SGA · Production Hardening Sprint

From Prototype
to Production.

The need to take our 20+ prototypes and develop them into fully automated production systems.

SGA Dental Partners · Scott Guest
Briefing for Karen Flug & Myles · May 13, 2026
01 · Current Projects
SGA · Production Hardening Sprint
02 / 11
02 · Today's Reality
The structural problem

Scott is the plumbing.

Every project above ground runs because I manually trigger it. Nothing is monitored. Nothing has a handoff path. As the portfolio grows, the burden compounds.

ABOVE GROUND · 20+ PROJECTS IN USE TODAY S-curve Karen · revenue/EBITDA Headcount staffing levels Labor Analysis hours vs production Doctor Barometer Paul · daily goal RIS · Replacement provider risk score HIS · Hiring Index pipeline health Doctor Days Audit network-wide days Compounder Mitch & Paul Fee Negotiations Pam · RCM Net Production Budget Ops · Finance Historical Pacing day-by-day trajectory OSA / Next 3rd capacity signal ROD Specialty specialty referrals Hygiene Dashboard hygiene KPIs Neurality Effect. AI voice agent Friday Analysis open Fridays Benefits Cluster benefits plans Schedule Designer provider templates Zoho Ticketing Ron · IT + more in flight audits · spikes · pilots GROUND LEVEL BELOW GROUND · INVISIBLE · HAND-CRANKED · ONE PERSON SCOTT Hand-cranking 20+ pumps every morning. No monitoring · No alerts · No handoff path PUMP PUMP
SGA · Production Hardening Sprint
03 / 11
03 · Structural Risk
If we don't build the plumbing

Today this looks like velocity. In six months it looks like collapse.

The portfolio is currently a stack of prototypes one person can carry. As the stack grows, the carry gets heavier — and the failure surface gets wider.

RISK 01 · PEOPLE
Bus Factor
Single point of failure
What breaks
If Scott is out for a week, half the reports stop. Nobody else can pick them up — nothing is documented or owned.
Today's mitigation
Scott doesn't take vacations.
RISK 02 · QUALITY
Silent Failures
No monitoring layer
What breaks
When a scraper breaks or a refresh fails, nobody knows until the meeting where the data is wrong. Trust in the data erodes.
Today's mitigation
Scott checks each system manually each morning.
RISK 03 · SCALE
No Compounding
Stack grows fragile
What breaks
Every new project adds load to the existing stack. In six months, Scott firefights prototypes instead of building new value.
Today's mitigation
Working longer hours. Not sustainable.
SGA · Production Hardening Sprint
04 / 11
04 · The Blueprint
What we need to build · the plumbing underneath

Build the plumbing once. Run on it forever.

Products are what Karen sees. The plumbing is what makes them run. Today, the bottom three layers don't exist as a real system — they exist in Scott's head. Click each layer to focus on it.

↑ ABOVE GROUND · PRODUCTS PEOPLE USE
↓ BELOW GROUND · THE PLUMBING (click each layer)
DAX Query System
635 validated measures · organized by domain · single source of truth for every product above ground
PARTIAL · Needs validation
Data Sources
Power BI · Dental Intel · PeerLogic · Neurality · Historical Pacing · S-Curve
LIVE · Connected
Automation Platform
Schedulers · retries · alerts · health monitoring  ·  This is the keystone.
NOT YET BUILT
Secure Environment
Auth · audit logs · key management · BAA-ready for PHI Phase B
NOT YET BUILT
SGA · Production Hardening Sprint
05 / 11
05 · Product · Headway
Product 01 · the people system

Managing our people in real time.

Human intelligence — at the practice, region, and network level.

Product · Beta
Headway
Capacity + Momentum: Making headway through smart people decisions.
A real-time view of hiring, retention, and capacity in one operating picture. Built from four signals — replacement risk, hiring pipeline, headcount, and labor cost — combined into one human-intelligence layer for HR, RODs, and OMs.
FOUR BUILDING BLOCKS FEED IN
01
Replacement Index (RIS)
Provider replacement risk — who's at risk of leaving, ranked.
02
Hiring Index (HIS)
Time-to-hire & pipeline health by role and region.
03
Headcount Analysis
Staffing levels by practice, role, and tenure band.
04
Labor Analysis
Hours vs. production — efficiency signal per provider.
WHAT IT ANSWERS
"Where are we understaffed?" · "Who's at risk of leaving?" · "Can we afford to hire?" · "Labor vs production — by region?"
SGA · Production Hardening Sprint
06 / 11
06 · Product · Daybreak
Product 02 · daily decision engine

The day's actions, written by AI, delivered every morning.

