Charities and not-for-profit · Playbook 56

Impact reporting from case notes

Outcome figures and themes drawn from caseworkers' notes for the trustees' report, funder reports and internal learning. De-identified, every figure traceable to records, nobody scored, and trustees approving what is published.

  • Generative AI
  • Predictive ML
  • Human in the loop
  • Charities SORP 2026
  • Charity Commission (CC3)
  • Serious incident reporting
  • UK GDPR
  • ICO anonymisation guidance

At a glance

The problem
The difference the charity makes is written down in thousands of case notes, but reports are built by hand from spreadsheets and memory
Typical volume
Around 30,000 people supported a year, 400 caseworkers, 40 funder reports and one trustees' annual report
Who's involved
Director of services, head of impact, service leads, caseworkers, data protection, finance, trustees
Pilot
8–10 weeks on one service's past year of notes, after a 3–4 week discovery
Platforms
AWS, Azure, or a sovereign or on-prem stack

An illustrative scenario. The numbers are typical examples, not a client case study. The cloud services and UK rules were checked on 9 November 2026.

The business problem

A support charity's case management system holds the evidence of its impact: what people came with, what was done, and what changed. Most of it is in free-text notes. Structured fields capture some outcomes, but they are filled in unevenly, and the detail that funders and trustees want sits in the narrative.

So reporting is manual. Each funder wants its own measures, periods and definitions. An impact team pulls spreadsheets, service leads read samples of notes, and figures are adjusted until they look right. The same number appears differently in two reports. Small services report counts so low that a reader could work out who is being described. Learning that should change services, such as a rise in a particular kind of need, is noticed months late, if at all.

The bar has also gone up. Under Charities SORP 2026, for periods starting on or after 1 January 2026, charities above £500,000 income must review performance against their objectives and explain the impact they are making, and the largest must also review fundraising performance. Impact claims need to be backed by evidence, and trustees put their names to them.

What AI changes

  • Outcome measures come from structured fields first. Where they are missing, a model reads the notes and tags outcomes against the charity's own outcome framework, which a person can check.
  • Every figure is calculated by a defined query on a dated snapshot of the records, not written by a language model, and can be traced back to the records behind it.
  • Notes are de-identified before analysis, and small numbers are suppressed or grouped before any figure leaves the impact team.
  • Emerging themes across services are surfaced for service leads to review, as internal learning, not as conclusions.
  • Draft narrative sections for the trustees' report and funder reports are written around the approved figures, quoting people only where there is consent.
  • Nobody is scored. The system never rates an individual service user, caseworker or team. Trustees approve the annual report, and the named grant holder approves each funder report.

How success is measured

Agree these, and measure the baseline, before anything is built. Example targets are typical, not promises.

Time to produce a funder report

From the request to a report ready for sign-off.

ExampleFrom 2 weeks to 2 days

Figures traceable to records

Share of published figures linked to a query, a snapshot date and a suppression rule.

ExampleEvery one

Tagging agreement

How often the model's outcome tags match a human-coded sample, reported by outcome and by service.

ExampleAbove an agreed threshold, reviewed quarterly

Suppression breaches

Published figures below the small-number threshold, or that could identify someone.

ExampleZero

Draft edit rate

How much the impact team changes the drafts. Very high means it isn't helping; near zero can mean drafts are approved unread.

ExampleStable, and reviewed monthly

Learning acted on

Themes reviewed by service leads that led to a recorded decision.

ExampleReported to trustees each quarter

UK regulation and controls

Cross-sector rules are on the UK regulation map.

  • Charity regulators (the SORP-making body)Charities SORP 2026 (periods beginning on or after 1 January 2026)

    Reporting is tiered by income: Tier 1 (up to £500,000) summarises main achievements; Tier 2 (to £15 million) must review performance against objectives and explain the charity's impact; Tier 3 adds a review of fundraising performance. Every figure in the impact section carries its source, period and method, and the report says what the data can't show.

  • Charity CommissionThe essential trustee (CC3, updated 16 September 2026) and the Commission's view on AI

    Trustees approve the annual report and can't delegate that responsibility. The Commission expects reasonable independent checks before relying on AI output for an important decision. Trustees see the method, the tagging accuracy and the limitations alongside the draft, not just the numbers.

