Decision Lag

Decision Lag detects high-risk decisions that are viewed repeatedly but not acted on, then automatically triggers Slack alerts and creates Salesforce follow-up tasks - powered by Tableau analytics.

Comment 2

Inspiration

During day-to-day work, I noticed something frustrating:
important insights get created, reviewed, and even discussed — but nothing happens afterwards.

People open dashboards, read reports, say “this is important”, and then move on.
The decision doesn’t get actioned, and the organization silently accumulates what I started calling Decision Debt.

This hackathon project was inspired by that gap between “awareness” and “action.”

Problem Statement

Organizations generate tons of insights, but they don’t track follow-through.
This creates:

  • missed revenue opportunities
  • delayed escalations
  • repeated reviews of the same issue
  • wasted analyst and executive time

I wanted to build a system that detects “decision lag” automatically and pushes teams toward action at the right moment.

What I Built (Decision Lag)

Decision Lag is a lightweight decision execution layer built on Salesforce.

It tracks:

  • decision insights
  • how many times they were viewed
  • their risk/impact
  • whether they were acted on

Then it automatically triggers:

  1. A Slack alert when risk is high and views keep increasing
  2. A Salesforce Task for the owner to take follow-up action
  3. Updates the record so we know an alert was already sent (avoiding spam)

This turns passive analytics into active execution.

How It Works (Architecture)

1) Salesforce as the System of Record

I created a custom object:

  • Decision_Insight__c

Key fields:

  • Title
  • Domain
  • Risk / Impact
  • View_Count
  • Slack_Alert_Sent
  • Status

2) Web App (“Decision Debt Studio”)

I built a web experience to list insights, simulate “views”, and show decision lag.

When a user views an insight:

  • the system increments view count
  • updates “last viewed”
  • calculates a debt score (MVP)

3) Automation using Salesforce Flow

A record-triggered flow monitors Decision_Insight__c.

If:

  • Risk >= threshold
  • Views >= threshold
  • Status != Actioned
  • Slack Alert not already sent

Then it performs:

  • Slack alert
  • Task creation
  • Record update (Slack_Alert_Sent = true)

What I Learned

This project taught me how quickly you can build an execution layer when you combine:

  • Salesforce as the core platform
  • Flow as an automation engine
  • Slack for real-time team engagement
  • Tableau Next for insight visibility / dashboards

It also made me appreciate how small product ideas (like “track view count + risk”) can create a measurable impact.

Challenges I Faced

Some challenges I ran into:

  • Flow debugging issues (rollback mode vs real runs)
  • Task creation worked but I initially thought it failed because tasks were created under a different user
  • Validation rules / required fields blocked record updates
  • Ensuring Slack alert + task creation happens only once (no alert spam)

Each issue was solved iteratively and improved the final architecture.

What’s Next

If I continue this project, I’d expand it to:

  • embed Tableau Next dashboards directly into the web app
  • detect “repeat viewing without action” patterns
  • add AI summarization (“what’s blocking this decision?”)
  • escalate alerts automatically if tasks remain uncompleted
  • More Insights and Dashboard

Outcome

Decision Lag reduces Decision Debt by making sure that: Insights don’t just get viewed — they get acted on.

Built With

Share this project:

Updates