How to Track Decisions from Meetings: The End of Zombie Decisions in Engineering

TL;DR

We've all been there: staring at a block of code, wondering why it was written that way, only to find the answer is buried in a Zoom recording from six months ago.

Manual decision logs fail because they require discipline nobody has. Fenlo automates decision capture from meetings and links decisions directly to the code they affect—so you never lose the "why" again.

---

The Zombie Decision Problem

The scariest thing in software development isn't a production bug. It's a Zombie Decision.

You know the one. It's a decision that walks among you, influencing your architecture and your roadmap, but nobody remembers where it came from.

• "Why are we using this specific library?"
• "Who decided to deprecate the v1 API?"
• "I thought we agreed to hold off on this feature?"

The answer is usually: "Oh, Dave mentioned it in the sprint planning meeting last November."

But Dave is gone. And the meeting recording is 60 minutes long and titled "Weekly Sync - Nov 12." Good luck finding that timestamp.

Zombie decisions are everywhere. They haunt your codebase, creating confusion, blocking progress, and causing teams to either:

1. Re-debate settled questions — Wasting hours discussing something that was already decided
2. Make contradictory decisions — Because nobody knows the original context
3. Blindly follow outdated choices — Because "that's just how we do it"

At Fenlo, we realized that engineering teams are drowning in lost context because our tools aren't talking to each other. Here's how to fix the "Decision Black Hole."

---

Why Decisions Get Lost

Engineering decisions don't disappear because teams are careless. They disappear because the systems we use to make decisions are disconnected from the systems where decisions matter.

Where Decisions Actually Happen

The problem is clear: 70% of decisions happen in meetings and chat—the two places with the worst searchability and no automatic connection to code.

The Disconnection Problem

Even when decisions are captured somewhere, they're disconnected from the work:

• The meeting recording sits in cloud storage
• The task lives in your project tracker
• The code lives in GitHub
• The discussion lives in Slack

These systems don't talk to each other. So when someone asks "why was this built this way?", they have to manually search across four different tools—if they even know where to look.

---

The Old Way: Manual Decision Tracking

If you search for "how to track engineering decisions," the standard advice is to add more process.

Approach 1: Architecture Decision Records (ADRs)

ADRs are the gold standard for decision documentation. You create a markdown file in your repo documenting:

• The context and problem
• The options considered
• The decision made
• The consequences

The theory: Decisions live with the code. Future engineers can find them.

The reality:

The fatal flaw: ADRs require someone to stop and write after the decision is made. In practice, this happens maybe 20% of the time. The other 80% of decisions become zombies.

Approach 2: Meeting Minutes

Someone (usually the person with the lowest commit count that week) is assigned to take notes.

The theory: Capture everything discussed, share with the team.

The reality:

• Notes are incomplete (the note-taker can't capture everything while participating)
• Notes aren't linked to tickets or code
• Notes become a graveyard nobody reads
• Quality depends entirely on who's taking notes
• The "why" behind decisions is often lost in summarization

Meeting minutes capture what was discussed. They rarely capture why decisions were made.

Approach 3: The "Just Slack Me" Approach

Decisions happen in threads. Fast, convenient, async-friendly.

The theory: It's documented in Slack, we can find it later.

The reality:

• Slack retention limits (90 days for free/standard plans)
• Impossible to search effectively after a few months
• Context is scattered across multiple threads
• No connection to the code or tasks that resulted

Trying to find a specific decision in Slack after 6 months is a special kind of torture.

Why Manual Approaches Always Fail

All manual approaches share the same fundamental problem: they require extra effort from people who don't have extra time.

Writing an ADR takes 15-30 minutes. Taking good meeting notes requires dedicated attention. Organizing Slack threads requires discipline nobody has.

When the choice is between documenting a decision and shipping the feature, shipping always wins.

---

The Fenlo Way: Automated Decision Capture

At Fenlo, we believe you shouldn't have to stop coding to document your decisions.