  • Charity CommissionSerious incident reporting (guidance updated 16 January 2026)

    A significant data breach or loss is a reportable serious incident, reported as soon as reasonably possible, and trustees decide whether it is significant. That stays true when a supplier causes it, and the Commission has reminded trustees of this. A personal data breach may also need reporting to the ICO within 72 hours. The incident plan covers the analysis store and the AI suppliers.

  • ICOUK GDPR, the Data Protection Act 2018 and statistical purposes

    Case notes hold special category data such as health, and often concern people in vulnerable circumstances. A DPIA settles the lawful basis and the Article 9 condition before the pilot. Since 5 February 2026 the UK GDPR defines statistical purposes: the output must be aggregate data that is not personal data, and the data must not be used for measures or decisions about a particular person. That is why nothing flows back into individual records.

  • ICOAnonymisation and pseudonymisation guidance (March 2025, under review)

    Pseudonymised notes are still personal data; only properly anonymised outputs fall outside the law. Names, contact details and other identifiers are removed or replaced before analysis, the re-identification key stays with the case management team, and small numbers are suppressed. The ICO consulted on research-specific guidance from August to October 2026, so the rules are reviewed when it is final.

  • Fundraising RegulatorCode of Fundraising Practice (in effect 1 November 2025)

    When impact figures and stories are reused in appeals or bids, they must not mislead and need appropriate evidence. Case studies must reflect real situations, say if details were changed, and have the person's clear and informed permission.

The solution

Showing AWSAzuresovereign / on-prem services. Switch with the toggle; every diagram and table on this page follows it.

Impact reporting from case notes: reference architecture
  1. 01Channels

    • Caseworkers' notes
    • Outcome questionnaires
    • Funder reporting templates
  2. 02Intake

    • API GatewayAPI ManagementKong, or your existing API gatewayScheduled extract
    • Comprehend PII detectionAzure Language in Foundry Tools (PII)Microsoft Presidio (open source)De-identification
    • S3 and Amazon AthenaMicrosoft FabricPostgreSQL on KubernetesAnalysis store and figures
  3. 03Intelligence

    • SageMaker AIAzure Machine LearningMLflow and Kubeflow on KubernetesOutcome tagging model
    • SageMaker AIAzure Machine LearningMLflow and Kubeflow on KubernetesTheme discovery
    • Agent
      Bedrock AgentCoreFoundry Agent ServiceLangGraph on Kubernetes, with open-weight models via vLLMReport drafting assistant
  4. 04Orchestration

    • Step FunctionsLogic Apps / Durable FunctionsTemporal, or n8nReporting workflow
    • EventBridge + SQSEvent Grid + Service BusApache Kafka or RabbitMQEvents and queues
  5. 05Systems of record

    • Case management system
    • Outcome framework and consent records
    • Finance system
  6. 06People

    • Person
      Impact analyst checks every figure
    • Person
      Service leads review themes
    • Person
      Trustees approve the report
  • Cloud or platform service (AWSAzuresovereign / on-prem)
  • AI agent
  • A person decides
  • Your existing systems
  • Outside your organisation

