Replaced legacy Excel-based deal tracking with a React management portal and Python data pipelines. 34% productivity lift across 100+ sales reps. 50%+ fewer manual handoffs.
As part of the platform team at a Fortune 500 research firm, I designed and built the internal sales operating system that replaced a sprawling Excel-based workflow. Two surfaces, one platform: a React management portal the sales team lives in, and a set of Python REST APIs that keep client data, pipeline state, and downstream enterprise dashboards in sync.
100+ sales reps were tracking deals, clients, pipeline stage, and account history across fragmented Excel files. Every team had its own version. Every quarter someone had to stitch them back together for leadership reporting. By the time a number reached an exec, it was already a few days stale and a few dozen edits behind.
Below the surface, the real cost was integration debt. Client data lived in one system, deal data in another, and the unstructured stuff, like call notes and account context, lived in a rep's head or in a one-off spreadsheet column. Manual handoffs between systems weren't a process problem; they were the process. People were the integration layer.
So the brief had two parts: give the sales team an interface they actually want to use, and replace the human integration layer with code.
A React-based portal where sales reps run their day: real-time deal tracking, client management, pipeline stage updates, and dashboards scoped to their book of business. Role-based views for reps, managers, and ops, with the same data shown through a different lens.
Forms beat formulas. Updates write through to a real schema instead of a tab in a workbook. Audit trail by default. Reports that were quarterly stitching jobs became live dashboards.
I redesigned the Oracle SQL schema so client information became a first-class model instead of a denormalized blob spread across 14 tabs. Then I built REST APIs (originally Java Spring, later Python services) that downstream enterprise dashboards, reporting tools, and BI systems could subscribe to.
Structured data and unstructured context (call notes, account history) flowed through the same contracts. Manual handoffs dropped 50%+, not because anyone tried harder, but because the work stopped existing.
The portal became the place sales work happens. Pipeline reporting that used to be a quarterly stitching job became a live view. Client data stopped diverging across tools. Sales productivity went up 34% because the team stopped doing data entry twice and started spending the saved hours on actual deals.
If your sales, ops, or revenue team is running on a maze of spreadsheets, manual handoffs, and "wait, which file is the latest?", this is the kind of internal tooling I ship. Real schema, real APIs, real interface, real adoption.
And what used to take a rep 30 minutes of copy-paste between systems is exactly the kind of work I now automate end-to-end with AI agents. Same problem the Excel-replacement story solved a decade ago, modern stack: agents that read the inbox, update the CRM, draft the follow-up, and hand the rep a finished thing instead of another form to fill.