I'm Medha Ashok, a digital product manager who works where strategy meets the machinery: shipping roadmaps, wiring AI and automation into the everyday, and making pages findable by people and by models alike.
Not a specialist bolted onto a team. Someone who moves between the roadmap, the analytics and the automation without losing the thread.
A short list on purpose. Each of these is a body of work, not a screenshot, so each gets its own case study.
Owning the product surface of a growing D2C brand, from roadmap and experimentation to the search strategy that decides who ever finds the store.
Read the case ↗Building the intelligence layer behind an events operation, turning scattered activity into automated reporting the team could actually decide with.
Read the case ↗Seven releases across the storefront, the app and the search layer: the problem, the evidence behind the call, what shipped, and what each launch taught me.
Rebuilt the shopping app on AppBrew, owning the user experience end to end: the pain points in the old app, the wireframes, the navigation, and the review rounds that got it shipped.
The existing app had collected friction. Navigation buried the journeys people actually came for, and the experience didn't match how parents shop: quickly, one-handed, often returning for the same few things.
I audited the old app and wrote down every pain point, then compared app-builder platforms before we committed. That gave us a clear list of what had to change and why, rather than a vague brief to rebuild.
I wireframed the new app: the navigation, the user flows and the way it should look and feel. The AppBrew team built it from that, and I reviewed, tested and audited each version, feeding changes back over several rounds before we released in stages.
The new app went live on both stores, with the journeys people use most moved to the surface and a phased rollout from 20% to 100% that kept any issue small.
After launch we found the app's search results didn't match the website's. It went straight to the top of the fix list, and the lesson stuck: when two surfaces share a catalogue, you still have to test them separately rather than assume they behave the same.
Rebuilt the page where the buying decision actually happens, wireframes to live, with an A/B analysis behind the call rather than an opinion.
The product detail page carried the most decision weight and the most drop-off. Trust and delivery signals sat too low, and the page didn't answer the questions parents actually ask before buying for a child.
Ran an A/B testing analysis across the product-to-checkout path and wrote it up properly, then read it against funnel drop-off between product view and cart to locate the real leak.
Wireframed the page, took it through design review, briefed the outsourced development team and staged the release. Scope was deliberately held to reordering the information hierarchy and surfacing trust and delivery first; personalisation was deferred.
The new page went live across the catalogue with the decision-critical information moved above the fold. Across the period this and the wider conversion work ran into, blended conversion rate rose 79% year on year.
The week after go-live was corrections, not celebration: a run of minor and major fixes caught under active monitoring. I now plan launch week as part of the launch, with a triage window and an owner booked in advance instead of improvised.
Designed and launched a full loyalty tier programme, then re-platformed it when the original tooling capped what the programme could become.
Repeat purchase is the whole game for a parenting brand, where customers move through predictable life stages. There was no structured reason to come back, and no mechanic rewarding the customers who already did.
Researched tier mechanics and points-expiry models across comparable programmes, then worked the earn and burn rules until they held up rather than copying a competitor's.
Wrote the strategy and tier logic, designed the member communications across WhatsApp, email and on-site banners, secured sign-off, and drove the vendor implementation through to launch.
The full tier system went live in January 2026, and the programme later moved onto a new loyalty platform with its journeys rebuilt, lifting the ceiling on what could be automated.
One configuration decision, points expiry, sat unresolved for months because it lived on a vendor's roadmap rather than ours. It is the clearest lesson I have in checking what a platform cannot do before designing around it. The migration was the fix, and the requirement now goes in the evaluation, not the backlog.
Replaced a sign-in that kept failing with OTP and social login, so customers could get in with a phone number or an existing account instead of another password.
Signing in asked for a password nobody remembered, and the existing login failed outright in specific, repeatable situations. Account friction sat exactly where intent was highest, and every failure was a session that never reached a cart.
Rather than treating it as one bug, I collected the precise failing cases and traced each across the authentication, WhatsApp messaging and telecom compliance layers, since an OTP only arrives if all three behave.
Reviewed the login apps on the market against those cases and recommended the change, then coordinated three vendors to ship phone-number OTP and social sign-in together, testing the edge cases before go-live rather than discovering them afterwards.
OTP and social login went live in February 2026, so customers sign in with a one-time code or an account they already have, and the sign-in step stopped losing people.
Sign-in was fixed, but customers could still be asked to log in a second time at checkout, a seam between two systems that each behaved correctly on their own. I now treat cross-vendor seams as their own test surface rather than assuming two working integrations make a working journey. That thread is still being pulled.
Ran the storefront rebuild end to end: a new theme, a redesigned collection page, and the navigation the whole catalogue hangs off.
The store had outgrown its theme. Pages were inconsistent, journeys like the size guide and FAQ were half finished, and a catalogue spanning maternity, newborn, kids and gifting had no coherent way in.
I audited the live site for UI, UX and functionality faults and kept one running issue log, researched how comparable stores structure collection pages, and had interactive mockups built so people could react to something real instead of a description.
Designed the collection and About Us page mockups, specified every change for the development agency, rebuilt the mega menu and its sub-collection structure, and created the browse routes customers actually use, Shop by Brand across 45 brands, shop by age and the gifting pages.
