AXIOM · Sales enablement · Astrodata × Omni

AXIOM healthcare demo guide.

A walkthrough and talk track for Omni sales in healthcare conversations. AXIOM is the proof of what’s possible when Tuva, Snowflake Cortex, and Omni come together.

Open the live demo

Or read the walkthrough below first, the guide doubles as your script. Prefer to watch? Here's a video walkthrough.

Her problem	AXIOM’s answer	Powered by
Doesn’t know who to call first	ML-ranked worklist by readmission probability	Cortex ML.CLASSIFY on Tuva data
Info scattered across 4 systems	Single patient profile with everything	Tuva’s 6 data marts normalized
No predictive signal	30-day readmission risk score, refreshed nightly	Cortex ML, native, no MLOps
Starts from scratch each call	AI-assisted care summaries and recommendations	Omni narratives
No cost or quality visibility	Population dashboard with anomaly detection	Cortex Anomaly Detection + Omni
Can’t ask ad-hoc questions	Conversational analytics across the data	Omni conversational analytics

Her problem

AXIOM’s answer

Doesn’t know who to call first

ML-ranked worklist by readmission probability

Cortex ML.CLASSIFY on Tuva data

Info scattered across 4 systems

Single patient profile with everything

Tuva’s 6 data marts normalized

No predictive signal

30-day readmission risk score, refreshed nightly

Cortex ML, native, no MLOps

Starts from scratch each call

AI-assisted care summaries and recommendations

Omni narratives

No cost or quality visibility

Population dashboard with anomaly detection

Cortex Anomaly Detection + Omni

Can’t ask ad-hoc questions

Conversational analytics across the data

Omni conversational analytics

05 · Screen by screen

What to show. What to say.

Screen 01

Population dashboard

Sarah’s morning view. The dashboard tells her where the problem is before she has to look for it.

AXIOM Population Health Intelligence dashboard with a regional × condition risk-density heatmap, top-line KPIs, and an AI Summary panel.

What the buyer is seeing

Population-level risk stratification across roughly 60,000 attributed lives. Top-line KPIs: 14,374 high-risk patients, $1,559 PMPM, 4.2 percent 30-day readmission rate. A risk-density heatmap breaks the population by region and chronic condition. An AI Summary panel on the right flags the highest-risk cohorts and the most urgent patterns. Click any cell to drill into that cohort.

What to say

Lead with

"Sarah doesn’t open three reports anymore. She opens this. Snowflake Cortex flagged the anomaly, and Omni wrote the summary and surfaced it all in one view."

If asked "What’s the metric here?"

PMPM means per-member-per-month spend. It’s how risk-bearing organizations track cost per patient.

If asked "Is this real-time?"

Refreshed nightly. Cortex re-runs the risk model and anomaly detection against the full cohort overnight.

Don’t promise

Live streaming. The demo data is synthetic and refreshed in batch.

Partner moment

Cortex flagged the anomaly. Omni wrote the narrative and rendered the surface. Tuva normalized the data underneath. Three partners in one view, no glue code required.

Screen 02

Cohort analysis

CHF cohort deep dive. Cost drivers, utilization patterns, and an AI-generated narrative.

What the buyer is seeing

Sarah clicked into the flagged CHF cohort. Top-line KPIs: cohort lives, PMPM, 30-day readmission rate, and average length of stay. A PMPM trend chart, cost decomposition by type showing where the spend is concentrated, and an Omni-generated narrative explaining the cost drivers. Blobby is active in the sidebar for follow-up questions, and a Create Campaign button is ready for the next step.

What to say

Lead with

"This is the moment Sarah goes from ‘who is at risk’ to ‘what is actually happening.’ She didn’t write SQL. She didn’t open a second tool. Omni took her one click deeper and wrote the explanation."

If asked "Where did the narrative come from?"

Omni’s AI layer, running against the Tuva data marts. The narrative regenerates as the cohort changes, no canned reports.

If asked "What’s the biggest cost driver?"

The cost decomposition chart shows the breakdown by type. The Omni narrative below the chart explains what’s driving the spend and where intervention priorities should focus.

Partner moment

Omni-generated commentary against the Tuva data foundation, rendered inside the same view as the chart. The whole "explain this chart to me" experience is one integration.

Screen 02·A

Create campaign

The writeback moment. Where Omni stops being a dashboard and becomes an action layer.

What the buyer is seeing

Sarah scoped a cohort and launched a targeted outreach campaign against the high-cost patients she just identified. KPIs, engagement funnel, and recent activity in one modal. The campaign writes back to her care team’s operational system.

What to say

Lead with

"This is the writeback story. Sarah scoped a cohort in Omni and the campaign she launched here is actively pushing interventions back into her team’s tools. Omni isn’t a read-only BI layer in this workflow, it’s a workflow tool."

If asked "What’s the writeback?"

The ability for the App to push structured data back into source systems from the Omni-driven analysis. Care campaigns, intervention assignments, status updates, all originated in the analytics layer.

If asked "Does this integrate with our existing systems?"

The writeback target is configurable. In a production deployment we’d connect it to the customer’s care management platform of record.

Partner moment

This is where Omni stops being a dashboard and starts being a workflow tool. Worth lingering on.

