Data Enablement · RACI Workshop

Operating Model · Value Chain · Responsibility Framework · April 2026

LIVE SESSION WEEK 3 Facilitated by Allata

The Operating Model

How Ascend's data delivery organisation works as a unit — clear roles, accountable functions, and the trifecta that produces trusted data products.

Frank Ingari — SVP Operations
COO Equivalent · Executive Business Champion · Data Governance Sponsor
Enterprise authority · Governance sponsor · Joint investment accountability with CTO
The enterprise authority layer. Cross-brand mandate — the only function positioned to align strategy, enforce governance, and hold the business accountable for outcomes across ATI, NHA, NASM, and Click Safety. Chairs the Data Governance Council, setting agenda and decision authority. Owns Enterprise Data Quality as a business-visible product, not a technical metric.

Joint investment accountability with Ash Siebecker (CTO): Ash holds the technology budget and the "what do we build" decision. Frank holds the business mandate and the "did it change anything" accountability. Neither can do the other's job — the Governance Council is where those two lenses meet. Frank is the business voice on whether outcomes are realised; Ash is the technology authority on what gets funded and built.

The difference between a RACI on paper and one that is enforced.
↓ Orchestrating function — owns intake, squad, delivery oversight and assurance ↓
Engineering
Ben York · Madhavi Varansi
Bronze → Silver → Gold pipeline
Data model design & Coalesce build
CI/CD, observability, platform standards
Data quality & data management
Data Architect — design thinking lead
The data product engineering team. Owns the full pipeline from Bronze to Gold — design, build, and governance of the 150+ table Coalesce environment. CI/CD, observability, and platform standards as the foundation for sustainable and scalable delivery. Data quality as an engineering discipline — systematic, measurable ownership, not reactive firefighting. Data contracts between layers — the formal handshake between pipeline and product.

New role — Data Architect: Strategic, not operational. Sets the design standard for the engineering squad — translating business requirements into data model into engineering task. Governs the Coalesce DAG. Owns data contracts between layers. Defines the technical architecture of the semantic layer — what entities, relationships, and metric definitions the semantic layer must expose, and the standards to which it is built. BI is responsible for expression and adoption; the Data Architect is responsible for what the layer structurally is. Transitions Engineering from a pipeline team to a data product engineering team. "I want the architect defining how we build data structures, not living in day-to-day validation." — Ben York
Product Management
Jennifer Rudesill · Kylie Humpert
Owns business relationship & demand intake
Determines change components & forms squads
Backlog management & prioritisation
Oversight & assurance of delivery
Data product definition & data contracts
Business Analyst (new role)
The orchestrating function. Owns the data product roadmap — translating business vision into a prioritised investment portfolio and sequencing delivery against strategic outcomes. The single front door for all data demand: business relationship management, structured intake, backlog ownership. Accountable for oversight and assurance of delivery — not doing the work, answerable that the right work gets done.

New role — Business Analyst: Sits in Jennifer's function. Faces the business, owns requirement translation, and creates lightweight prototypes that become the contract between business intent and technical delivery. Near-term: BA earns the business relationship alongside Derek. Medium-term: BA fully owns front-door intake. "We don't have business analysts operating between these three groups." — Ben York
Engineering & BI partner under Product
BI · AI · User Experience
Derek Webb · BI Analysts
Semantic layer expression & user adoption
Dashboard & report delivery
AI product consumption layer
CoE: BI · AI · User Experience
Delivery partner under Product
↔ Data Architect touchpoint
Defines the technical semantic layer — structure, entities, metric definitions, build standards. BI owns expression and user adoption on top of that foundation.
The intelligence and experience layer — and an evolving Centre of Excellence. Owns the expression of the semantic layer — how Gold data is surfaced, organised, and made consumable through Tableau, AI products, and embedded analytics. Owns dashboard and report delivery, business metric definitions as experienced by the user, and the adoption of data products across Ascend's brands. The measure of success: not just a working dashboard, but whether the business makes better decisions because of it.

Semantic layer responsibility split: The Data Architect defines the technical architecture — what the semantic layer structurally is, what entities and metrics it must expose, and the standards to which it is built. BI owns the expression and adoption — how those structures are implemented in Tableau, how they are presented to users, and whether users trust and use them.

The transition: Near-term, strategic partner in all requirement sessions alongside the BA — no longer the informal intake point for all data demand. Long-term, as Laudio and AI-embedded analytics mature, BI owns the full front-end experience: the CoE for Business Intelligence, AI consumption, and User Experience across Ascend's brands.
SOURCE SYSTEMSCRM · Apps · Brands
BRONZEAs-is ingestion
SILVERBrand → Domain
GOLDBusiness-ready
SEMANTICTableau + AI
BUSINESSDecisions & Insights

Value Chain — Product Usage

Product Usage is Ascend's most requested data product. It tracks how institutions, students and faculty engage with ATI, NHA, NASM and other learning platforms. It is the clearest example of what the operating model needs to fix.

How Product Usage change requests flow today

Reconstructed from three weeks of discovery. This is what produced a dashboard that took a year to deploy, only served ATI, and didn't connect to the data.

