Home Services Work Press My Story The Desk Book a Call
The Work › US SaaS Co.

Nobody needed
another tool.
They needed one
built for them.

17 people. Two locations. Three signal sources. No shared truth. I mapped how they actually worked. Then built a system around that reality.

Scroll
17→1
People, two locations, one shared pipeline. No signal falls through.
5 surfaces
Den · Inbox · Pipeline · Pulse · Repository. Each replacing a different coordination failure.
0
Morning standup meetings. Async Pulse replaced them entirely.

The situation

What every morning looked like

Every morning started the same way. Open Slack. Scroll back through the overnight messages. Reconstruct what was urgent, what was waiting, what someone had asked and nobody had answered. Attend a standup to say it all out loud. By 10am, nothing from that standup was written anywhere. The frustrating part wasn't the number of apps. It was the uncertainty. Something could have slipped overnight, and there was no reliable way to know.

Meanwhile, signals were arriving in three places simultaneously. A bug reported by an operator in Gmail. A feature request buried in a Slack thread. A decision from last week's call, transcribed by Fireflies, never actioned. The ones that got picked up were the ones that reached the right person at the right moment. The others disappeared.

Features moved through stages (Discovery, In Progress, Review, Done), but the movement was invisible. Nobody knew SSO Login had been sitting in Review for nine days. Nobody knew Ryan had seven active tasks while another engineer had two. That information existed. It just lived nowhere anyone could see it.

Team size
17 people across 2 locations
Product lines
Growth · Enterprise
Apps
Mobile App · Admin App · Client App
Signal sources
Slack · Gmail · Fireflies
My role
AI Transformation Lead

How I worked

I don't start with tools.
I start with how you work.

Most teams I work with can already feel that something is broken before they can name it. They operate every morning on partial information, and they know it. That's where I start. Every project runs through the same four phases. No phase gets skipped. The order never changes.

01
Diagnose

Two weeks. No code. I sit with the team and map exactly what breaks every morning: which tools are open, which signals get missed, which workarounds people built because the official process didn't work.

02
Define

Before anything gets built, I define what "fixed" looks like for each person at each desk. What does a developer's morning look like when this works? That answer drives every decision in the build.

03
Build

I build in phases, not one big launch. Each surface ships, gets used by real people, and comes back with real feedback. The final system reflects that loop, not a spec written in week one.

04
Embed

A system people don't own doesn't stick. The last phase trains the team on the why, not just the how. They understand it well enough to adapt it, without calling me.

What changed

Four things that used to break
every single day.

01
Morning chaos

The day started
in four different apps.

Before
4 apps. 1 standup. Nothing written.
After
Den. One place. Day starts there.
02
Signals falling through

The request existed.
Nobody saw it.

Before
3 sources. No order. Luck, not process.
After
Inbox. Classified on arrival. Nothing slips.
03
No shared view

Nobody knew what
anyone was doing.

Before
Blockers invisible. Load invisible. Status in Slack.
After
Pipeline. 17 people. One board. AI flags blockers.
04
Daily standup drain

A 30-minute meeting
to say what a system should.

Before
17 people. 30 mins. Spoken. Nothing tracked.
After
Pulse. Async. Auto-published. Meeting cancelled.

One example from the build

Den: The private workstation.

Every developer's day starts here now. Private, incognito, and completely theirs before they join the shared pipeline.

Surface 01 | Den

The private workstation.
Incognito. Yours alone.

No private space existed. Planning happened across Notion, personal notes, and calendar, all disconnected from the actual pipeline. The Pipeline showed 17 people's work. There was no way to think before broadcasting.

Den is a private, incognito workstation inside the system. Tasks assigned to you, @mentions, personal cards, and cards you're collaborating on all flow into your Den inbox. You decide which 3 to commit to. You lock them as your Daily Scrum and broadcast to the team. The workspace stays yours: Excel, Word, calendar, sticky notes, reminders, all inside Den, no app switching.

