My Ownership
Product direction through release.
As project owner, I owned problem definition, roadmap, backlog, product architecture, UX/UI, build specifications, release planning, and hands-on QA.
A B2B SaaS platform that gives impact investors and the companies they back one place to track, verify, and report on impact and financial metrics. AI structures existing portfolio data before both sides have joined.
AI Summary
The Brief
Impact investors and the companies they back were managing everything across disconnected tools: investor updates in email threads, metrics in spreadsheets, reports in shared drives with no common format or version control. The core problem was a cold-start gap: an investor tracking a portfolio company that isn't yet on the platform had no structured way to capture historical data. As project owner, designer, and AI developer, I owned the backlog, roadmap, product architecture, UX and interface design, delivery, and hands-on QA. I directed AI coding agents through structured specifications, then reviewed every change in the live product before it shipped. Invized is now a responsive SaaS platform used by investors to track and verify impact metrics across their portfolios.
My Ownership
As project owner, I owned problem definition, roadmap, backlog, product architecture, UX/UI, build specifications, release planning, and hands-on QA.
Collaboration
I translated Resonary stakeholder feedback into milestones and acceptance criteria, directed AI coding agents through scoped builds, and reviewed every production release.
Core Constraints
The product had to support investor-first adoption, inconsistent source files, permission-sensitive portfolio data, and continuous releases without breaking shared workflows.
Invized impact and financial metrics dashboard for investors and portfolio companies
Context
Impact investors need consistent, verifiable data across their portfolio companies to report to limited partners and make capital allocation decisions. Portfolio companies report in different formats, leaving investors to reconcile the data themselves. An investor tracking a company that hasn't joined the platform faces the same problem: historical data is buried in PDFs, quarterly updates, and email attachments with no structured home.
Portfolio companies have no clean way to share structured metrics with investors, manage their cap table, or control access to sensitive documents. They're maintaining relationships across email and spreadsheets, and every reporting cycle is a manual rebuild from scratch. The problem affected both sides of the relationship, so the product needed to work for both.
Product Design
Invized solves that gap with an AI document-extraction flow: an investor uploads a portfolio company's existing reports and investor updates, the AI extracts each metric with a reference back to its source, the investor reviews and confirms each value, and confirmed data folds into their portfolio. No manual re-entry. No guessing what the numbers mean.
Portfolio companies use a dedicated workspace to report metrics, manage their cap table, and share reports and documents with investors through a controlled access model. The provenance model distinguishes company-reported data from investor-entered data across every metric, so the source of every number is always traceable. Invitations, notifications, a rich-text report editor, and AI usage controls support the wider workflow.
Step 01
I mapped the investor and portfolio-company journeys, including the cold-start case where only one side is on the platform. That work clarified the data model, provenance requirements, and highest-risk workflows. I translated the findings into a prioritized backlog and milestone roadmap covering metrics, reports, invitations, and AI document extraction.
Step 02
I designed the information architecture and end-to-end flows for onboarding, metric collection, AI-assisted review, reporting, and collaboration. A reusable system for tables, modals, badges, empty states, and status cues kept patterns consistent as the product grew. Edge cases and permissions were resolved at the flow level before being translated into build specifications.
Step 03
I directed AI coding agents through scoped specifications and reviewed each increment in the live product. Product use and QA exposed both interaction friction and performance issues. One investigation traced slow pages to cross-region infrastructure rather than the interface. Fixes, new insights, and stakeholder feedback flowed back into the backlog for the next release milestone.
Step 04
Every release went through hands-on QA against acceptance criteria, responsive states, permissions, data edge cases, and shared-component regressions. I batched work in Monday.com, verified each deployment in production, and used product behavior and feedback to prioritize follow-up improvements. The product continues to evolve through the same discovery, design, delivery, and learning loop.
Research Synthesis
Journey mapping showed that investors needed value before every portfolio company joined. That finding made investor-created company records and historical imports part of the core experience.
Metrics could come from companies, investors, or uploaded documents. The interface needed to preserve the source of every number instead of presenting all data as equally verified.
The workflow analysis connected metric tracking, document storage, and reporting into one system so each reporting cycle could reuse structured data instead of starting from another spreadsheet.
Decision-Making
Requiring both parties to join would create a dead end. I prioritized investor-created records that can later connect to a company workspace, accepting more permission complexity in exchange for immediate value.
Fully automatic imports would be faster but harder to trust. The selected flow pairs every extracted value with its source and requires review before it becomes portfolio data.
Reusable tables, statuses, permissions, and empty states were defined before feature expansion so new workflows could ship without creating a different interaction model each time.
Process Evidence
These slots are reserved for the research, flow, and iteration screenshots that will document how the product changed.
MVP milestones, incoming bugs, and resolved issues.
Investor-first adoption through connected collaboration.
Pain points, goals, behaviours, and system constraints.
Invized is a live, responsive SaaS platform with a consistent component
system, source-backed AI document extraction, structured metric tracking,
collaborative reporting, and clear data provenance.
Beta access has reached more than 2,000 users, and over 200 reports have
been created. The new workflows reduced average data-entry time by 34%.
Structured specifications, written QA checklists, and batched releases
helped me move quickly with AI coding agents while maintaining a clear,
dependable product experience.
