DIP
NetBramha Studios · Internal Platform
Product Portfolio · Sarath MS · 2026
Delivery
Intelligence
Platform
A role-aware operations system designed to reduce coordination overhead and give every person in the team the right view of work — without needing to ask anyone for an update.
Problem
Six spreadsheets, three Slack threads, no single view of what was happening
Solution
Role-aware platform where each persona sees exactly what they need to act
Impact
Weekly reporting overhead dropped, KRA reviews moved off email, timesheets consolidated
Scale
40+ active users · 7 role personas · 6 modules in production
Product goal — the north star
"Give every role a clear line of sight — so decisions happen at the right level, not after a meeting to find the information."
Before DIP, visibility required asking. A PM wanted resource allocation? They pinged a PM lead. Leadership wanted project health? Someone compiled a slide deck. Designers weren't sure about their KRA progress until the quarterly catch-up. The goal was to make the right information available to the right person without them needing to request it. Not just data access — the kind of view that actually changes how you make decisions each week.
Before and after — what actually changed
Before DIP
×Resource allocation lived in a shared spreadsheet with no conflict detection — double-bookings discovered in standups
×KRA reviews ran over email threads. No consistency, no comparative history, no way to see if last quarter improved
×Three separate timesheet systems, one of which was still a Google Sheet from 2021
×Delivery status required a PM to compile a slide deck before every leadership review
×Leave requests went to an HR inbox and were tracked manually in a Notion table
After DIP
✓Gantt-based allocation with live conflict detection. PMs see availability before assigning. Double-bookings caught before they happen.
✓Structured KRA reviews with AI coaching (Mira) and blind peer feedback. Historical ratings queryable. Reviews happen on schedule.
✓One timesheet interface aggregating all three legacy collections. Two-click logging. Compliance visible to PMs in real time.
✓Leadership dashboard with RAG project health, revenue-at-risk, and hygiene scores. No slides needed.
✓Self-serve leave requests with delegated approval rights per role. HR no longer manually maintains a tracking table.
Outcomes — what the system changed
🕐
Coordination overhead dropped
The recurring "where are we on X?" messages in Slack nearly disappeared for projects tracked in DIP. PMs stopped compiling status updates — they shared a link instead.
Before: Weekly slide deck prep before every leadership review
☑
KRA reviews actually happened
Review completion went from ad-hoc (some quarters skipped entirely) to structured and on schedule. The structured format also made manager vs. self-perception gaps visible for the first time.
Before: Email threads, inconsistent timing, no history
📋
Timesheet consolidation
Three separate systems collapsed into one. Designers log in two clicks. PMs can see team compliance without chasing individuals. The 2021 Google Sheet was finally retired.
Before: Three systems, one of which was a 2021 spreadsheet
Who uses it — and what decisions it supports
Leadership
CEO / COO
Portfolio health, RAG status, revenue-at-risk, team utilisation
Which projects need intervention before they slip
Stopped asking PMs for weekly slide decks
Project Manager
PM
Resource allocation, delivery hygiene scores, KRA compliance
When to flag risk; how to allocate without over-committing a designer
Allocation conflicts surfaced before assignments are confirmed
Designer
UXD / VD
Own KRA progress, timesheet, leave balance
What to prioritise this week; where they stand against their KRA
KRA progress visible continuously, not just at review time
HR
People
Leave calendar, team headcount, KRA completion rates
Capacity planning; identifying review gaps before they escalate
Retired the manual Notion tracking table entirely
Program Manager
Delivery
Cross-project portfolio view, revenue-at-risk, team health
Which engagements need attention; where to place senior resource
Single view replaces three separate spreadsheet reports
AI layer — Mira, the coaching engine
Mira — Claude Sonnet as performance coach
The problem with KRA reviews isn't data collection. It's that managers say different things, in different formats, with different levels of specificity — and the person receiving the feedback can't compare it to anything. Mira solves a consistency problem, not a content problem.
✦Reduces subjectivity. Ratings feed into a structured coaching prompt. The response is calibrated to the score pattern, not the manager's writing style. Two people with the same score get comparable guidance.
✦Surfaces the gap. Mira compares self-rating vs. manager rating per dimension and explicitly names where perception diverges — the conversation that most reviews avoid.
✦Cached per FY. Coaching output is stored against the review. Historical comparison is possible. "Did Q3 feedback land?" is now a question the system can help answer.
✦Not a replacement for the manager. Mira generates context, not verdicts. The manager still owns the review conversation — they just walk in better prepared.