Impact
Morning context-switching eliminated. Each person's day starts in one private place: their own tasks, their own tools, before they join the shared pipeline.
The Problem
Four apps open before the day started: Notion for notes, Google Calendar for reminders, Slack for @mentions, Pipeline for tasks. No single place. No private space.
Den
Everything in one private surface. Tasks filtered to yours. Tools embedded. Daily Scrum broadcast with one click. Team sees your commitment. Process stays private.
TaskFlow · Den
Den · Your Private Workstation
Private · Incognito
My Cards
"

I didn't realise how much time I was spending just figuring out where to start each morning.

Developer · US SaaS Co. · Week Three

Surface 02 · Inbox

Signals were arriving.
Nobody had one place to act on them.

External signals scattered across three apps. Gmail held client requests. Slack carried operational noise. Fireflies captured meeting action items that nobody tracked. Three apps open every morning before you could know what needed attention. The signal existed. The problem was where it lived.

The Inbox pulls every inbound signal into one surface. AI reads each one on arrival and classifies it: needs a task, waiting on you, FYI, or noise. Deadline-bound items are flagged. Meeting action items that never became tasks are caught. Signals tied to open Pipeline items are linked, not duplicated. You open one surface. You know what arrived. You act.

Impact
7 signals. Classified. Actioned. Done in under 4 minutes. Three apps closed.
The Problem
Slack open. Gmail open. Fireflies notes open. 20 minutes reconstructing what arrived overnight and whether any of it mattered. No single place to decide.
The Fix
Inbox opens. AI has already read everything. 5 signals. 3 need a task. 1 waiting on you. 1 FYI. Actions taken in 4 minutes. Apps closed.
TaskFlow · Inbox
AI Triage · 7 signals · 4 need action · 2 waiting on you · 1 FYI
GMAIL legal@client.com
Needs task
User data export request, 30-day SLA window
Requesting full account export: usage history, billing records, support interactions. Compliance window opens on receipt.
↳ AI: Deadline-bound · Assign before end of week Create Task →
SLACK #dev-ops · Ryan Carter
Needs task
Webhook retry still failing after last night's fix
↳ AI: Linked to overdue Pipeline item · Escalate or reassign Create Task →
FIREFLIES Sprint Planning · Yesterday
Needs task
Action item: Ryan to review Data Sync accuracy before Thursday
↳ AI: Untracked meeting action item · Never became a task Assign to Ryan →
FIREFLIES Product Review · Mon
Needs task
Decision: Surge Pricing Engine to move to In Progress this week
↳ AI: Decision logged · No Pipeline task created yet Create Task →
SLACK #general · Sarah Mitchell
Waiting on you
Anyone have the Fireflies transcript from the mobile API call?
↳ AI: You were tagged · Reply or forward Reply →
GMAIL contracts@enterprise-co.com
Waiting on you
MSA renewal, awaiting your countersignature
↳ AI: Sent 3 days ago · No reply logged Reply →
GMAIL updates@partner.io
FYI
Q2 integration summary, no action required
↳ AI: Informational · Archived
TaskFlow · Pipeline
Discovery 3
Real-time Data Sync Improvements
Growth Mobile App
Surge Pricing Engine
Growth
Driver Rating System Overhaul
Growth Rider App
In Progress 4
SSO Login · Admin App
Enterprise Operator
AI Booking Suggestions
Growth
Billing Portal Upgrade
Enterprise
Mobile Push v2
Growth Mobile App
Review 3
Webhook Retry Logic
Enterprise
Workflow Optimisation v2
Enterprise
API Rate Limiting
Enterprise
Done/Launch 8
Push Notification · Mobile
Growth
Analytics Dashboard v2
Enterprise
Route Optimisation Engine
Growth Operator
✦ AI Board Summary
2 overdue tasks SSO Login (Critical) needs immediate attention
Workload imbalance Ryan has 7 active tasks. Consider redistributing.
3 tasks ready to advance All subtasks complete: "Webhook Retry Logic" and 2 more
"Review" is piling up 5 tasks in this stage. Possible bottleneck.
Surface 03 · Pipeline

The shared board.
Seventeen people. One truth.