Components, side by side

Impact reporting from case notes: components on AWS, Azure and a sovereign or on-prem stack
LayerAWSAzureSovereign / on-prem
Intake: scheduled extractA dated snapshot, taken through the case management supplier's published API or export. Agree in discovery what the supplier allows.API Gateway, S3API Management, Blob StorageKong or your existing API gateway; Ceph or IONOS S3 Object Storage
De-identificationOff-the-shelf detection misses local place names, nicknames and rare details. Add rules for them and test on real notes.Amazon Comprehend PII detection, plus the charity's own rulesAzure Language in Foundry Tools (PII detection), plus the charity's own rulesOpen-source PII detection (e.g. Microsoft Presidio), plus the charity's own rules
Analysis store and figuresEach published figure is a named, versioned query with its snapshot date and suppression rule. The language model never calculates a number.Amazon S3 and Amazon AthenaMicrosoft FabricPostgreSQL on Kubernetes
Outcome tagging (predictive ML)A classifier trained on notes coded by caseworkers against the charity's outcome framework. It tags notes, never people, and its accuracy is reported by service.Amazon SageMaker AIAzure Machine LearningMLflow and Kubeflow on Kubernetes
Theme discoveryAmazon Comprehend topic modelling closed on 30 April 2026. Themes are prompts for service leads, not findings.Open-source topic modelling (e.g. BERTopic) on SageMaker AIOpen-source topic modelling (e.g. BERTopic) on Azure Machine LearningOpen-source topic modelling (e.g. BERTopic) on Kubernetes
DraftingBoth clouds offer a choice of models, including Claude. Drafts insert approved figures; they don't generate them.Amazon BedrockMicrosoft FoundryOpen-weight models served with vLLM on your own GPUs, or IONOS AI Model Hub
Methods, framework and case studies (retrieval)The outcome framework, definitions, last year's report and consented case studies, so drafts use the charity's own terms.Bedrock Knowledge BasesFoundry IQ, built on Azure AI SearchOpenSearch, or PostgreSQL with pgvector
Agent runtimeThe original Bedrock Agents closed to new customers on 30 July 2026; new builds use AgentCore.Amazon Bedrock AgentCoreFoundry Agent ServiceAn open-source agent framework (e.g. LangGraph) on Kubernetes
WorkflowRuns each reporting cycle: snapshot, de-identify, tag, calculate, suppress, draft, review.AWS Step FunctionsLogic Apps or Durable FunctionsTemporal, or n8n
Middleware and eventsAmazon EventBridge, SQSEvent Grid, Service BusApache Kafka or RabbitMQ
Human reviewReviewers see each figure beside its query, its snapshot and the tagging accuracy for that service.In the impact team's existing reporting tools and documentsIn the impact team's existing reporting tools and documentsIn the impact team's existing reporting tools and documents
GuardrailsBlocks drafts containing a number not in the approved figures, a name, or a quote without a consent record.Bedrock GuardrailsFoundry guardrails (Azure AI Content Safety)Open-source guardrails (e.g. NeMo Guardrails, Llama Guard)
Identity, keys, data protectionThe re-identification key never leaves the case management side. Analysts see de-identified data only.IAM Identity Center, KMS, MacieEntra ID, Key Vault, PurviewKeycloak or your existing Active Directory; OpenBao or HashiCorp Vault; an HSM
Monitoring and auditLog every snapshot, model version, query, draft, edit and approval for each report.CloudWatch, CloudTrailAzure Monitor, Log AnalyticsPrometheus, Grafana, OpenTelemetry
UK data residencyIONOS AI Model Hub runs in German data centres: EU, not UK, residency. Commissioners' and funders' contracts may require UK processing.eu-west-2 (London)UK South / UK WestYour own UK data centre, or a UK or EU provider

On a phone, the table shows the platform you pick. Sovereign / on-prem means open, portable components you can run in your own data centre or with a UK or European provider; IONOS is shown as one example, and others exist. What sovereign cloud means →

Hybrid and on-prem

Many organisations keep core systems and sensitive data on their own servers. This is how the same design works across both.

Stays on-prem

The case management system, identifiable notes and the re-identification key stay where they are, whether on the charity's own servers or with its supplier. De-identification runs on that side, so identifiable notes never reach the AI.

Uses cloud AI

Outcome tagging, theme discovery, figures and drafting run on de-identified data in a UK cloud region, or on the charity's own GPUs if the trustees decide notes shouldn't leave its estate even in de-identified form.

How they connect

A scheduled, encrypted transfer over a site-to-site VPN, or AWS Direct Connect or Azure ExpressRoute for larger charities. Only de-identified batches cross the link, and results come back as figures and drafts, never as changes to records.

If the link drops

Nothing a service user depends on stops: casework carries on as normal. Reporting is a batch job with deadlines weeks away, so the next snapshot simply runs when the link returns.

The agents

What each agent may do through its tools, what it may not, and where a person signs off. The limits are enforced by the platform, not the prompt.

Report drafting assistant

May do

  • Read the approved, suppressed figures for the reporting period
  • Look up the outcome framework, definitions and last year's report
  • Draft narrative sections for the trustees' report and funder reports
  • Use case studies and quotes that have a current consent record
  • List changes from last year and the limits of the data

May not do

  • Calculate, change or estimate a figure
  • Open identifiable case notes
  • Score or rank an individual service user, caseworker or team
  • Write anything back to a case record
  • Publish a report or send it to a funder

Where a person signs offThe head of impact checks every figure and draft. Trustees approve the annual report, and the named grant holder approves each funder report. The agent's permissions are enforced by the platform's identity and policy controls, not by instructions in a prompt.