Sunrise Action: Daybreak tells every OM and ROD exactly what to do today — with a feedback loop that makes tomorrow's brief sharper. → sga-daybreak-v2.pages.dev

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 delta 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). The feedback loop is built in — tomorrow's brief is sharper than today's.
SGA · Production Hardening Sprint
07 / 11
07 · Product · Lookout
Product 03 · self-serve, ad-hoc reporting (future)

See for yourself — without waiting on Scott.

Lookout becomes possible only after the plumbing is real. It's the leverage payoff.

Product · Future
Lookout
Ad-hoc questions. Validated answers. Instant.
Anyone in the org asks a question — Mitch, Pam, Ron, an ROD — and Lookout assembles a validated answer from pre-built DAX building blocks. Drag-and-drop assembly. Instant answers. No Scott. No tickets. No waiting.
HOW IT WORKS · WHAT IT UNLOCKS
01 · ASK
Anyone asks a question
"Production by region, last 12 weeks." · "Which payers paid slowest this month?" · "Open tickets > 7 days, by location."
02 · ASSEMBLE
Pre-validated DAX blocks
Drag-and-drop pieces from the validated DAX query library. No raw SQL. No guessing definitions.
03 · ANSWER
Validated, instant
Drillable report rendered in seconds. They own it. They re-run it. Scott never gets pulled in.
TODAY VS WITH LOOKOUT
3 questions = 3 Slack messages → 3 days. With Lookout: 3 questions = 3 self-serve queries → 3 minutes.
SGA · Production Hardening Sprint
08 / 11
08 · The Ask
What I need · from Myles and Karen

Three to four weeks. Focused time. One sprint.

A moratorium on new project requests so the existing portfolio gets a real foundation.

01

Validate all DAX queries

635 measures. Confirm definitions, alignment with PBI source of truth, and reusability as building blocks across every product above ground.

02

Build the automation infrastructure — the keystone

Schedulers, retries, alerts, health monitoring. This layer has never been built. It is what makes everything else durable. Without it, nothing else matters.

03

Wire the existing 20+ projects to the new infrastructure

Headway, Daybreak, every individual report. They stop being hand-cranked and start being real systems.

04

Documentation, ownership, & handoff

Runbooks. Ownership maps. Clear escalation paths. The next person can run this — and Scott can take a vacation without the portfolio collapsing.

SGA · Production Hardening Sprint
09 / 11
09 · Sprint Plan
Three to four weeks · what gets done

Validate. Build. Wire. Hand off.

The red Automation Platform bar spans the full sprint — that's the keystone everything else hangs on.

WORKSTREAM
Week 1 · Validate
Week 2 · Build
Week 3 · Wire
Week 4 · Ship
Validate 635 DAX queries
DAX validation sprint
Document data sources
Source inventory + access
BUILD automation infra
AUTOMATION PLATFORM · the keystone · never built before
Secure environment
Auth + audit + key management
Wire Headway
Headway on the new plumbing
Wire Daybreak
Daybreak on the new plumbing
Migrate 17 individual tools
Migrate prototypes to infra
Monitoring + alerting
Health checks live
Docs + handoff prep
Runbooks + ownership
SGA · Production Hardening Sprint
10 / 11
10 · After the Sprint
The payoff · what 3–4 weeks of focus actually buys

The next year is leverage, not firefighting.

Outcome 01

Scale & leverage

Scott stops being the pump. New work no longer crushes existing work. The portfolio compounds instead of decaying. Each new project gets cheaper because it inherits the plumbing.

Outcome 02

Real monitoring

Failures alert. Health is visible. Nobody finds out about a broken refresh in the meeting where the data is wrong. The data layer is trustworthy by default.

Outcome 03

Ownership & handoff

Documented systems. Clear runbooks. A second engineer can take over any tool. Scott can take time off — the data still flows. The org doesn't depend on one person.

Outcome 04

Lookout becomes possible

Once building blocks are validated and automated, the team can run their own analyses. Self-serve reporting moves from "someday" to "this quarter." The next product gets built on the foundation, not on Scott.

SGA · Production Hardening Sprint
11 / 11