SGA · VP of AI Portfolio · Proof Point 01

Operation
Daybreak

AI-written ops intelligence, delivered at dawn — to every tier that runs the network.

SGA Dental Partners · Scott Guest
Prepared April 22, 2026 · Internal
Agenda
What we'll walk through · Operation Daybreak

From raw ops data to tier-tuned briefs.

The audiences, the data, the engine, the loops, and what each tier actually gets in their inbox.

01
The Vision
One AI-written brief. Three audiences. Every morning before anyone logs in.
02
Three Audiences
OM, ROD, and Ops Leadership — what each tier needs and doesn't.
03
Data & Flow
Left: what streams in. Right: who gets what. Today and next.
04
The Decision Engine
Inside the Daybreak core — Ingest · Score · Rank · Write · Deliver.
05
The Three Loops
Feedback · self-improvement · self-healing — how the system gets sharper.
06
What Each Tier Sees
Sample outputs for OM, ROD, and Ops Leadership — side by side.
SGA · Operation Daybreak
02 / 09
01 · The Vision
What Daybreak Is
An AI-written ops brief, delivered every morning — tuned to the tier that receives it.
Practices covered today
248
OM Morning Brief live · runs ~5:30am daily
Audience tiers
3
OM · ROD · Ops Leadership
Data sources planned
5 +
PBI · Tessera · Pacing · OSA · S-Curve
Status
Live
Tier 1 (OM) operational · ROD + Ops next
The Problem We're Solving
Dashboards are built for one persona, then bent to serve three. OMs need today's actions, RODs need this week's patterns, leadership needs the network story. None of them have time to extract those views from raw BI. Daybreak writes each one, every morning, automatically.
Why It's A Daybreak, Not A Dashboard
A dashboard shows you yesterday. Daybreak tells you what to do today — and records whether you did it and whether it worked. That feedback sharpens tomorrow's brief. It's a teammate on a cadence, not a report.
SGA · Operation Daybreak
03 / 09
02 · Three Audiences
One engine, three tuned audiences
Different grain. Different format. Different cadence. Same underlying intelligence.

Each tier gets a brief written for their altitude — the detail an OM needs would drown a network exec; the pattern a leader needs doesn't help an OM tomorrow morning.

TIER 01 · PRACTICE
Office Manager
248 office managers · one brief per practice
Grain
Single practice · daily
Format
One-page Morning Brief — printable, pin-to-wall, mobile-readable. Headline + score + top 3 actions today + quote + principle.
Decisions It Drives
"What do I coach on at the huddle?" · "Who do I call?" · "What do I close out before end of day?"
Cadence
Daily
Length
1 page
TIER 02 · REGION
Regional Ops Director
~15 RODs · one brief per region (15–25 practices)
Grain
Region · weekly rollup + daily exceptions
Format
Regional dashboard — practice grid with tier semantics, movers list, weekly exceptions, visit-planning ranker. Where to spend Friday.
Decisions It Drives
"Which practice needs me this week?" · "Who's trending down quietly?" · "Where should I put pressure or support?"
Cadence
Daily + Wk
Practices
15–25
TIER 03 · NETWORK
Ops Leadership
Karen + ops leadership team · one brief per network
Grain
All 265 practices · weekly + monthly
Format
Exec brief — network health score, systemic signals, outliers, pacing-vs-target, S-Curve position, strategic asks.
Decisions It Drives
"What systemic issue is emerging?" · "Where are we ahead/behind plan?" · "What should we invest in next quarter?"
Cadence
Weekly
Scope
All 265
SGA · Operation Daybreak
04 / 09
03 · Data & Flow
How Daybreak is wired — data left, engine center, audiences right
Six source systems flow into one engine. Three audiences come out the other side.

Green = live today. Gold = next wave (Q2). The engine is identical for all three audiences — only the final writer and the delivery format differ.