The new theme went live in December 2025 with a rebuilt navigation, working size guide, FAQ and About pages, cleaned-up 404s and redirects, and a faster site after CSS and JS minification and lazy loading.
The fix list kept growing after go-live, and for a while I was chasing changes across chat threads. The fixes only started closing reliably once every issue had one owner and one place to live. A single shared log, not a faster inbox.
Launched a baby registry so expectant parents could build and share gift lists, then made it something customers could navigate without help.
Baby showers and newborn gifting are a natural fit for this catalogue, but there was no way for a parent to build a list, or for friends and family to buy from one.
Researched how registry journeys work elsewhere and what expectant parents actually need from one: building it, sharing it, and knowing what has already been bought. I mapped the journey before choosing the tooling.
Configured and integrated the registry apps, built the registry landing and must-haves pages, optimised them so the registry could be found in search, fixed the flows that broke, and wrote step-by-step visual guides so customers could self-serve.
A new gifting journey went live end to end, with its own landing page, a curated must-haves page and customer-facing how-to guides.
The first version shipped with a registry owner page that rendered blank for some users. I found it by recording myself walking the live journey, not the staging one. Screen-recording the real flow is now how I check anything customer facing before I call it done.
Made the storefront readable by AI assistants before most retailers had heard of the idea, alongside the technical SEO foundation underneath it.
Discovery was starting to happen inside answer engines and AI assistants, where a store either exposes clean machine-readable context or is simply invisible. Meanwhile the site's own technical health was holding back classic search.
I worked out what each engine actually reads, sitemaps and robots for crawlers, plain-language context files for models, then audited what the store was exposing and where the technical gaps were.
Published llms.txt, robots.txt, XML sitemaps and a written store description, registered the store as an OpenAI merchant, submitted it to search engines beyond Google, and cleaned up schema, alt text, metadata, 404s and redirects. Later extended to Shopify's agents.md.
The catalogue became legible to AI assistants and answer engines as well as crawlers, and the site's technical health score moved from 65 to 85.
Being early means there is no benchmark. I could not prove the return the way I can for a conversion change, which made it a hard sell internally. I now pair it with weekly AI-visibility tracking, so the bet is measured rather than assumed.
Medha Ashok did not start in a product org. She started where a lot of the most useful product people quietly start: in search. At AdaptNXT, an SEO internship was less about keywords than about a first, unglamorous truth. A product nobody can find may as well not exist.
That lesson followed her. As a News Analyst at Cognizant she learned to read signal out of noise at volume; at Ekya Schools, running search as an SEO executive, she learned that ranking is a downstream effect of decisions made much earlier, in structure, in content, in intent. Marketing kept handing her problems that were really product problems in disguise.
At Showmaker the work turned technical and operational: AI, analytics, automation. Instead of reporting on what had happened, she started building the systems that reported themselves, freeing a team to spend its attention on decisions rather than spreadsheets. This was the turn, from describing outcomes to designing the machinery that produces them.
Today, as Digital Product Manager at The Mom Store, the threads pull together. Roadmap and experimentation sit beside conversion and checkout. An AI layer and automation quietly handle the repetitive middle. And search has widened into three fronts at once: classic SEO, answer engines (AEO) and the generative surfaces (GEO) that increasingly decide what a customer sees before they ever reach a page. Much of the work is leadership too: keeping a blend of in-house specialists and external partners, across content, SEO and development, moving on the same roadmap.
The throughline isn't a job title. It's a way of working: find where the business actually loses momentum, then build the smallest system that fixes it for good, with product thinking, data and increasingly with AI.
Eight overlapping practices. Most engagements pull from several at once, which is rather the point.
Roadmaps, prioritisation and delivery with cross functional teams, from discovery to ship.
Putting models and automated workflows to work on the repetitive middle of the business.
Being found by search engines, answer engines and generative engines alike.
Framing the problem, sizing the bet, and sequencing the work that follows.
Conversion, checkout and the customer journey: the mechanics of a store that grows.
AI assisted dashboards and measurement that make the next decision obvious.
Design, wireframing and front end build, from first sketch to shipped page.
Moving teams from manual and reactive to systematic and measured.
A reverse-chronological look at the roles that shaped how I work.
Owning the digital product surface of a growing D2C brand: roadmap and experimentation, SEO/AEO/GEO, conversion and checkout optimisation, AI-assisted automation, and analytics dashboards that keep the team pointed at the right work.
A freelance engagement building the intelligence and automation layer behind an events operation: AI systems, automated reporting and workflows that replaced manual reporting and sharpened daily decisions.
Ran search as a discipline across technical and content SEO, structure and intent, learning that rankings are the downstream effect of decisions made much earlier in the product and content.
Read signal out of high volume information streams, the analytical groundwork that later became a habit of measuring before deciding.
First professional exposure to search, and the founding lesson of the whole career: a product nobody can find may as well not exist.
Formal grounding in the commercial side of digital: marketing strategy, ecommerce, analytics and the business logic underneath the product decisions.
A working stack, not a wish list. AI runs through how I research, design, build and ship, every day.
Have a product, a store, or a search problem worth solving?