access nor fluent output establishes learning.

HCD.7:5.4 - Project Access Missing before Practice

A learner and teacher are available, but the planned workplace task requires project data and permission for a small safe trial that the organization has not granted. The arrangement returns the exact data, authorization, owner, affected task, and retry condition. A stipulated teaching case can preserve an earlier practice use, but it cannot be reported as the workplace result. HCD.8 can compare obtaining an alternative contribution with building an environment or provider capability when that exact gap must be closed for the selected practice.

HCD.7:5.5 - Prepare Mentor Help before Counting It as Capacity

Two prospective mentors will help groups compare rival diagnostic explanations. Give each a qualified case, the intended personal action, representative learner mistakes and the limits of the mentor’s authority. Each independently selects a response to the same two learner attempts, explains its grounds and identifies the question that would require a specialist. A prepared teacher compares their responses against the case criteria, corrects incompatible advice and observes another attempt where needed. HCD.8 supplies the obtain-or-build and operating-test work if that competence is still missing.

For the pump case, an AI suggestion asks the learner to take a fresh pressure reading. The mentor must notice that this variant already supplies the reading and allows no further worthwhile observation. They reject that suggestion, use the qualified reference to ask about the pressure implication, and leave the model-selection operation with the learner. If the subject implication itself is unresolved, the mentor returns that exact question rather than improvising a criterion.

This is an application of §4.4’s existing human-check arrangement. Access to the suggestion tool adds no qualified mentor until the person can select, adapt or reject its output in the actual task. Reserve preparation, checking, live help and fallback separately. For example, two thirty-minute mentor preparation attempts and forty minutes of teacher comparison allocate 100 provider-minutes before live delivery. Include those minutes once in the product’s fixed preparation; count live windows and subsequent maintenance where they actually fall.

Tutor CoPilot, November 2025 revision supplies a bounded example of tutor-facing suggestions built from expert reasoning with human selection and editing. Its paid professional tutors received training; the study does not establish that unprepared volunteers can replace them, nor that exit-ticket effects show distant transfer or independent tutor learning. Its tool price is not a complete support cost. The diagnostic preparation above is an authored application, not that study’s observed result.

HCD.7:6 - Bias-Annotation

Lenses: Gov, Arch, Onto/Epist, Prag, Did. Scope: arranging actually available provider, access, tool, AI, environment, and human-check contributions for bounded capability-development use.

Recurring bias	Likely drift	Repair
provider-label bias	“Teacher,” “mentor,” “community,” or “platform” is assumed to supply all functions.	Allocate contribution-specific competence, authority, timing, capacity, and constraints.
aggregate-hours bias	A total or per-person quotient is treated as service availability.	Test feedback deadlines, concurrency, individual work, and peak periods.
automation-free-capacity bias	AI assistance is counted without prompting, checking, correction, escalation, or fallback.	Put the human-check contribution and its workload in the allocation.
peer-substitution bias	Peer participation replaces independent or competent assessment.	State the peer’s learning and support contributions and retain the required assessor result.
access-by-account bias	A login, licence, or repository membership is treated as usable access.	Verify data, permissions, interface, timing, accessibility, and task compatibility.
class-average bias	A class arrangement is assumed suitable for every participant.	Recheck personal task, participation, access, support, and accommodation conditions.
nominal-fallback bias	An unprepared substitute or delayed source is called resilient capacity.	Test the fallback against the same competence, access, timing, and authority positions.
supported-capability bias	Equipped output or supervised participation becomes a human capability claim.	Preserve support conditions and route stronger evidence to HCD.11–HCD.13.

HCD.7:7 - Conformance Checklist

The human holder or learner class, participation basis, programme or practice, later-Work target, calendar, evidence uses, and receiving decision are explicit.
Every required explanation, preparation, access, feedback, checking, assessment, specialist, coordination, environment, and fallback contribution has its own row.
Provider competence, scope, authority, and independence are supported for the exact contribution rather than inferred from a role label.
Data, source, project, tool, AI, interface, privacy, confidentiality, accessibility, safety, legal, and organizational conditions are explicit where material.
Total capacity includes preparation, delivery, review, correction, coordination, disagreement, fallback, and competing work.
Peak periods, concurrency, feedback deadlines, and handoff times fit the actual provider windows.
Peer, teacher, assessor, specialist, organization, source, tool, and AI contributions remain distinct.
AI and tool rows state permitted work, prohibited data or use, human checks, checking capacity, attribution, and fallback.
Intended help, permitted help, actual availability, received help, joint performance, and independent performance are not conflated.
A fallback satisfies the same relevant competence, access, capacity, timing, and authority conditions or is reported as unavailable.
A class-level arrangement retains personal-fit and participation questions for the named learner.
The disposition says which bounded use can start and returns each exact missing contribution without claiming provider creation, service operation, learning, transfer, or retention.

