# Next best action design
> Recommendation logic your medical team can explain and defend.
Source: https://www.visfo.health/projects/next-best-action-design
Tags: Sprint

_Recommendation logic your medical team can explain and defend._

We design the rules that decide what happens next with a given HCP, need, objective weighting, recency and channel, and pressure-test them against real history before anything is automated.

## What it is

The decision logic behind the recommendation: given what we know about this clinician, what is the most useful next interaction, and why. Written as rules a medical reviewer can read, not a model nobody can interrogate.

Built before any platform work, so the thinking is agreed and tested rather than encoded by default.

## How we run it

1. **Frame** — Set the boundaries: what medical may act on, what it may not, and where compliance draws the line.

2. **Ingest** — Take the need map, objective weightings, engagement history and channel constraints as inputs.

3. **Synthesise** — Write the ranking logic and back-test it against historical interactions to see what it would have recommended and whether that holds up.

4. **Deliver** — Ship the rule set, the back-test read-out, the exception handling and an implementation spec if it goes on to a platform.

## What you get

- Documented recommendation logic in plain language

- Back-test against historical engagement with the disagreements explained

- Guardrails and exclusions agreed with compliance

- Explainability format: why this HCP, why now, why this topic

- Implementation spec ready for build if you take it further