DATA SOURCES DAYBREAK DECISION ENGINE AUDIENCE OUTPUTS Power BI Financials · production · collections · labor LIVE Tessera AI intelligence & signals platform LIVE Historical Pacing Day-by-day trajectory vs. prior years LIVE S-Curve Revenue + EBITDA target curves (Karen) Q2 OSA Online Schedule Availability — capacity signal Q2 Dental Intel Reappointment · recall · hygiene metrics Q2 + MORE CANDIDATES · SEE SIDEBAR DAYBREAK · DECISION ENGINE 01 · INGEST Pull all sources · normalize to practice-grain events 02 · SCORE 6-domain health score · trend · pacing delta 03 · RANK Top actions per entity (OM/ROD/Ops grain) 04 · WRITE Claude Opus 4.7 drafts narrative per tier 05 · DELIVER Email · print · dashboard — per audience MEMORY past decisions + outcomes feed Score & Rank TIER 01 OM Morning Brief 248 practices · one page · daily ● LIVE TODAY TIER 02 ROD Regional Brief ~15 regions · daily + weekly ◐ NEXT WAVE TIER 03 Leadership Exec Brief Network-wide · weekly ◐ NEXT WAVE ↻ FEEDBACK · actions & outcomes → MEMORY
SGA · Operation Daybreak
05 / 09
04 · The Decision Engine
Inside the Engine · how the morning's intelligence gets assembled
Five stages, left to right — every brief, every morning, the same pipeline.

The only variable is the tier grain (practice · region · network). The pipeline below runs once per tier per morning.

INPUT Source feed PBI · Tessera · Pacing OSA · S-Curve · DI ~5:30 AM · daily 01 Ingest Pull all sources Normalize to practice Time-align to yesterday CHECK: freshness · completeness source availability ~30 seconds 02 Score 6 health domains Production · Collections Schedule · Retention Hygiene · Labor COMPUTE: score · trend · pacing Δ vs. prior year + S-Curve ~2 minutes 03 Rank Prioritize actions Candidate actions → impact × confidence × tier-fit filter USES MEMORY: which actions worked for this OM previously? ~1 minute 04 Write Generate narrative Claude Opus 4.7 tier-specific prompt tone · altitude · length INCLUDES: quote · principle · trend rationale per action ~3 minutes 05 Deliver Route to audience OM · email + PDF ROD · dashboard Ops · exec brief TIMING: in inbox by 6:15 AM CT before huddle window ~90 seconds MEMORY Every past decision + outcome · feeds Score (trend), Rank (what worked before), Write (what resonated) ↻ OUTCOMES CAPTURED · FEEDBACK → MEMORY (next morning) TOTAL RUNTIME ~7 min per tier · parallel
01 INGEST
Pull all sources. Normalize. Fail loudly if anything is stale or missing.
02 SCORE
6-domain health score + trend + delta vs. S-Curve and prior-year pacing.
03 RANK
Generate candidate actions. Rank by expected impact × confidence × tier fit.
04 WRITE
Claude Opus 4.7 writes the narrative — quote, principle, rationale per action.
05 DELIVER
Inbox by 6:15 AM CT. OM prints it. ROD scrolls it. Ops forwards it.
SGA · Operation Daybreak
06 / 09
05 · The Three Loops
How Daybreak gets smarter · three loops, three cadences
Daybreak runs three loops simultaneously — each at a different speed.

Feedback (daily) corrects individual decisions. Self-improvement (weekly) sharpens the model. Self-healing (monthly) catches drift before it becomes damage.

Loop 02WEEKLY
Self-Improvement
The engine reads its own history and upgrades the ranker itself.
STEP 01DECISIONS STEP 02PATTERNS STEP 03MODEL UPD. STEP 04SHARPER SELF- IMPROVEMENT full model
Every Friday, Daybreak reads ALL week's decisions + outcomes, surfaces new patterns ("morning recalls move hygiene recovery in Region 3"), and updates the ranker. Day 30: decent. Day 90: sharp.
Loop 03MONTHLY
Self-Healing
Daybreak watches itself — and catches drift before anyone notices.
STEP 01PREDICT STEP 02OBSERVE STEP 03DRIFT? STEP 04RECALIBRATE SELF-HEALING drift correction
Daybreak compares its predictions against reality monthly. If accuracy drops (new comp system, Power BI schema change, unusual spike), it flags itself and pauses recommendations until recalibrated. No silent degradation.
Why This Matters
Without loops, Daybreak is an AI-written PDF. With loops, it's a teammate that improves alongside the OMs, RODs, and ops leaders using it — and catches its own mistakes before they embarrass anyone.
Activation Schedule
Loop 01 is live today — OMs rate actions via a thumbs rail. Loop 02 activates once we have 4+ weeks of labeled data. Loop 03 activates once baseline accuracy is stable enough to detect meaningful drift.
SGA · Operation Daybreak
07 / 09
06 · What Each Tier Sees
The same intelligence · three very different artifacts
One engine writes three briefs. Each tier receives exactly what's actionable at their altitude.