01
Business request arrives
Customer Success / Sales
Someone asks: "Can I see which students aren't using the platform?" No formal intake. Verbally raised — usually to Derek.
Output: Verbal ask, no written requirement
02
Derek absorbs the request
BI Team (Derek)
Derek becomes the default intake. He tries to scope it, but lacks business and engineering partnership on the source system knowledge to fully define it. He becomes accountable for things he can't control.
Output: Informal understanding, no PRD
03
BA function: nobody
GAP — role doesn't exist
No business analyst translates the requirement into data model components. Engineers go directly to data stewards, not business stakeholders. Wrong source of truth.
Output: A list of questions. No prototype.
04
Engineering builds in isolation
Engineering (Madhavi)
Source data tables found. Build starts. No data architect governs the model. Only ATI considered for scope.
Output: ATI-only data model. No contract.
05
BI builds dashboard separately
BI Team
Dashboard created with limited contact with Engineering. UAT occurs in production environment. Data quality issues surface. Business logic in Tableau that should be in Gold.
Output: Long dashboard development cycle. Analysts wrangling and validating data.
06
ATI delivered — other brands in backlog
Everyone / No-one
ATI-only dashboard in Tableau. Other brands (NHA) sit in backlog with no clear plan on how to incorporate. Architecture was never designed for scale.
Output: ATI only. Trusted but fragile for scale.
What went wrong
No single front door for requirements — demand defaulted to Derek informally
No BA to translate business question into data model components
Engineering built without understanding the consumption need
BI and Engineering worked in parallel with no handoff mechanism
No data contract — gold layer design was improvised, not governed
Cross-brand scope (NHA, NASM) never considered — ATI defaulted as the only brand
No quality gate — issues discovered by business users, not caught in pipeline
How a Product Usage change request flows in target state

PM holds accountability end-to-end — from intake to value realisation. Engineering and BI are delivery partners. Frank provides cross-brand authority at key decisions.

Frank Ingari — Executive Authority Layer
Cross-brand scope decisions · Governance enforcement · Joint investment accountability with CTO (Ash Siebecker)
↓ Accountable: Step 2 (scope) ↓ Consulted: Steps 4, 5 ↓ Accountable: Steps 6, 7 (value & ROI)
Product Management — Accountability Spine End-to-end ownership · Jennifer Rudesill · Kylie Humpert · BA (new)
Owns
Intake & demand management
Manages
Cross-brand scope process
Owns
BA translation + prototype sign-off
Accepts
Data contract before build starts
Oversees
Engineering delivery vs contract
Oversees
BI delivery vs contract
Gates + Loop ↺
UAT + value realisation → feeds back to intake
01
Structured intake
PM — BA (new role)
Business submits structured intake form. BA confirms scope, schedules discovery. Derek notified as BI partner.
Output: Intake record, scope confirmed
02
Cross-brand scope
Frank (A) · PM (R)
ATI-only or cross-brand? Frank decides. PM manages. NHA, NASM scoped in or explicitly out from day one.
Output: Scope decision, signed off
03
BA translates & prototypes
PM — BA (new role)
BA works with business to define the metric and decision. Creates lightweight Tableau wireframe. The prototype is the contract.
Output: PRD + prototype = success criterion
04
Squad formed, contract signed
PM (A) · Data Architect (R)
TPM forms the squad. Data Architect authors the contract — what each layer must deliver. PM accepts it. Build cannot start without this.
Output: Signed data contract. Sprint plan.
05
Engineering: Bronze → Gold
Engineering (A/R)
Engineers build to contract. Data Architect governs Gold model. DQ checks shift left. All brands in scope from day one.
Output: Gold layer, DQ-validated, all brands
06
BI: Gold → Semantic layer
BI Team (A/R)
Derek's team expresses the semantic layer in Tableau and AI products, built to the Data Architect's technical spec. Business logic stays in Gold. BI focused on expression quality and user adoption.
Output: Report ready for UAT
07
PM gates against prototype
PM (Accountable)
Does delivery match the prototype? PM leads UAT. No deviation without formal change request. Sign-off, then release. Value tracking begins.
Output: Approved. Agreed. Trusted.
Value realisation loop — Did the product deliver the promised outcome? PM tracks post-release value and feeds findings back into the next intake cycle. The loop is what makes the model self-improving.
What this changes
One front door — demand no longer defaults to Derek
Cross-brand scope decided at intake, not discovered at UAT
The prototype is the contract — no ambiguity about what "done" means
Engineering builds to a spec, not a list of questions
BI focuses on delivery quality — not intake, not blame absorption
DQ failures caught in pipeline — business users see trusted data
PM closes the loop — value realisation tracked and fed back into portfolio decisions

RACI Workshop — Live Session

Use the agreement dropdowns to record the room's position. A row turns green when all four functions agree.

EX
Frank Ingari
SVP Operations · COO Equivalent
Business champion, cross-brand authority, governance sponsor, Enterprise DQ owner
PM
Jennifer Rudesill
Director Product Mgmt · Kylie Humpert
Orchestrating function. Intake, squad, delivery oversight, assurance
BA role (new)
EN
Ben York · Madhavi Varansi
VP Engineering · Sr. Manager
Bronze→Gold pipeline, platform, data modeling, observability, data quality
Data Architect (new)
BI
Derek Webb
BI Director · BI Team
Semantic layer expression & user adoption, dashboard delivery, BI · AI · UX CoE
Key: A — Accountable R — Responsible A/R — Both C — Consulted C+ — Deep Consult I — Informed JC — Joint Council ★ Recommendation
Session Agreement Progress 0 / 30 activities agreed
Add New Activity
Activity EX
Frank		 PM
Jennifer		 EN
Ben/Madhavi		 BI
Derek		 EX agrees PM agrees EN agrees BI agrees Edit