Enterprise Workflow Automation

Discovery and Process Mapping

We spent the first two weeks on-site (one week at each office) interviewing team members across roles — from the founder and practice leads to junior consultants and the accounts coordinator. We weren't just mapping workflows on paper; we sat with the operations lead while she did her Friday reporting and documented every tab switch, copy-paste, and manual lookup in real time.

Two findings from discovery reshaped the project scope significantly:

• The team's real workflow differed from what leadership described.Leadership believed consultants updated the project management tool daily. In practice, most updated it once a week — if at all — and the real project status lived in WhatsApp group chats. Any platform we built had to make status updates faster than sending a WhatsApp message, or people simply wouldn't use it.
• The operations lead's Excel workbook was far more sophisticated than anyone realised.It didn't just track resource allocation — it calculated blended billing rates, flagged projects approaching budget limits, and generated a utilization heatmap leadership relied on for staffing decisions. We had to replicate this logic exactly, not approximately, in the new system.

Iterative Build

We structured the build into three releases over 12 weeks:

• Release 1 — Client engagement & project tracking (Weeks 1–4). The core module. We replaced the CRM and project management tool with a unified view where each client engagement had its own workspace — pipeline status, staffing assignments, task board, and document links. Consultants could update project status in under 30 seconds, faster than typing a WhatsApp message. We piloted with a small group and iterated on feedback before rolling out company-wide.
• Release 2 — Resource allocation & the profitability engine (Weeks 5–8).The hardest module. We reverse-engineered the operations lead's Excel formulas into a rules engine that calculated blended billing rates, tracked actual vs. budgeted hours, and flagged engagements trending toward margin erosion. She validated every calculation against her spreadsheet until the numbers matched exactly.
• Release 3 — Invoicing, dashboards & executive reporting (Weeks 9–12).We integrated with the firm's accounting software via API, built a one-click invoice generator tied to logged hours and contract terms, and created the dashboards that gave leadership their real-time view of revenue, utilization, and pipeline health.

Data Migration and Transition

Data migration was more complex than initially scoped — which is typical in projects like this. The CRM data exported cleanly, but the project management tool's export had inconsistencies: project names didn't match between systems, some closed engagements in one tool still showed as active in another, and historical task assignments were incomplete.

We built a migration pipeline in Python that pulled from the existing tools' APIs, cross-referenced records using client name and engagement date as matching keys, flagged conflicts for manual review, and loaded clean data into PostgreSQL. The operations lead reviewed the flagged records manually — a process that took a couple of days but ensured no historical data was lost or misattributed.

For the transition itself, we ran both systems in parallel for several weeks. The firm continued using their existing tools while the team got familiar with the new platform. Adoption was tracked daily and ramped quickly — within a few weeks, the majority of project status updates were happening in the new system, and the old tools were effectively abandoned. We didn't force anyone to switch — the platform was simply faster and more useful than the tools it replaced.