KRA dimensions tracked
28 monthly KRAs across 4 behaviour buckets — Craft, Accountability, Curiosity, Value
Feedback model
Self + Manager + Blind Peer — three independent signals per cycle
Positioning
Decision support. Not performance management automation.
Product decisions — what we chose and why
Decision
Flexible KRA uploads over a rigid schema
Why
KRA formats vary significantly by role. Enforcing one schema would have required 6 role-specific templates on day one — blocking launch.
Trade-off
Some inconsistency in rating data structure across early cycles. Cross-role comparison required more normalisation.
Decision
Role-based views with strict data separation
Why
In a 40-person studio, visibility is political. If designers could see each other's KRA scores, trust in the system collapses. Access needs to match org norms, not just user stories.
Trade-off
More complex data model and auth logic. Seven role types, each with separate view logic. Higher build cost for a cleaner product.
Decision
Aggregate legacy timesheets rather than migrate
Why
Migration requires trust in the new system before it exists. Aggregation lets the team use DIP on day one while historical data remains accessible.
Trade-off
Three-collection data model adds read complexity. A full migration was deferred to V3 — by which point the team had enough confidence to move.
Scope decisions — what we chose not to build
Scope discipline is as important as what you ship. These were deliberate exclusions — each with a reason, not a deferrral.
No real-time collaboration
Concurrent editing on allocation or KRA data creates conflict resolution complexity. The team didn't need Google Docs — they needed a single source of truth. Those are different problems.
No unified UI for all roles
A single dashboard that shows everything to everyone is a dashboard that helps no one. Separate views per role were more expensive to build and more useful to use.
No automated performance scoring
Mira generates coaching, not scores. Automating the verdict would have removed the one thing managers still own — and removed accountability for the decision.
No rigid KRA templates
Templates would have forced role standardisation before anyone had agreed on what each role looks like. Flexibility first, standardisation later — after patterns emerged from real use.
No client-facing portal
Clients ask for visibility. Giving it to them changes how PMs behave — they optimise for what the client sees, not what's actually true. Kept DIP internal deliberately.
No notification fatigue
Early builds had email alerts for every status change. Turned off within a week. The platform succeeds when people check it because it's useful — not because it nudges them to.
How it evolved — three versions, one direction
V1
Mid 2025 · Foundation
■Resource allocation Gantt — replace the shared spreadsheet
■Basic timesheet logging for designers
■Google OAuth — no custom auth
■Leadership dashboard skeleton
Focus: Replace the spreadsheets
V2
Late 2025 · Intelligence
■Full KRA review system with 28 dimensions
■Blind peer feedback added alongside manager review
■Leave management replacing the HR Notion table
■RAG status and hygiene scores for PMs
Focus: Make it a decision tool
V3
Early 2026 · AI layer
■Mira: Claude Sonnet coaching layer on KRA output
■Timesheet V3 aggregating all three legacy systems
■Revenue-at-risk view for Leadership
■Kubernetes deployment — moved off shared hosting
Focus: Reduce subjectivity, improve foresight
Adoption — how the team actually changed
Migration from Sheets and Slack
Adoption was the hardest part. The team had years of muscle memory with spreadsheets. DIP didn't win by being better on a feature checklist — it won by removing friction at the moment of the task. A designer logging two hours doesn't want a form. They want to click, confirm, and close the tab.
✓Resource allocation spreadsheet: retired in Week 3 after V1 launch
✓KRA email threads: replaced entirely in V2. No exceptions made for legacy process.
✓HR Notion table: closed when leave management went live. HR migrated voluntarily.
✓Leadership review slide deck: last one was compiled in August 2025. None since.
✓2021 timesheet spreadsheet: finally retired in V3 after aggregation was verified accurate.
40+
Active users across all roles
7
Role-based personas with separate views and access
5
Legacy tools retired since DIP launched
Where it goes next — product direction, not a feature list
Activity tracking
→ shifting to
Delivery prediction
Once there's enough timesheet and allocation history, the question changes from "what happened?" to "what's about to happen?" Risk flags before a PM needs to raise them.
Individual KRA snapshots
→ shifting to
Team capability mapping
When enough KRA cycles have completed, patterns emerge at the org level. Which capabilities are growing? Where are consistent gaps? That's a hiring and training question, not just a review question.
Reporting on what happened
→ shifting to
Supporting what to do next
The platform currently tells you the state of things. The next version should tell you what to decide. Revenue-at-risk is information. "These three projects need attention this week" is decision support.
Stack — technology choices
Claude Sonnet
AI coaching (Mira)