Middleware and integration

The case management system stays the record. A scheduled job takes a dated snapshot through the supplier's published API or export, de-identifies it on the charity's side, and sends only the de-identified batch onward. Nothing is ever written back.

Figures have a lineage. Each one names its query, snapshot date, outcome framework version, model version and suppression rule, so any number in a published report can be reproduced or corrected later.

The AI is optional. If it's paused, the impact team runs the same queries on structured fields and writes the narrative themselves, exactly as before, with less of the notes' detail.

Process and people

  • The outcome framework and its definitions are owned by the head of impact and approved by trustees. Changes are versioned.
  • Caseworkers code a sample of notes each quarter, which is used to train the tagging model and to check its accuracy.
  • A small-number rule, for example no figure below 10, applied before anything leaves the impact team.
  • A written rule, approved by trustees, that outputs are never used to judge individual staff or service users.
  • Caseworkers and service users are told how notes are used for reporting, in plain language, in the privacy notice.
  • A data breach plan that covers the analysis store and the AI suppliers, with trustees deciding on serious incident reports.

Delivery plan

  1. 3–4 weeks

    Discovery

    Map today's reporting, agree the outcome framework, audit note quality and consent records, and confirm what the case management supplier allows.

  2. Gate · must be true to passOutcome framework agreed, DPIA started, extract route confirmed, suppression rule set.

  3. 8–10 weeks

    Pilot

    One service, last year's notes: de-identify, tag, calculate and draft, then compare with a human-coded sample and last year's reports.

  4. Gate · must be true to passDe-identification tested on real notes, tagging accuracy above the agreed threshold, every pilot figure reproducible, DPO sign-off.

  5. 4–6 weeks

    Production

    Live for that service's funder reports, with monitoring, support and a rollback plan.

  6. Gate · must be true to passMonitoring, support model, rollback plan and impact team training in place.

  7. Ongoing

    Scale

    Add services one at a time, then the trustees' annual report impact section for the year.

  8. Gate · must be true to passEach new service passes the same pilot gate; trustees see the method before the first annual report.

Roles and RACI

Who is responsible, accountable, consulted and informed, by activity
ActivityDirector of servicesHead of impactSolutions architectI can fill thisTechnical PMI can fill thisEngineersI can fill thisDPOTrustees
Business case and benefitsAccountableResponsibleConsultedConsultedInformedConsultedInformed
Outcome framework and suppression rulesConsultedAccountable and ResponsibleConsultedInformedConsultedConsultedConsulted
Target architectureInformedConsultedAccountable and ResponsibleConsultedConsultedConsultedInformed
DPIA and privacy noticeConsultedConsultedConsultedInformedInformedAccountable and ResponsibleInformed
Build, de-identification and integrationInformedConsultedAccountableResponsibleResponsibleConsultedInformed
Pilot and accuracy testingConsultedAccountableConsultedResponsibleResponsibleConsultedInformed
Trustees' report approvalResponsibleResponsibleInformedInformedInformedConsultedAccountable
Run, monitor and improveConsultedAccountableConsultedConsultedResponsibleConsultedInformed

Business case and benefits

Director of services
A · Accountable
Head of impact
R · Responsible
Solutions architect
C · Consulted
Technical PM
C · Consulted
Engineers
I · Informed
DPO
C · Consulted
Trustees
I · Informed

Outcome framework and suppression rules

Director of services
C · Consulted
Head of impact
A/R · Accountable and Responsible
Solutions architect
C · Consulted
Technical PM
I · Informed
Engineers
C · Consulted
DPO
C · Consulted
Trustees
C · Consulted

Target architecture

Director of services
I · Informed
Head of impact
C · Consulted
Solutions architect
A/R · Accountable and Responsible
Technical PM
C · Consulted
Engineers
C · Consulted
DPO
C · Consulted
Trustees
I · Informed

DPIA and privacy notice

Director of services
C · Consulted
Head of impact
C · Consulted
Solutions architect
C · Consulted
Technical PM
I · Informed
Engineers
I · Informed
DPO
A/R · Accountable and Responsible
Trustees
I · Informed