Feature status on a 17-person team required asking. Overloaded engineers were invisible until 1:1s. Stalled tasks only surfaced when someone chased them on Slack. Nobody knew how long anything had been waiting.

The Pipeline is a 4-stage Kanban where health is measured in days in stage, not progress percentage. Green at 0 to 2. Amber at 3 to 6. Red at 7 or more. No one has to ask how long something has been waiting. Above the board, AI Board Summary runs continuously: overdue tasks flagged by name, workload imbalance detected by count, bottlenecks called out by stage accumulation.

Impact
No chasing. No status meetings to generate visibility. Bottlenecks visible before they become crises.
The Problem
Feature status discovered in meetings. Overloaded engineers found in 1:1s. Stalled tasks only visible when someone chased them on Slack.
The Fix
Health dots visible for every task. AI surfaces bottlenecks, imbalances, and ready-to-advance items automatically. No chasing. No meetings to generate visibility.
Surface 04 · Pulse

The async standup.
The meeting that stopped being necessary.

The standup meeting existed because status had nowhere better to live. Verbal. Forgotten by 10am. No record of who committed to what. Managers chased updates in 1:1s. The meeting was not the problem. The missing structure was.

Each morning: three written commitments, timestamped, visible to everyone. Each evening: a Wrap auto-generates, committed vs done vs carried vs dropped, by task name, by person, permanently searchable. No meeting needed to produce this. The data was already in the Pipeline.

Impact
The standup became optional in week one. By week two, nobody scheduled it.
The Problem
30-minute standup. Verbal status. Nothing written. By 10am, no one remembered who committed to what.
The Fix
Daily Scrum broadcast at 9am: 3 commitments written, timestamped, searchable. Evening Wrap: committed vs done vs carried vs dropped. By name. By task.
TaskFlow · Pulse
Today · 6
Daily Scrum
Wrap
All team
PW
Priyankka Wadhwa DAILY SCRUM 9:04am
1SSO Login · Admin App (Critical review)
2Webhook Retry Logic · final QA sign-off
3AI Booking Suggestions · team alignment call
RC
Ryan Carter WRAP 6:12pm
✓ Done 2 · committed 3 (2 closed)
✓ Webhook Retry Logic · PHP Backend
✓ SSO endpoint scaffolding
→ Workflow Optimisation unit tests CARRIED
SM
Sarah Mitchell DAILY SCRUM 9:22am
1Admin App QA · Enterprise flows
2Push notification regression testing
Asking: anyone have the Fireflies transcript from last Wednesday's mobile API call?
JT
Jessica Torres WRAP 5:58pm
✓ Done 3 · committed 3 (all closed)
✓ Workflow Optimisation v2 · QA complete
✓ API Rate Limiting · staging deploy
✓ Billing Portal · spec signed off
TaskFlow · Repository
All time
Growth
Enterprise
Slack
Critical
Data Sync · 3 versions
v3
Data Sync · Accuracy Overhaul
Growth · Mobile App · Launched Apr 2026 · Ryan Carter
Slack
v2
Data Sync · Background Mode Fix
Growth · Mobile App · Launched Feb 2026
Gmail
v1
Data Sync · Initial Release
Growth · Mobile App · Launched Nov 2025
Fireflies
Workflow Optimisation · 2 versions
v2
Workflow Optimisation · ML Rework
Enterprise · Client App · Launched Mar 2026 · Jessica Torres
Slack
v1
Workflow Optimisation · Original
Enterprise · Client App · Launched Jan 2026
Gmail
SSO Login · 2 versions
v2
SSO Login · Admin App Rollout
Enterprise · Admin App · Launched Apr 2026 · Ryan Carter
Gmail
v1
SSO Login · Operator App
Enterprise · Operator · Launched Feb 2026
Slack
Push Notifications · 2 versions
v2
Push Notifications · Ride Status Alerts
Growth · Mobile App · Launched Mar 2026 · Sarah Mitchell
Fireflies
v1
Push Notifications · Booking Confirmed
Growth · Mobile App · Launched Dec 2025
Gmail
Surface 05 · Repository

The permanent archive.
Everything shipped. Nothing lost.

