TCAA live clinic pathway

AI workflow clinic for humanitarian teams.

A facilitated seven-part clinic where a team maps one real workflow, designs a controlled AI-supported step, tests failure modes, runs a reversible pilot, and makes an evidence-based adoption decision.

Preview session Intake map Back to training

Learning time5.5 guided hours

Learning checks7 microchecks + 12 practice

Final draft16 professional scenarios

OutputWorkflow Clinic Pack

Humanitarian team mapping and testing an operational workflow — One workflow, visible decisions, controlled assistance, tested failure modes, and a human-owned scale or stop decision.

Clinic modules

Move from AI curiosity to a safer operating workflow.

Orientation

Frame the clinic decision

Define the operational problem, accountable owner, authorized boundary, intended benefit, exclusions, and stop conditions.

Part 1

Map the real workflow

Trace purpose, users, inputs, systems, handovers, delays, workarounds, risks, decisions, outputs, and baseline evidence.

Part 2

Write the Prompt Operating Procedure

Control approved inputs, evidence rules, prohibited behavior, output, review, records, ownership, and escalation.

Part 3

Assign human roles and authority

Define operators, reviewers, approvers, data owners, support, incident routes, separation of duties, and stop authority.

Part 4

Run a reversible pilot

Limit users, data, duration, and outputs; preserve rollback; test realistic and adversarial cases; monitor controls.

Part 5

Measure quality, risk, burden, and adoption

Compare baseline and pilot evidence across usefulness, source accuracy, rework, incidents, equity, and staff experience.

Part 6

Defend the Workflow Clinic Pack

Present evidence and decide to adopt, revise, restrict, pause, or stop, with residual risk and review triggers visible.

Practical outcomes

Teams should leave with something they can use immediately.

Current-state workflow map.
AI use-case and risk selection matrix.
Prompt and review template pack.
Role, approval, and upload-boundary rules.
Testing checklist for common AI failure modes.
First-month team implementation checklist.

Clinic evidence preview

See what a team should prepare before a serious AI workflow clinic.

Assessment blueprint

A public outline of the workflow evidence, use-case controls, facilitation checks, and launch boundary. It does not include the private facilitator key or scoring decisions.

Download blueprint

Assignment brief

A practical workflow implementation task using anonymized or redacted examples only. Real sensitive records must never be submitted through public previews or unsafe tools.

Open assignment brief

Public clinic materials

Eleven resources make the clinic operational.

Session previewSix-step walkthrough from workflow selection to implementation plan. Workflow intake mapCapture workflow owner, inputs, outputs, handoffs, risks, and AI-support candidates. Implementation checklistTurn the clinic result into owners, review gates, controls, and first-month actions. Facilitator session planWorkshop structure for team clinics, organization cohorts, and adapted in-person sessions. Assessment blueprintPublic assessment shape without private facilitator key or scoring decisions. Assignment briefBuild and defend the complete Workflow Clinic Pack with safe examples. Current-state mapCapture actual steps, handovers, workarounds, baseline, risk, and decisions. Prompt operating procedureControl purpose, approved input, instructions, evidence, review, version, and escalation. Responsibility matrixKeep authorization, review, approval, incident response, and stop authority with named people. Workflow test matrixTest normal, boundary, multilingual, missing, conflicting, and adversarial cases. Pilot evidence registerRecord quality, time, rework, risk, incidents, burden, and decisions case by case. Decision recordAuthorize adopt, revise, restrict, pause, or stop with evidence and review triggers.

Delivery boundary

A clinic can improve a workflow, but it cannot certify an organization overnight.

This pathway should be marketed as practical implementation support, not as a formal audit, legal compliance certification, or guarantee that AI use is safe in every context. Paid clinic delivery should only begin after session intake, payment or invoice handling, participant onboarding, materials, data-protection boundaries, follow-up expectations, and admin review are fully tested.

Delivery options

Useful when teams need a real workflow fixed, not another generic AI talk.

Single-team clinicFor one organization or department working on a selected workflow. Organization cohortFor local organizations that want shared AI workflow practice and implementation support. In-person workshopFor adapted field, headquarters, consortium, or training-of-trainers sessions.

Related learning

Use the clinic after teams know the basic safety boundaries.

Responsible AI for NGOs

Use this pathway first when a team still needs AI policy, acceptable-use rules, and risk controls.

Open Responsible AI

AI, Data Protection, and Digital Security

Use this before the clinic when workflows involve sensitive documents, partner records, field data, or risky file sharing.

Open data protection pathway