Joe Salmon - Business & Performance Business Partner Application
Public aggregate demonstration reports
Six worked examples of agent-assisted analytical briefs using public aggregate NHS data for Dorset HealthCare.
Each brief is a first draft for human review. They are not official Dorset HealthCare reports and should not be used for operational decision-making.
The purpose of this page is to show how agentic AI could support business and performance work when it is used with clear sources, reproducible scripts, explicit caveats and human sign-off.
A Business & Performance Business Partner needs to turn data into clear, useful performance intelligence.
That means more than producing numbers. It means explaining:
These examples show how an AI agent could help produce a structured first draft, while keeping the human responsible for validation, judgement and sign-off.
Each brief follows a slimmed, review-first structure:
This creates a clear audit trail from public data to draft narrative.
Agent-assisted brief
First-draft performance brief from public NHS Oversight Framework data with enriched standard/peer/trend/validation columns. Priority flags for long-stay inpatients, UCR, crisis access and cost index.
View worked exampleAgent-assisted brief
Six-month MHSDS access and activity profile (Provider/RDY rows only). Stock-vs-activity insight, MHS69 validation flag and desired-direction column in the trend summary.
View worked exampleAgent-assisted brief
CSDS community activity profile with MoM-up / six-month-down headline pattern, “Other” category data-quality flag and CHS waiting-list signpost.
View worked exampleAgent-assisted brief
NHS Talking Therapies access and waits brief with 75% six-week and 95% eighteen-week standards, self-referral insight and corrected M019–M022 waiting-band totals. See how a Report Analysis Agent would review a flawed draft before publication.
View worked exampleSource map
A source applicability map for KO41a, ERIC, DSPT, FFT and CQC context — not a scorecard. Traceable KO41a/ERIC values, currentness-risk column and who-to-contact guidance.
View worked exampleApplicability check
Public source applicability check for A&E, DM01 and KH03. Explains grouping rationale, audiology waiting-list dominance and why A&E rows validate service model rather than ED performance.
View worked exampleThe briefs were created using a human-directed agent workflow.
Public NHS sources were downloaded and catalogued in DATA_SOURCE_REGISTER.csv.
R scripts filtered the relevant Dorset HealthCare rows and created processed public-data extracts.
The report rendering script then produced each brief using the same structure: question, data used, key findings, agent summary, human checks and verification notes.
This means the reports are not just manually written text. They are part of a repeatable workflow that links source data, processing and narrative.
Every brief ends with caveats and a human review gate.
Supporting evidence: regenerate briefs with Rscript site/R/03_render_public_reports.R (runs post-render validation automatically). Update processed extracts in public-data/processed/ first. Standalone validation: Rscript site/R/04_validate_public_reports.R after sourcing from the render script.
Performance reports are most useful when they are clear, accurate and honest about uncertainty.
Agentic AI can help with the first draft: organising the data, structuring the findings, drafting the narrative and highlighting caveats.
But the value comes from the governed workflow around it.
In a live NHS setting, an accountable person would still need to confirm the data source, check the definitions, validate the figures, agree the interpretation and decide what action is needed.
This page shows how AI could support that process without replacing professional judgement.