HCD.7:8 - Common Anti-Patterns and How to Avoid Them

Anti-pattern	Why it fails	Repair
“We have 20 teacher hours per learner”	A costing allocation is mistaken for individual contact and peak capacity.	Reconstruct the actual shared, individual, preparation, checking, and final-week rows.
“One meeting supports twice the class”	Individual checking and conversations grow even when group time does not.	Add competent capacity, divide the group, redesign the contribution, or delay.
“Peers can grade it”	Peer review may lack domain competence, calibration, independence, or authority.	Use peers to prepare feedback and preserve the qualified assessor contribution.
“The platform gives feedback”	The provider, information content, latency, and correction opportunity are hidden.	Name the exact feedback contribution and its evidence, timing, and recipient.
“AI saves the teacher’s time”	Prompting, checking, correction, data limits, and escalations are omitted.	Compare the complete human-plus-tool workload and preserve consequential judgement.
“A licence means access”	Credentials can coexist with missing data, permission, interface, or usable timing.	Test access for the actual task and window.
“Supervised participation proves capability”	Participation, entrusted access, and supported output answer different questions.	Record the support configuration and obtain the applicable performance evidence.
“We should train every provider”	Existing qualified contributions are ignored and HCD.8 becomes ritual.	Use the available provider; build only the exact missing capability.

HCD.7:9 - Consequences

The practitioner gains a launch boundary grounded in real contributions rather than names or totals. Capacity bottlenecks, data restrictions, assessment independence, AI checking, and fallback become visible early enough to redesign the arrangement or return a precise missing result.

The cost is service-level detail. Provider competence, access, schedules, peak loads, handoffs, and privacy conditions must be checked, and some attractive programme designs will remain disabled. A class design may still need personal adaptation, and an enabled arrangement still supplies no evidence that learning occurred.

HCD.7:10 - Rationale

Development support is a configuration of contributions. Its feasibility depends on the relation among task, provider, recipient, timing, access, capacity, and consequence. Provider names and aggregate resources erase those relations and make substitution look easier than it is.

Separating intended, permitted, available, and received help protects both design and evidence. Explicit human checking prevents automation from hiding labour or reallocating consequential judgement without notice. A precise missing-contribution return gives HCD.8 or another owner a real obtain-or-build question instead of a generic demand for more support.

HCD.7:11 - SoTA-Echoing

The working question is: How should a practitioner establish that the people, access, tools, AI, and other support needed by a development use are actually available at the required competence, time, capacity, and consequence boundary?

The selected best-known line treats support as a configuration of distinct provider, access, coordination, and checking contributions rather than as a provider label. Brooks and colleagues’ 2024 qualitative study of learning-health-system communities (SOC-05) makes people, platform, facilitation, leadership, and resource contributions separately visible; it supplies arrangement questions, not a causal capability effect. Matsuo and Tsukube’s 2020 cognitive-apprenticeship review (SOC-02) supports distinguishing modelling, coaching, scaffolding, articulation, reflection, and exploration while leaving their sequence and workplace transfer conditional. HCD.7 adapts those contributions into the allocation and service tests in §§4.2–4.5.

The serious ordinary alternative is label-and-total planning: name a teacher, mentor, community, apprenticeship, or platform, total the available hours, and assume the functions will be supplied. It is cheaper to plan, but it hides competence, individual work, peak load, access, privacy, assessment independence, and fallback. HCD.7 rejects the provider label and aggregate total as sufficiency evidence while retaining them as inputs to a complete contribution account.

Tullis, Goldstone, and Hanson’s 2015 experiment (COG-11) is direct limit evidence that assistance improving practice can impair later unassisted performance and that support scheduling matters in its laboratory motor task; it does not prescribe a universal fading schedule or occupational arrangement. Wiese, Kilty, and Bennett’s 2018 clinical realist synthesis (SOC-03) shows that supervised participation and graduated access depend on task risk, trust, learner position, and service pressure; its clinical governance does not transfer wholesale. Dell’Acqua and colleagues’ bounded field experiment (AI-03) supports independent error challenge under specific task and model conditions, not a stable universal AI frontier. These limits produce the support-state, human-check, and changed-support boundaries in §§4.3–4.7.

The selected line costs more coordination than naming providers and adding hours, but it preserves the exact result, peak feasibility, authority, privacy, human judgement, and fallback that the cheaper plan loses. Reopen it when a provider, task, capacity window, access rule, AI condition, evidence use, or serious source changes, or when a lower-effort arrangement test preserves the same decision-bearing values.

HCD.7:12 - Relations

HCD.2: supplies intended contribution and access allocations for programme candidates; HCD.7 establishes what is actually available.
HCD.5: supplies Method-specific provider, support, task, risk, and evidence needs.
HCD.6: supplies task materials, permitted help, criteria, feedback opportunities, and domain-provider requirements.
HCD.8: receives an exact unavailable provider, environment, assessment, or support contribution for obtain-or-build comparison and the preparation or operating test actually needed. Already established usable contributions can be arranged directly through HCD.7.
HCD.9: uses the enabled arrangement and records help actually received, feedback delivered, corrections, and provider failures during practice.
HCD.10: may change support across practice episodes for a stated development or evidence purpose; the revised arrangement must remain feasible.
HCD.11–HCD.13: interpret supported, independent, transferred, delayed, and changed-support performance; availability alone establishes none of them.
HCD.14: may revise only the affected support assumption or programme component from compatible evidence.
Learning-product practices: can specify concrete materials, interactions, and schedules; publication or access does not establish provider competence or service operation.
Organization, operations, platform, privacy, safety, legal, clinical, employment, accessibility, provider, and target-domain practices: retain their own capacity, authority, competence, data, risk, and correctness decisions.