Mockups below. The OM brief is in production today; the ROD and Leadership briefs are designed and shipping next.

● TIER 01 · OMDAILY · 6:15 AM
OM Morning Brief
Stone Creek Dental · Memphis APR 22 · WED
78
Practice Health
▲ +3 vs. 7-day avg
6 DOMAINS
Today's top 3 actions
01
Call the 4 hygiene patients who no-showed Monday — all reappointable
02
Coach Jasmine on tx presentation — perio case value trending low
03
Confirm tomorrow's 3 high-value implant consults by noon
"You don't rise to the level of your goals. You fall to the level of your systems." — fix the huddle before the numbers.
Headline + score + 3 concrete actions + coaching quote + principle. Prints on one page. OM pins it to the huddle wall.
● TIER 02 · RODDAILY + WEEKLY
ROD Regional Brief
Region 03 · Mid-South 18 PRACTICES
92
88
81
79
76
71
70
68
67
65
58
55
52
49
42
38
–
–
–
–
This week's movers
Cedar Grove▲ +14
Elm Station▲ +9
Briar Mills▼ -11
Oak Ridge▼ -8
Regional health heatmap (tier-colored) + movers list + who-to-visit-Friday ranker. Daily glance; weekly deep brief.
● TIER 03 · OPSWEEKLY · MON AM
Leadership Exec Brief
Network Health · Apr 15–21
81.4
▲ +1.2 vs. prior week · ahead of S-curve
Systemic signals this week
Hygiene reappointment up network-wide — OSA roll-out in Region 7 driving it
Labor % drift in 11 practices — cluster in Southeast · investigate
3 practices have been sub-50 for 14+ days · escalate to ROD visit
ASK: Budget 2 days from RevOps to validate the Southeast labor cluster — may be a coding mismatch vs. real variance.
Network score + systemic signals + persistent outliers + one clear strategic ask. Ready-to-forward to Karen.
SGA · Operation Daybreak
08 / 09
07 · Roadmap & Commitments
Phased rollout · audiences · data sources · loops
Tier 1 is live. Two more tiers and four more data sources in the next quarter.

Each row below is a concrete commitment — who owns it, when it lands, and what status tag applies.

#DeliverableOwnerTargetStatus
1OM Morning Brief — 248 practices, daily at 6:15 AM CT, PBI + Historical Pacing + Tessera feeds.VP of AITodayLIVE
2Loop 01 (feedback) — OM thumbs rail on each action; outcomes written to Memory; ranker reads labels from Day 1.VP of AINowNOW
3ROD Regional Brief — regional heatmap, movers list, Friday-visit ranker. 15 regions.VP of AI4 weeksQ2
4S-Curve integration — pacing-vs-target delta surfaces in all three tier briefs. Ties Daybreak to Karen's $2M EBITDA runway.VP of AI + RevOps6 weeksQ2
5OSA integration — Online Schedule Availability as a leading capacity/demand signal in Score and Rank.VP of AI + OSA team8 weeksQ2
6Ops Leadership Exec Brief — weekly network brief for Karen + ops leadership. Tier 3 goes live.VP of AI10 weeksQ2
7Loop 02 (self-improvement) — weekly ranker retrain on accumulated labels; Loop 03 (self-healing) drift detection online.VP of AIQ3 W1Q3
8Dental Intel integration — reappointment & recall signals as inputs to Hygiene and Retention domains.VP of AIQ3 W3Q3
CANDIDATES IN THE HOPPER — Swell (patient reviews/sentiment), Modento (patient comms), Labor Analytics (timecard detail), SMILE Program (culture/engagement), Provider Level Budget (Gen4 CDM). Each is a candidate signal — we'll add in order of marginal value to Rank & Write, not all at once.
SGA · Operation Daybreak
09 / 09
1 / 9