Shipped features left no institutional trace. "What did we ship this quarter?" required searching Slack history or asking someone who might remember. The decision history, why Data Sync went through three iterations, which signal triggered each one, existed nowhere.

The Repository archives every launched feature with its product line, app, assignee, source signal, and full version lineage. Data Sync v1 to v2 to v3, the complete evolution visible in one view. Filterable by date, priority, source, and product line.

Impact
The retrospective question that took a 20-minute meeting now runs in seconds.
The Problem
Shipped features existed nowhere. "What did we ship?" required searching Slack history. Version history was in your head.
The Fix
Every launch archived with version lineage, source signal, assignee, product line. Full history searchable in seconds.

The outcome

What actually changed

$340K
Saved in 90 days. Not from a new tool. From removing the work the team should never have been doing.
0
Morning standup meetings. Pulse replaced them with async Daily Scrum broadcasts.
3
Signal sources (Slack, Gmail, Fireflies) consolidated into one Inbox triage queue.
17
People, one shared pipeline, health visible for every task without asking anyone.
∞
Launch history. Every feature, every version, every source signal, archived forever.

The metric that doesn't show up on a dashboard: the first Monday morning when nobody asked 'what happened to that feature?' Because everyone already knew.

The decisions behind the build

Every design choice was deliberate

TaskFlow is not generic project management configured for US SaaS Co. Here are three decisions that define how it works.

Why 3 commits, not 5
Five feels achievable. Three forces a choice. The discipline of picking 3 and locking them creates the focus that a standup is supposed to produce but rarely does. The AI surfaces candidates by criticality (overdue and critical tasks appear first), so the selection is informed, not arbitrary.
Why health is days in stage, not progress %
Progress percentage requires manual input and is almost always optimistic. Days in stage is automatic and honest. A feature that has been in Review for 9 days has a red dot regardless of what anyone says about progress. The signal is structural, not subjective.
Why Wrap shows carried and dropped separately
Carried means the task continues tomorrow. Dropped means something displaced it. The distinction matters. A manager who sees "dropped" can ask why in the next interaction, not with blame, but with information. Wrapping everything as "not done" loses the signal about what actually happened.
"

AI implementation isn't about which tool you choose. It's about whether the tool was built for the way your team actually works. Or whether your team is expected to adapt to a tool built for someone else.

Priyankka Wadhwa · Let's Execute AI

A note

If you're reading this and recognising your own morning routine in what I described. That recognition is the diagnostic. You don't need a formal audit to know which part is broken. You already know. The question is whether you do something about it before the workarounds become permanent.

Every build I do starts with two hours at one desk. Not a proposal. Not a strategy deck. Two hours mapping exactly what breaks and in what order.

What this page doesn't show is the conversation in week three when a developer says: I didn't realise how much time I was spending just figuring out where to start each morning. That is what changes first. The $340K comes later.

Priyankka Wadhwa · Let's Execute AI

Before you decide

See the system
running live.

This is not a mockup. The system built for US SaaS Co. runs live. Before you commission your own build, book 30 minutes to see exactly how it works.

Live Demo · 30 min

Walk through the working system. See the four surfaces. Ask how it handles your team's specific coordination failures.

Book a demo call →

No pitch. No deck. Just the system.

Free Download

10-Step AI Execution Framework

The same sequence I ran on this project. Use it as a checklist before your next AI build, or as a diagnostic if something you've started isn't moving.

Download PDF

Work with me

Your team has the
same coordination failures.

Every team has a version of the standup problem, the dropped signal problem, the disappeared work problem. Left alone, the workarounds become the system. The standup stays. The dropped signals multiply. By the time it shows up as a people problem, it has been a coordination problem for two years. The build looks different each time. The starting point is always the same.

Start with a diagnostic → See the services →
LET'S EXECUTE AI · BOOK A CALL · PW