Skip to content

Enterprise AI products,from first prototype to real-world adoption.

Part product strategist, part hands-on builder. I prototype AI ideas myself to decide which ones earn a roadmap slot, and I lead the product, engineering, and QA team that takes them live across FOX Sports, FOX News, and FOX Entertainment.

I came up through business analysis and solution architecture before running product, so my judgment starts in the workflow: how the work really happens, what data and integrations it touches, and whether a prototype earns a place on the roadmap. I still do that hands-on work myself, and I pair it with leading a 14-person product, engineering, and quality organization.

Featured workDownload resume
Location
Phoenix, AZ
Current role
Director, Product · Fox Corporation
Focus
AI products · Platforms · Adoption
Building for
FOX Sports · FOX News · FOX Entertainment

The scale I work at

Teams, talent pools, and stakeholder groups large enough that product decisions show up in daily operations.

  • 14people

    Product, engineering, and QA org

    A distributed team I lead across time zones, covering enterprise applications and AI Solutions Engineering for three FOX businesses.

  • 9,000+freelancers

    Talent pool behind AI crew matching

    The scale of freelance crew my AI recommendation concept ranks by location, sport experience, past collaborations, and preferred-hire signals.

  • ~200stakeholders

    Operators I built for

    FOX Sports and Entertainment users whose live-production and service needs I turned into a prioritized platform roadmap.

  • >50%efficiency gain

    Workflow redesign at Cisco

    Reported operational efficiency improvement from aligning requirements and redesigning Salesforce workflows.

How I decide what to build

Most of product is choosing what not to do. These are the rules I hold myself to.

  • 01

    Start from the workflow, not the model

    I look for the expensive, repeated decision inside someone's day before I pick a technique.

    In practice: AI Car Shares targets rental car spend on productions, and grouping freelancers by arrival time is how it gets there.

  • 02

    Prototype to kill ideas cheaply

    I test AI concepts hands-on for operational value, technical feasibility, enterprise controls, and adoption before they get roadmap time.

    In practice: AI Crew Recommendations was designed and demonstrated as a working concept.

  • 03

    Use signals people can explain

    A recommendation gets adopted when the people using it can see why it was made.

    In practice: Crew suggestions rank on location, sport-specific experience, past collaborations, and preferred-hire status, each a reason a person would give for a pick.

  • 04

    Design adoption in from day one

    Launch is the start of the product, so the feedback loop ships with it.

    In practice: FOX News Field Intelligence went live with issues flowing from Slack to AI-assisted support to Jira.

  • 05

    Build platforms, retire one-offs

    Every new feature should simplify the system around it, not add another island.

    In practice: Playbook Event Hub consolidated event information as part of a plan to reduce fragmented processes and platform-license reliance.

  1. 01

    Start from the workflow, not the model

    I look for the expensive, repeated decision inside someone's day before I pick a technique.

    In practice: AI Car Shares targets rental car spend on productions, and grouping freelancers by arrival time is how it gets there.

  2. 02

    Prototype to kill ideas cheaply

    I test AI concepts hands-on for operational value, technical feasibility, enterprise controls, and adoption before they get roadmap time.

    In practice: AI Crew Recommendations was designed and demonstrated as a working concept.

  3. 03

    Use signals people can explain

    A recommendation gets adopted when the people using it can see why it was made.

    In practice: Crew suggestions rank on location, sport-specific experience, past collaborations, and preferred-hire status, each a reason a person would give for a pick.

  4. 04

    Design adoption in from day one

    Launch is the start of the product, so the feedback loop ships with it.

    In practice: FOX News Field Intelligence went live with issues flowing from Slack to AI-assisted support to Jira.

  5. 05

    Build platforms, retire one-offs

    Every new feature should simplify the system around it, not add another island.

    In practice: Playbook Event Hub consolidated event information as part of a plan to reduce fragmented processes and platform-license reliance.

Featured work

The problem, the calls I made, and what changed. AI products, platforms, and launches across Fox and Cisco.

AI search · Crew logistics

Featured build

CrewFlow AI

Hands-on build · Media & Entertainment, live sports

AI search is the product

Example: “Who can cover Jets at Chargers on Sunday?”

  • Ask in plain English, not filters

    A first-time scheduler shouldn't have to learn the crewing system's data model.

  • The AI explains its pick

    People trust a recommendation when they can see the reasons behind it.

  • A person always decides

    The AI recommends, a coordinator approves, and every decision is logged.

  • Plugs in, doesn't replace

    It sits read-only on top of the crewing system a team already uses, so nobody has to migrate.

Crew logistics for live sports where the week is the home screen: one US map with every event and every crew member, and an AI search bar that answers staffing questions in plain English.

Why it mattered

Crews get scheduled one event at a time, so nobody sees the whole week. Putting the week on one map, and letting people ask for what they need instead of filtering for it, turns staffing from a hunt into a decision.