The problem isn't that teams aren't writing things down. They are—in Slack, in meeting transcripts, in commit messages. The problem is that this "writing" is disconnected from the code it affects.

Fenlo connects everything automatically.

1. Meeting Intelligence: Auto-Extract Decisions from Calls

Instead of relying on manual notes, Fenlo uses Meeting Intelligence to automatically process your meeting recordings.

When you have that 45-minute debate about database sharding on Zoom, Fenlo doesn't just transcribe—it extracts structure:

From a typical architecture discussion, Fenlo captures:

• The Problem: "Database latency is increasing as we scale"
• Options Discussed: "Vertical scaling vs. horizontal sharding vs. read replicas"
• The Decision: "Implement sharding by Tenant ID"
• The Reasoning: "Tenant isolation provides natural boundaries; consistent with our multi-tenant architecture"
• Action Items: "Alex to prototype by Friday; Sarah to update the data model doc"
• Participants: Who was in the room and who drove the decision

This happens automatically. No one needs to take notes. No one needs to write an ADR after the meeting. The decision is captured in real-time.

2. The Knowledge Graph: Linking Talk to Code

A decision captured in a meeting is useless if it isn't connected to the work.

Fenlo's Knowledge Graph creates these connections automatically:

Meeting: "Architecture Review - Database Scaling"
  └── Decision: "Implement Tenant ID Sharding"
        ├── Task: "ARCH-234: Implement database sharding"
        │     └── PR: "#892: Add tenant-based sharding layer"
        │           └── Files: db/sharding.ts, models/tenant.ts
        ├── Action Item: "Alex to prototype by Friday"
        └── Related Discussion: Slack thread #database-scaling

Every decision is linked to:

• The meeting where it was made
• The task that implements it
• The code that results
• The people who have context

3. The "Why" Search: Natural Language Queries

Traditional search fails for decisions because it's keyword-based. Search "database decision" and you get hundreds of irrelevant results.

Fenlo's AI-powered search understands intent:

"Why did we decide to shard by tenant ID?"

Fenlo doesn't just return keyword matches. It traverses the Knowledge Graph to find:

1. The meeting where sharding was discussed
2. The specific timestamp with the decision
3. The reasoning that was stated
4. The alternatives that were considered
5. The people who made the decision

Result: Complete context in seconds, not hours of archaeology.

4. Automatic Decision Timeline

For any piece of code or system, Fenlo can show you the complete decision history:

Payment Service Decision Timeline:

New engineers can see the complete evolution of any system—every decision, every change, every "why"—without digging through old recordings.

---

Real-World Example: The Sharding Decision

Let's trace how a decision flows through Fenlo:

The Meeting

Your team has a 45-minute Architecture Review discussing database scaling. Multiple approaches are debated. Eventually, the team agrees: shard by Tenant ID.

Without Fenlo:

• The recording sits in Google Drive
• Someone might write notes (probably not)
• The decision lives in people's heads
• Six months later, nobody remembers why

With Fenlo:

• Meeting is automatically transcribed
• Decision is extracted: "Implement Tenant ID sharding"
• Reasoning is captured: "Tenant isolation provides natural boundaries"
• Alternatives discussed are logged: "Considered vertical scaling, rejected due to cost"
• Timestamp is marked: 23:45 in the recording

The Implementation

Alex creates a task and starts implementing. When the PR is merged, Fenlo automatically links:

• The PR → to the Task → to the Decision → to the Meeting

Three Months Later

A new engineer is reviewing the sharding code. They wonder why Tenant ID was chosen as the shard key.

Without Fenlo:

• They search the wiki (nothing)
• They search Slack (too much noise)
• They ask around ("I think Dave decided this?")
• They spend 2 hours and still don't have full context

With Fenlo:

• They click on the code
• They see the linked Decision
• They watch the 2-minute meeting snippet where it was explained
• They understand the full context in 5 minutes

The zombie decision becomes a documented, traceable choice.