Build, de-identification and integration

Director of services
I · Informed
Head of impact
C · Consulted
Solutions architect
A · Accountable
Technical PM
R · Responsible
Engineers
R · Responsible
DPO
C · Consulted
Trustees
I · Informed

Pilot and accuracy testing

Director of services
C · Consulted
Head of impact
A · Accountable
Solutions architect
C · Consulted
Technical PM
R · Responsible
Engineers
R · Responsible
DPO
C · Consulted
Trustees
I · Informed

Trustees' report approval

Director of services
R · Responsible
Head of impact
R · Responsible
Solutions architect
I · Informed
Technical PM
I · Informed
Engineers
I · Informed
DPO
C · Consulted
Trustees
A · Accountable

Run, monitor and improve

Director of services
C · Consulted
Head of impact
A · Accountable
Solutions architect
C · Consulted
Technical PM
C · Consulted
Engineers
R · Responsible
DPO
C · Consulted
Trustees
I · Informed

R ResponsibleA AccountableC ConsultedI Informed· not involved

Where I'd fit

Option A

Lead the architecture

If there's no architect on the programme, I'd own the design end to end: the extract from your case management system, de-identification, the figures and their lineage, the models, the drafting and the gates. I'd report to your director of services and make sure trustees can see how every figure was made.

Option B

Support your architect

If your partner or your own team already has an architect, I'd take on the data pipeline, de-identification and the integration with your case management system under their lead, and give them a second senior opinion on data protection and the model testing.

Where this goes wrong

  1. Goes wrong

    The language model writes numbers, and one is wrong in the trustees' report.

    How the design avoids it

    Figures come only from versioned queries. Guardrails block any number that isn't in the approved set.

  2. Goes wrong

    De-identification misses local details, and a small service's story identifies someone.

    How the design avoids it

    Charity-specific rules on top of standard detection, testing on real notes, and a small-number rule before anything leaves the team.

  3. Goes wrong

    Tagging works for one service and quietly fails for another.

    How the design avoids it

    Accuracy reported by service against a human-coded sample, and services added one at a time.

  4. Goes wrong

    Outcome data drifts into judging caseworkers.

    How the design avoids it

    A trustee-approved rule, no staff or team scoring in the design, and aggregate outputs only.

  5. Goes wrong

    Thin notes produce confident-looking figures.

    How the design avoids it

    Report coverage alongside every figure, and say plainly in the report what the data can't show.

On the horizon

Changes already announced or under way that affect this design. More in the UK horizon scan.

  1. Year ends from 31 December 2026
    The first trustees' annual reports under Charities SORP 2026, due within 10 months of the year end. Tier 2 and Tier 3 charities must explain their impact, not just list activities.

    Run the pipeline in shadow on this year's notes now, and agree the outcome measures with trustees before the year closes.

  2. Late 2026 to 2027
    The ICO's final guidance on anonymisation and pseudonymisation for research, archiving and statistics, after consultation from August to October 2026, alongside its updated guidance on the research and statistics provisions.

    Keep de-identification and suppression rules as configuration, so they can be tightened without a rebuild.

  3. 2026–2027
    The ICO's statutory code of practice on AI and automated decision-making, required by regulations in force since May 2026.

    Keep the records that show outputs are aggregate and never used for decisions about individuals.

  4. Late 2026–2028
    The Cyber Security and Resilience Bill, still going through Parliament, would bring medium and large managed service providers into regulation, with duties set by secondary legislation.

    Expect your IT, case management and AI suppliers' duties to change, and review contracts and the breach plan with them.

Questions to ask your partner or vendor

Good answers are specific. Vague answers to any of these are worth following up.

Copy them into your next supplier meeting

  1. Can every figure in the report be reproduced from the records, and by whom?
  2. How accurate is the tagging for each service, and how do you know?
  3. What stops a small number or a detail from identifying someone?
  4. How is it guaranteed that nobody, staff or service user, gets scored?
  5. Where are the notes processed, and is anything used to train anyone's model?
  6. Who owns the outcome framework, models and queries after you leave?

Planning something like this?

Whether you need someone to lead the design or to support the architect you already have, let's talk it through.

Sam Naseem
Sam NaseemBirmingham, UK · UK & Europe
Available now