Case study
MedGrocer Technology & Product Process Audit
A one-month consultancy audit identifying improvements across MedGrocer's product ownership, preview deployments, QA, AI-assisted development, and release management.
My role
Responsibilities
Stack
Year
Timeline
The opportunity
Returning with an outside perspective
After working at MedGrocer from 2019 to 2023, I returned in 2025 for a one-month consultancy engagement focused on auditing the company’s technology and design processes.
As the organization and its digital platforms had grown, product decisions, development handovers, releases, and QA activities had become increasingly distributed across teams. The engagement provided an opportunity to examine how work moved from business requirements to design, development, testing, and release.
My objective was to identify the operational issues creating the most friction and recommend practical improvements that could be introduced without disrupting ongoing product delivery.
Want to see the thinking behind the outcome?
Jump ahead to the research and product decisions that shaped the final experience.
Solution & results
A practical roadmap for modernizing product delivery
The final deliverables included a written audit report, a presentation to leadership, and a prioritized implementation roadmap built around four areas:
- Create dedicated product ownership between business units and the technology team.
- Introduce automated, isolated preview environments for earlier QA and stakeholder review.
- Expand and standardize the use of AI-assisted development and code review tools.
- Use release trains, feature flags, and selective deployments to reduce release risk.
The roadmap connected each recommendation to the workflow issue it addressed, its expected organizational value, and the operational changes required for adoption.
A later independent audit evaluated the same four areas and provided a useful snapshot of the organization’s subsequent progress:
- AI-assisted development received the strongest implementation score at 8.1/10. Cursor and CodeRabbit had become expected parts of code-quality workflows, although planning and prompt practices still required standardization.
- Dev-to-QA handover and preview workflows scored 6.75/10. Pre-production previews and some automated triggers were present, but isolated environments during active development remained limited.
- Release trains and feature flags scored 5.75/10. Several platforms had adopted controlled release practices, while broader development still relied on bulk deployments.
- Product management alignment scored 4.25/10. Some integration had taken place, but development leads continued to absorb product responsibilities and a dedicated product manager had not yet been established.
Because this assessment was conducted independently, I present it as validation of the importance and continued relevance of the four areas identified during my engagement, rather than direct attribution to my recommendations alone.
- 4
- Focus group discussions
- 4
- Strategic recommendations
- 3
- Final leadership deliverables
- 1 month
- Consultancy engagement
Discovery
Four discussions, four recurring delivery gaps
I conducted four focus group discussions for each of their major ongoing projects to understand how team members experienced product planning, development, quality assurance, deployment, and release management.
Rather than evaluating individual performance, I looked for recurring issues across workflows. I compared how participants described ownership, handovers, testing constraints, release practices, and their use of emerging development tools.
The discussions revealed that several challenges were connected. Unclear product ownership affected task quality and turnover, deployment limitations slowed QA, large releases increased operational risk, and AI adoption varied significantly between team members.
Methods
- Focus group discussions
- Workflow analysis
- Process mapping
- Thematic analysis
- Recommendation prioritization
Product ownership was distributed across technical roles
Product coordination and business correspondence were frequently absorbed by development leads alongside their existing responsibilities. Without dedicated product ownership, requirements and user stories could reach the technology team without consistent cross-functional alignment.
QA depended heavily on shared environments
The team lacked a consistent way to review isolated changes while development was still in progress. This pushed more functional and integration testing into shared staging environments, increasing coordination overhead and extending feedback cycles.
Releases were larger and more tightly coupled than necessary
Much of the development workflow relied on bulk deployments. Without broader use of feature flags and selective releases, teams had fewer options for separating completed work, controlling exposure, or reducing the risk associated with larger deployments.
AI adoption was promising but inconsistent
Some developers had begun using tools such as Cursor and CodeRabbit, but adoption was limited and often depended on individual initiative. The opportunity was not simply to introduce AI tools, but to develop shared practices for using them safely and consistently.
The audit showed that the team did not need one isolated process fix. It needed a more cohesive product delivery model connecting business requirements, technical planning, development, QA, and controlled releases.
I consolidated the findings into four recommendations selected for their potential to reduce coordination overhead, shorten feedback cycles, improve release safety, and support the team’s ability to scale.
Designing the experience
Prioritizing changes with organization-wide leverage
Because the engagement was limited to one month, the goal was not to redesign every internal process. I prioritized four changes that addressed the most frequently observed sources of friction and could create benefits across multiple stages of product delivery.
Establish dedicated product management
I recommended hiring a dedicated product manager to own correspondence between business units and the technology team.
This role would translate business requirements into a unified product direction, improve task readiness before development, coordinate priorities across stakeholders, and allow development leads to focus more consistently on technical leadership.
Introduce automated preview deployments
I recommended using Vercel preview deployments for eligible web applications so that each proposed change could be reviewed in an isolated environment before release.
Giving developers, designers, product stakeholders, and QA personnel access to the same preview would allow feedback to begin earlier and reduce dependence on a centralized staging environment.
Standardize AI-assisted development
I advocated for wider adoption of tools such as Cursor and CodeRabbit across the development team.
The recommendation included treating AI as a shared engineering capability rather than an individual preference, with common expectations around code review, validation, security, and prompt practices.
Adopt release trains and feature flags
I recommended introducing predictable release trains alongside feature flags and selective deployments.
This approach would allow completed features to be merged and deployed without immediately exposing them to every user. It would also reduce reliance on large bundled releases and provide greater control over testing, rollout, and rollback.