Guides to the medication journey.
These guides explain how post-discharge follow-up, medication adherence and pharmacy call handling actually work — the mechanisms, the failure modes, and the cases where automation is the wrong answer. They are written to be useful whether or not you ever buy anything from OneDose.
Where a guide would normally quote a statistic we cannot attribute to a source, it says so instead. That is a deliberate policy, explained in each guide.
The guides
How AI reduces 30-day readmissions
The causal chain from a missed dose to a readmission, where AI actually intervenes in it, and what the intervention cannot fix.
Improving medication adherence at 90 days
Adherence does not decay evenly. Why the 90-day mark is where chronic therapy is won or lost, and what actually moves it.
What is continuity-of-care automation?
A definition, the handoffs it covers, and the distinction between automating continuity and automating the clinical decisions inside it.
Transitions of care: the nurse-time problem
Why post-discharge follow-up is rationed, why the rationing is rational, and why that makes it an arithmetic problem rather than a clinical one.
How to cut pharmacy inbound call volume with AI
What is actually in the queue, why IVRs and callbacks do not reduce it, and the difference between deflecting a call and resolving it.
Are AI voice agents HIPAA-compliant?
Compliance is a property of a deployment, not a badge on a product. What HIPAA actually requires of a voice agent, and the questions to ask a vendor.
What clinical questions should a voice agent escalate?
The escalation boundary is the most important design decision in a pharmacy voice agent. Where to draw it, and why it should be drawn conservatively.
Why do these guides have no statistics in them?
Because the numbers you would expect — readmission reduction, adherence lift, share of calls handled — are figures OneDose cannot currently attribute to a named deployment or a published study. Publishing them anyway would make these guides less useful, not more: a statistic without a source is not evidence, it is decoration that happens to have a decimal point.
So each guide argues from mechanism, and where a number genuinely belongs, it says what it cannot tell you and why. When a figure can be attributed, it will appear with the deployment attached.