Screen 03

High-risk patient worklist

Sarah’s daily queue. Cortex ranks the panel by 30-day readmission probability.

What the buyer is seeing

Sarah’s active CHF panel, ranked by Cortex readmission probability. Filter controls for risk tier, region, and condition scope the view to her Southeast CHF patients. Each row shows the patient, their risk score, current medication, adherence, and tier. Click any patient to open their full profile.

What to say

Lead with

"Omni dashboards showed Sarah the cohort. Cortex tells her where to start. The model running underneath this list is Snowflake Cortex ML, classifying every patient on her panel by 30-day readmission probability and re-ranking nightly."

If asked "What’s HCC?"

Hierarchical Condition Category. A risk-adjustment score weighting medical complexity. Higher score, more clinically complex patient.

If asked "What makes Cortex’s ranking better?"

Today most care managers prioritize by gut, recency of discharge, or a supervisor’s flag. Cortex ranks by a trained model against the full longitudinal claims record. It’s the difference between guessing and knowing.

Don’t promise

Per-patient AI explanations inside the row expansion. The care summaries are scoped to the pathway, not generated per individual.

Partner moment

"Ranked by Snowflake Cortex readmission probability" appears directly under the worklist title. The clearest partner attribution moment in the app, point to it.

Screen 04

Patient profile

The full picture, in one place. Then take action.

What the buyer is seeing

One screen for the top-ranked patient, Jacob Jimenez: header KPIs (readmit risk, HCC score, average monthly cost, attributed PCP), a paginated clinical timeline showing the full encounter history, and a Blobby Narrative at the bottom summarizing his clinical and financial picture in plain language.

What to say

Lead with

"This is the screen Sarah lives in. Everything she used to assemble from four systems, claims, labs, pharmacy, quality measures, is in one Omni view, normalized by Tuva. The narrative summarizing her clinical and financial picture was generated by Omni."

If asked "What’s a HEDIS gap?"

HEDIS is the standard quality measure set health plans use to demonstrate care quality. A gap means a measure that hasn’t been satisfied yet, a missed screening, an overdue lab. Closing gaps proactively is one of Sarah’s core KPIs.

If asked "What’s the writeback?"

When Sarah hits Approve & Deploy, the App writes the intervention plan back into her team’s operational system. Care coordinators, pharmacists, and the attributed PCP all see assignments in the tools they already use.

Don’t promise

That the AI narrative is generated per-patient in real time. For the demo, the prose is templated around the pathway. Per-patient generation is part of the production build.

Partner moment

Three partners stacked on one screen. Tuva normalized the data. Cortex ranks the risk. Omni renders the canvas, writes the narrative, and triggers the writeback.

Screen 05

Care pathway live

Jacob Jimenez’s pathway, projected outcomes, and a cohort comparison. Patient-level action meets population-level context.

AXIOM care pathway for Jacob Jimenez showing projected outcomes and cohort comparison.

What the buyer is seeing

A projected-outcomes header: if Jacob completes the pathway, his 12-month cost drops by $744 against a $4,037 trajectory, his 30-day readmit risk falls from 77 to 54 percent, and three tasks remain to assign. Below, his CHF Post-Discharge Transition pathway log with step-by-step progress, a Schedule Now button on the overdue medication reconciliation, and on the right, a cohort comparison showing how similar patients in the Southeast CHF panel are faring through the same protocol.

What to say

Lead with

"This is where the story resolves. Jacob’s pathway, his projected outcomes, his overdue step, all on one screen. Sarah can schedule the reconciliation right here. And on the right, she can see how similar patients in her panel are moving through the same protocol."

If asked "What can she do on this page?"

Schedule overdue steps directly (the Schedule Now button), view projected outcomes if the patient completes, and compare against the cohort on the right. She can also ask Blobby for projected savings or bottleneck steps.

If asked "Is this connected to the patient’s chart?"

In production, yes. The pathway integrates with the customer’s EHR or care management system of record. The demo shows the writeback shape; the production integration is part of the deployment.

If asked "How does the team know what to do?"

Each step has an assigned role, a target date, and the underlying intervention rationale. Care managers are accountable for completion; the pathway tracks status as the patient works through it.

Partner moment

The operational payoff. The story didn’t end at "here’s an insight," it ended at "here’s a patient with a projected outcome, an overdue step with a Schedule Now button, and a cohort benchmark for context."

Time	Action	What to say
0:00	Frame the buyer	"This is a care manager workflow. She’s an RN managing 200 high-risk patients. Her job is to prevent readmissions. Today she does it across four systems."
1:00	Anomaly to cohort	Show Population Dashboard. "Snowflake’s anomaly detection flagged this. Omni surfaced it. One click drills in."
2:00	Cohort to action	Show Cohort Analysis, then click Create Campaign. "Omni wrote the explanation. The campaign writes back to the care team’s tools."
3:00	The Cortex moment	Show Worklist. "Cortex ranked these 72 patients by readmission probability. This is the model running underneath every row."
4:00	The full picture	Open the top patient. "Everything from four systems in one view. Tuva normalized. Omni wrote the narrative. One click deploys the pathway."
5:00	The payoff	Show Care Pathway. "The interventions are now operational. Hospitalist, pharmacist, cardiologist, all assigned. That’s the integration story."

Time

Action

What to say