Problem
Staffing tools are built around one event at a time. Coordinators jump between screens to work out who is nearby, who is free, and who fits, and the tradeoffs across a full week of events stay out of view.
Key decisions
  • Made search the front door. The model turns a question into validated filters and the matching engine does the deciding, so every answer is grounded in real availability.
  • Kept a person in charge. The AI recommends and explains, a coordinator approves in two clicks, and every decision is logged.
  • Designed for trust over cleverness. The AI can't invent dollar figures, because every answer is checked against the matching engine, and if the AI is down the app falls back to basic matching.
  • Plugs in on top of the crewing system a team already uses, through a read-only adapter layer, instead of asking anyone to migrate.

Under the hood

  • Open-weight LLM (gpt-oss-120b on Groq) turns questions into validated filters
  • Every answer checked against the matching engine
  • Basic matching fallback when the AI is unavailable
  • Read-only adapter layer that plugs in on top of your crewing system
  • Python and SQLite backend, strict TypeScript and MapLibre front end
  • 141 automated tests, built on open-source tools, Docker on a free host

Outcome

A working product anyone can try, with demo data from the public 2026 NFL schedule. The real exercise was product judgment: where AI earns its place, and where a person should stay in charge.

AI searchLLMsHuman in the loopLive sportsMaps
Try the live demoSee the LinkedIn post

Runs on a free host, so the first load can take about a minute to wake up.

  • AI · Recommendations

    AI Crew Recommendations

    Fox Corporation · Senior Product Manager

    9,000+

    freelancers in the pool

    Ranked crew suggestions for live productions, drawn from a pool of more than 9,000 freelance professionals.

    Why it mattered

    With 9,000+ freelancers, good staffing can't rely on memory. Ranking crew on the signals behind a good pick makes strong staffing decisions repeatable.

    Problem
    Staffing a live event means choosing crew from 9,000+ freelancers, and the right pick depends on location, sport-specific experience, and who has worked well together before. That context is hard to hold in anyone's head.
    Key decisions
    • Chose four signals operators already trust: location, sport-specific experience, previous production-team collaborations, and preferred-hire status.
    • Built and demonstrated the recommendations in context of the staffing workflow, so the value was visible to the people making the calls.
    Outcome
    A demonstrated recommendation approach that turns crew selection from recall into a ranked shortlist, built to improve staffing decisions at scale.
    RecommendationsLLMsPrototypingLive production

  • AI · Cost optimization

    AI Car Shares for Freelancers

    Fox Corporation · Recent

    An AI tool that groups production freelancers into shared rental cars based on when they arrive.

    Why it mattered

    Rental cars are a recurring cost on every production. Grouping freelancers who arrive around the same time turns that cost into a simple matching problem.

    Problem
    Freelancers travel to productions separately, and when each one picks up a rental car, transportation spend and coordination work pile up for every event.
    Key decisions
    • Matched freelancers into car shares by travel arrival time, the signal that decides whether sharing a car is practical.
    • Generated car-share recommendations automatically, replacing manual back-and-forth coordination.
    • Positioned it within a broader push to automate production workflows and speed up operational decisions.
    Outcome
    Fewer rental cars per production and lower rental costs, with less time spent coordinating travel.
    AI matchingWorkflow automationCost reduction

  • Social monitoring

    Stormwatch

    Fox Corporation · Distribution Marketing

    A social monitoring app that helps Fox's Distribution Marketing team follow the social conversation that matters to their work.

    Why it mattered

    Distribution Marketing needs a fast, reliable read on social conversation. A monitor built around the team's own questions is more useful than one that tracks everything.

    Problem
    Social conversation moves fast and spreads across platforms. Distribution Marketing needed a dependable way to see relevant signals without tracking them by hand.
    Key decisions
    • Framed the product around what Distribution Marketing needs to see and act on.
    • Shaped the app around the team's monitoring workflow.
    Outcome
    One place for Distribution Marketing to keep watch on social conversation and signals.
    Social listeningMarketingInternal product

  • Launch · Operations

    FOX News Field Intelligence

    Fox Corporation · Director, Product

    Product development and go-live for a FOX News field operations product, with a feedback loop built in from launch day.

    Why it mattered

    A new field product earns trust in its first weeks, so how feedback gets handled matters as much as the launch itself.

    Problem
    An operations product earns trust in its first weeks. Issues and requests from field users had to land somewhere visible and get resolved quickly.
    Key decisions
    • Owned product development through go-live operations rather than handing off at launch.
    • Designed one intake path for issues and enhancement requests: Slack, then AI-assisted support, then Jira tracking.
    Outcome
    A launched product with a repeatable loop from user feedback to tracked work, so it keeps improving after go-live.
    Go-liveAI-assisted supportFeedback loops

  • Platform · Live production

    Playbook

    Fox Corporation · Senior Product Manager

    Thousands

    live events and studio shows supported

    The Salesforce-powered system FOX uses to plan, staff, and run thousands of live sports events and studio shows.

    Why it mattered

    Playbook is where live production gets planned, staffed, and run, so every improvement reaches thousands of events and shows. It's also where AI features meet the people who will use them.

