Strategy · Product · Organizational Effectiveness

I build the systems behind business transformation

A systems builder with an Industrial-Organizational psychology foundation, bringing leaders and teams together to build great things.

About

I diagnose like a scientist and build like an engineer.

As a former EY management consultant trained in Industrial-Organizational psychology, I sit at the multi-functional crossroad of product and people. I build frameworks and models for strategy, but I'm still here when it comes time for implementation and execution. I work with teams to make long-lasting impact that sticks.

At Comcast, I led a team that consolidated three siloed groups into one center of excellence that handles over half a million service inquiries annually. Through aligned standards, processes and tools, we improved their on-time SLA from 43% at 48 hours to 92% within 2 hours.

I've also stood up an entire organization from scratch — designing the full employee lifecycle for a 329-person team that plans and executes the network upgrades reaching six million customers a year. From hiring and onboarding through training, performance, and the operating model that holds it together, I built how the team works so the change could scale and last.

Jonathan Shan
Experience
Comcast EY MTA New York City Transit
Education
New York University UC San Diego
Automation & internal tooling Org design & change management ML / predictive modeling Service & workflow design Identity & access Cross-functional leadership
Work Projects

Case studies in bringing systems into focus.

Eight years at Comcast building the tools, automation, and organizations that make a 28M-customer business run — framed the way a consultant would.

Case 01 · Service Journey COE

Three divisions, one center of excellence.

Situation

Three divisions (Northeast, Central, West) handled service-availability inquiries in isolation — different processes, different tools, no shared definitions. Roughly 550K inquiries a year, and only 43% met a 48-hour SLA.

Confounds

National scale across separate P&L groups, entrenched local ways of working, a legacy ticketing system, address-matching errors driving rework, and no single owner for the end-to-end journey.

Outcomes

Consolidated into one national COE with shared playbooks and journey maps, an evolved ticketing system, and better address matching — moving on-time SLA from 43% at 48 hours to 92% within 2 hours, and lifting revenue 17%.

Org designChange managementService designProcess standardizationCross-functional leadership
SLA time
2h
On-time SLA
92%
SLA windowtarget: 2 hours
Share of 550K inquiries meeting SLA506K on time
48h 2h

Compressing the SLA window from 48 hours to 2 didn't trade off attainment — it more than doubled it. Standardization made faster and better the same thing.

Case 02 · Project Genesis

A 329-person organization, from zero.

Situation

A major network transformation needed an execution organization that didn't exist — a team to plan and run upgrades reaching 6M customers a year.

Confounds

A net-new build with no existing lifecycle, roles, or training. The org had to be hired, onboarded, trained, and evaluated at speed — while the transformation it served was already in motion.

Outcomes

Stood up and scaled a 329-person organization with a full employee lifecycle — sourcing, onboarding, training, and performance scoring — operational and delivering against the transformation timeline.

Organizational buildWorkforce designTraining & enablementOperating modelI-O psychology
329
full-time team members,
hired and trained from scratch
6M customers reached annually
Case 03 · Predictive AI

Teaching the workflow to forecast itself.

Situation

Construction completion dates were unreliable — hurting forecasting, planning, and the customer experience. Teams needed a dependable signal for when work would actually finish.

Confounds

Building useful features from operational data; getting field teams to trust and adopt a model's recommendations; and tying predictions into an existing Salesforce-based workflow without disrupting how people already worked.

Outcomes

Shipped an "Estimated Construction Completion Date" random-forest model improving accuracy 18% over baseline — served live into the Salesforce workflow as recommendations field teams could act on.

Predictive modeling (ML)Feature engineeringDecision-support toolingSalesforce integration
+18%
completion-date accuracy
over baseline
random-forest recommendations
served live into Salesforce
Case 04 · Data Quality

Messy upstream data, reconciled at the source.

Situation

Node mismatches between the billing and telemetry systems that should have agreed led to correlation challenges, unnecessary truck rolls and forced teams into manual reconciliation and research of mismatches.

Confounds

Reconciling records across systems that never shared a clean key; designing automation resilient to messy, real-world inputs; and fixing the root cause rather than the symptom so the backlog wouldn't simply return.

Outcomes

Built workflow automation and data-quality checks that cut node mismatches from 520K to 45K — a ~91% reduction — retiring manual reconciliation and giving downstream teams data they could trust.