---

Implementing Decision Tracking with Fenlo

Here's how to get started:

Step 1: Connect Your Meeting Platform

Link your Zoom, Google Meet, or other meeting recordings to Fenlo. Past recordings can be processed retroactively.

Step 2: Let Fenlo Process Your Meetings

Upload recordings or connect live. Fenlo automatically:

• Transcribes with speaker identification
• Extracts decisions and action items
• Identifies topics and entities discussed

Step 3: Review and Refine

Fenlo's AI is good, but not perfect. Review extracted decisions and confirm or adjust. Over time, the system learns your team's patterns.

Step 4: Connect Your Code and Tasks

Link GitHub and your task tracker. Now decisions automatically connect to the work that implements them.

Step 5: Search and Discover

When someone asks "why did we build it this way?", they can query Fenlo and get the answer—with full context and citations.

---

The ROI of Automated Decision Tracking

Time Saved

Quality Improved

Strategic Value

• Faster onboarding: New engineers understand "why" without archaeology
• Better decisions: Teams can see what was considered before, avoiding repeated debates
• Preserved institutional knowledge: Decisions don't leave when people leave
• Audit trail: For compliance-sensitive decisions, complete traceability

---

Stop Relying on Human Memory

Your brain is for solving complex engineering problems—not for remembering which Tuesday in 2023 you decided to use a specific library.

If you're a Tool-Fatigued Tech Lead constantly re-explaining the same decisions, or an Overwhelmed CTO watching context disappear when people leave, you know this pain.

Stop treating decision tracking as a manual chore. Let the tools do the heavy lifting.

Fenlo captures decisions automatically, links them to your code, and makes them searchable forever. No extra process. No discipline required. Just context that's always there when you need it.

---

Conclusion: Kill the Zombie Decisions

Zombie decisions haunt every engineering organization. They cause confusion, waste time, and lead to contradictory choices.

The solution isn't more process or better discipline. The solution is automation.

Fenlo captures decisions from where they actually happen—meetings and discussions. It links them to the code they affect. It makes them searchable with natural language.

The result: When someone asks "why was this built this way?", the answer is always one click away.

Stop letting decisions die in meeting recordings. Stop losing the "why" when people leave. Stop re-debating questions that were already settled.

Start using Fenlo and kill the zombie decisions for good.

---

Key Takeaways

70% of engineering decisions happen in meetings and chat—the hardest places to search

Decisions made in Zoom calls and Slack threads are nearly impossible to find later. Meeting recordings are unsearchable; chat threads scroll away. Fenlo's Meeting Intelligence automatically extracts decisions from recordings and makes them searchable, with full context preserved.

Manual decision tracking (ADRs, meeting notes) fails because it requires discipline nobody has

Writing an ADR takes 15-30 minutes. Taking good meeting notes requires dedicated attention. When the choice is between documenting and shipping, shipping wins. Fenlo captures decisions automatically—no extra effort, no discipline required.

Fenlo links decisions to code through a Knowledge Graph

A decision captured in a meeting is useless if it's not connected to the work. Fenlo's Knowledge Graph automatically links decisions → tasks → PRs → code. When someone looks at code, they can see the meeting where the approach was decided.

Natural language search finds decisions by intent, not just keywords

Traditional search fails for decisions because "database decision" returns hundreds of irrelevant results. Fenlo's AI search understands intent: ask "Why did we choose PostgreSQL?" and get the actual decision with reasoning, alternatives considered, and the people involved.

Decision tracking ROI: from 30-60 minutes of archaeology to 2 minutes of lookup

Finding a past decision traditionally takes 30-60 minutes of searching across tools. With Fenlo, it takes 2 minutes. Multiply by the dozens of times per month this happens, and the time savings are substantial—plus the quality improvement of having complete context.

---

Ready to kill the zombie decisions in your organization?

Try Fenlo free →