    Problem
    Live broadcast runs on hard deadlines. Planning, staffing, and operations tools have to be dependable for every event, every week.
    Key decisions
    • Owned the strategy and roadmap for the full live-production ecosystem.
    • Ran two-week Agile delivery across engineering, analysis, QA, and vendor teams, with release quality as a priority.
    • Brought AI concepts, including crew recommendations and car-share matching, into the live-production roadmap.
    Outcome
    A platform that supports thousands of events and studio shows and gives AI features a place to reach real operators.
    Product strategySalesforceAgile delivery

  • Modernization

    Playbook Event Hub

    Fox Corporation · Senior Product Manager

    One operational view of event information, and the first step in a plan to simplify the platform and reduce license reliance.

    Why it mattered

    Consolidating event information helps operators today and moves the platform toward fewer fragmented processes and less license reliance.

    Problem
    Event information lived in fragmented processes, and the platform carried growing license dependence.
    Key decisions
    • Consolidated event information into a single operational view.
    • Tied the delivery to a broader modernization plan: decouple custom workflows, simplify architecture, and reduce long-term licensing dependencies.
    Outcome
    A consolidated view for operators and a clear path toward a simpler, less license-bound platform.
    Platform modernizationArchitectureOperations

  • Cross-system workflow

    FIFA World Cup 2026 Workforce Tracking

    Fox Corporation · Senior Product Manager

    Workforce-tracking capabilities for FIFA World Cup 2026, built with HR and operations teams.

    Why it mattered

    At World Cup scale, workforce data has to move cleanly between Playbook, Smartsheet, and Workday, the systems HR and operations work in.

    Problem
    A tournament at World Cup scale needs workforce data to flow cleanly across the systems different teams already rely on.
    Key decisions
    • Partnered with HR and operational stakeholders to define requirements.
    • Designed workflows that span Playbook, Smartsheet, and Workday instead of forcing a single tool.
    Outcome
    Requirements and cross-system workflows that support workforce tracking for one of the biggest events on the calendar.
    IntegrationsHR systemsRequirements

  • Process redesign

    Cisco Workflow Optimization

    Cisco Systems (Amensys Inc.) · Salesforce Business Analyst

    >50%

    efficiency improvement

    Salesforce workflow and data operations redesign that reported more than 50% efficiency improvement.

    Why it mattered

    Aligning business and technical teams on requirements, then redesigning the workflows, is what drove a reported efficiency gain of more than 50%.

    Problem
    Business and technical teams were working from misaligned requirements, and the workflows around them were inefficient.
    Key decisions
    • Led stakeholder discovery and solution design across administration, workflows, dashboards, and automation.
    • Improved data quality with Data Loader, DemandTools, SQL, and reporting.
    Outcome
    Reported operational efficiency improvements of more than 50%.
    DiscoverySalesforceSQL

Career timeline

From mapping workflows to running product and AI for three FOX businesses. Each step added a layer I still use every day.

Director, Product

Fox Corporation · Tempe, AZ

Jul 2026 — Present

Running product and AI solutions for three FOX businesses.

  • Set product strategy and investment priorities for enterprise applications and AI Solutions Engineering across FOX Sports, FOX News, and FOX Entertainment.
  • Lead a distributed org of 14 across product, engineering, and quality, from discovery through delivery.
  • Took FOX News Field Intelligence through development and go-live, with a Slack to AI-assisted support to Jira feedback loop.
  • Prototype AI concepts myself to test value, feasibility, controls, and adoption before they reach the roadmap.
  • Shape the modernization strategy for Salesforce-based apps: decouple custom workflows, simplify architecture, cut license dependence.

Capabilities

Product judgment and technical depth, in the same person.

AI product strategy

Deciding which AI is worth building, and making it land.

  • Generative AI and LLM products
  • AI agents and recommendations
  • Workflow automation
  • Value, feasibility, and controls assessment
  • Adoption metrics

Hands-on building

I prototype the idea before I ask a team to build it.

  • GPT, Claude, Gemini
  • Cursor, Replit, Figma Make
  • Rapid prototyping
  • SQL and Databricks
  • Solution architecture

Platforms and integration

Knowing where a workflow should live and what to retire.

  • Salesforce Service Cloud, Field Service, Data Cloud
  • Enterprise integrations and APIs
  • Platform modernization
  • Data migration and quality
  • Tableau reporting

Leading product orgs

Setting direction and shipping with distributed teams.

  • Strategy, roadmaps, portfolio prioritization
  • Business cases and ROI
  • Product discovery
  • Distributed and vendor team leadership
  • Agile delivery, UAT, release governance
  • Executive stakeholder alignment

Education & certifications

  • Master of Science, Information Systems & Management

    Marist University, New York · 2016

  • Bachelor's Degree, Computer Science

    Jawaharlal Technological University, India · 2013

  • Salesforce Certified Administrator
  • Salesforce Cloud Consultant
  • Salesforce Field Service Consultant
  • Salesforce Platform App Builder
  • Salesforce AI Associate
  • Harvard Managing Others Certificate

Book a call

Grab 30 minutes for a virtual coffee. Happy to chat about AI, products, or whatever you're building.

Contact

Want to talk AI products, platform strategy, or a problem you're working through? I'm always glad to compare notes. Based in Phoenix, AZ.

Headshot of Ashok Volati
Use email app