Data qualityWorkflow automationSystems integrationProcess design
−91%
node mismatches,
520K → 45K
manual reconciliation
retired at the source
Identity & access
2FA/MFA for 47M identities

Owned multi-year roadmaps for service availability, identity, and Flex (Xumo) — driving MFA adoption across 47M identities under government compliance.

Reporting alignment
One set of north-star KPIs

Aligned BI reporting across teams and divisions — shared definitions, consolidated dashboards, and feedback loops that gave executives a single source of truth.

Design operations
$2M saved, in-sourced

Consolidated three orgs and in-sourced production from external vendors, saving $2M a year and building a skills scorecard for 260+ employees.

Comcast tool · Locates correlation
What actually moves job outcomes

An internal decision-support tool that surfaces which factors most affect job outcomes — turning operational signal into guidance field teams can act on.

Personal Projects

Things I build on my own time.

Design and systems work outside the day job — proof that I don't just direct the build, I do it.

Design & front-end build

Queen Fashion — a luxury atelier, online.

A bilingual (IT/EN) marketing site I designed and built for a Prato-based women's ready-to-wear atelier: editorial heritage storytelling, a process walkthrough, a showroom gallery, and an appointment-request flow — all on a warm, restrained luxury system.

Open the live site ↗

queenfashionsrl.com

Live embed — if it doesn't load in a sandboxed preview, the site renders when this page is hosted, or open it directly above.

Live site · IT / EN
Queen Fashion
A bilingual luxury atelier site — heritage storytelling, showroom gallery, and an appointment-request flow. Best viewed full-screen.
Open the live site ↗
Rapid tooling · Plant 51

Shipping a useful tool in days, not months.

A building delivery & lookup guide for my own residential building — the practical, hard-to-find details residents and couriers actually need to get a package to the right place.

The point isn't the scope, it's the approach: solving an everyday problem without the $15k overhead — building and shipping it in days instead of months.

Open the live tool ↗

plant51buildinglookup.netlify.app

Live embed — if it doesn't load in a sandboxed preview, the tool renders when this page is hosted, or open it directly above.

Live tool · shipped in days
Plant 51 Building Lookup
A building delivery & lookup guide — an everyday problem solved without the $15k overhead. Best viewed full-screen.
Open the live tool ↗
Volunteer

The human end of everything I believe about people.

Systems are only ever in service of people. This is where that conviction shows up most directly.

Crisis Text Line

Volunteer Crisis Counselor

I volunteer with Crisis Text Line, supporting people through their hardest moments one conversation at a time. It's the most direct version of the listening, empathy, and grace-under-pressure that the rest of my work only gestures at — and a constant reminder of who the systems are for.

Good Karma Bikes

Volunteer · community bike shop

I volunteer with Good Karma Bikes, a San Jose nonprofit that turns donated bikes into affordable transportation and hands-on job training for youth and neighbors working toward something better. Tangible, mechanical, and local — a different register from the rest of my work, and one of the most grounding.

Personal project · wellbeing

Daily Journal — a tool that notices patterns.

Built in the same spirit: a small journaling app that pairs reflection with light analytics — a streak, a weekly count, a mood mix. The same instinct behind the engagement and pulse surveys I've run for thousands of employees, shrunk to a tool of one.

Daily Journal
How did today go?
My Journey

A decade, in steps.

Scroll across →

2012
B.S. Psychology & Neuroscience
UC San Diego
2014
Change Management / OD
NY Metropolitan Transportation Authority
Drove adoption of an Enterprise Asset Management system — people, process, and technology together.
2015
M.S. Industrial-Organizational Psychology
New York University
2015–18
Senior Consultant, People Advisory
Ernst & Young
Led change on digital transformations; built a 6,000-leader development program (97% favorable).
2018
Director, Strategic Programs
Comcast
Cross-functional transformation across Business, Product, Development, and Analytics.
2021
Senior Director, Next-Gen Access Network
Comcast
Project Genesis, the Service Journey COE, and predictive operations.
2025
Executive Director
Comcast
Leading strategic programs across automation, identity, and system architecture.
Selected impact
329
person organization
built from scratch
4392%
on-time SLA,
48 hrs → 2 hrs
550K
annual inquiries
through one COE
$2M
annual savings from
in-sourcing
Connect

Have a transformation that has to stick?

I partner with leaders to turn big, messy change into systems that last — strategy through execution, diagnosed like a scientist and built like an engineer. If that's the work in front of you, let's talk.

jonathancshan@gmail.com