Product strategy · Growth · Operations · Applied AI

Building products that make complex operations feel simple.

I’m Bhaskar Ashoka, a Sydney based product leader with 14+ years across omnichannel retail, aviation, payments and B2B SaaS. I connect customer experience, operating models and commercial outcomes, then turn the strategy into a measurable launch.

Explore my workTry the AI prototypesDownload portfolio PDF
681 → 755Australian Direct to Boot sites, F22–F25
8 → 4 minApproximate median customer handover time
+12%Relative 90 day repeat in targeted member cohorts
45%Approximate Qantas front end latency reduction
Experience

From business case to rollout.

My work spans customer adoption, service operations, API enabled journeys, and the measurement needed to decide whether a product is ready to scale.

2021–2025 · Woolworths Group

Direct to Boot and personalisation

Built a growth roadmap around reach, trial, repeat and speed. Worked with Product, Engineering, CRM, Rewards, Finance and Store Operations to scale site readiness and improve the collection journey.

2025 · Qantas

Cancellation and refunds

Owned domestic cancellation and automated refund discovery, launch economics, exception design and API aligned requirements in Online Booking Core.

2020–2021 · Integrated Research

Payment monitoring SaaS

Consolidated telemetry and severity based alerts across 50+ enterprise accounts; resume reports mean time to detect improving from approximately 15.2 to 10 minutes.

2012–2020 · Fluke / J.P. Morgan

Connected assets and payment flows

Delivered asset health workflows and B2B payment mapping, validation and reconciliation requirements across enterprise platforms.

Selected case studies

What changed, and how I approached it.

01 · Omnichannel retail

Scaling Direct to Boot without losing service quality

ProblemPickup demand and site coverage grew, while inconsistent arrival signals and store readiness added friction at handover.
My roleOwned the growth roadmap and launch measures; aligned customer, store and platform teams on eligibility, readiness and exception handling.
DecisionsUsed catchment demand, bay readiness, capacity, training, arrival signalling and service gates to decide launch, hold or remediate. Paired consent based location arrival with a manual fallback.
ResultNetwork expanded from 681 sites in F22 to 755 in F25. Median customer handover time fell approximately from 8 to 4 minutes. The two trends are associated with the roadmap; this page does not attribute the full network growth to one intervention.
EconomicsFour minutes saved per eligible collection is a capacity measure. Annual labour savings require observed order volume, paid hours, staffing flexibility and Finance validation; I do not claim a realised payroll saving from handover time alone.
02 · Retention

Testing Rewards for repeat use

Hypothesis: Targeted bonus points and re engagement would grow repeat use more efficiently than blanket discounting.

Approach: Segment eligible members, create a test and control, track 90 day purchase alongside offer cost and fulfilment quality.

Result reported on resume: 31.0% to 34.7% repeat purchase in targeted cohorts: +3.7 percentage points, approximately +12% relative. This is a cohort result, not a companywide lift.

03 · Aviation

Native cancellation and refund journey

Problem: Hosted screen handoffs made customer journeys slower and exceptions harder to manage.

Approach: Built the commercial case; mapped voucher and card refund economics, API boundaries and 24 high risk scenarios; set performance budgets with architecture and engineering.

Outcome reported on resume: Approximate p95 front end latency reduced from 5.5 to 3.0 seconds, about 45%. The 35% eligible contact deflection figure was an estimated opportunity, not a measured result.

AI Lab · interactive portfolio prototypes

Two end to end decision workflows.

These browser based prototypes use synthetic examples and explicit rules. They demonstrate orchestration, evidence, guardrails, approval and measurement. They do not connect to employer systems, execute actions or call a live language model.

Retail Operations Copilot

Detect service risk → investigate → estimate impact → ask for approval → measure

Scenario: a pickup location shows rising handover time. Change the inputs to see how the proposed action changes.

Sample baseline target: 4 minutes. Minutes released are theoretical capacity, not verified labour savings. Pilot requires store owner approval and privacy review for location events.

Payments Migration Copilot

Map requirements → detect risk → propose migration → approve a test plan → monitor

Scenario: a merchant wants a safer path from legacy billing to modern payment options. Select the constraints; the prototype generates a staged recommendation.

Conceptual product design only. No account details or payment data are collected or sent anywhere. Actual migration needs scheme, security, privacy, legal and provider review.

Workflow architecture

  1. Collect structured inputs and check ranges.
  2. Retrieve permissioned evidence and cite its source.
  3. Use specialist analysis to propose options and quantify uncertainty.
  4. Apply deterministic policy and safety checks.
  5. Present an action for a named human owner.
  6. Execute only after approval; log decision and measure actual outcome.

Production design checklist

Start with read only integrations, scoped credentials, redaction, audit logs and evaluation cases. Add tool calling behind a server, require human approval for operational or payment actions, and monitor errors, drift, cost and value realised. The downloadable deck includes the rollout plan and success measures.

About

Product leadership with technical depth.

PhD in Computer Science, University of Otago (2012). I have led products and requirements across retail fulfilment, aviation servicing, payments monitoring, industrial IoT and commercial payment flows.

What I bring to a team

Customer discovery, commercial cases, launch readiness, experimentation, API and exception mapping, cross functional alignment, and operational measurement. I am especially interested in applied AI that gives teams a clear, reviewable decision and a way to verify its result.

Download the full presentation