[ NOTES 07 / 08 ] M E S Field notes
000%Scroll
00:00:00 UTCDistributed
[ 07 — Notes ] Essays · monthly

Notes from the field.

What we've learned shipping agents into the messy reality of other people's companies. Specific, opinionated, occasionally wrong.

Continue
All · 26 Architecture Evals Ops Sales Support Hiring
Featured

Read this first.

/analytics
/insights
/reports
/overview
/dashboard
/metrics
/stats
/admin
/cockpit
/agent-mgr
FEATURED ESSAY · ARCHITECTURE
The dashboard
is a graveyard.
9 MIN READ · PRANAY NAVOTH
ArchitecturePranay NavothMay 2026

Every "AI dashboard" we've ever shipped has been used for two weeks and forgotten.

The agents that endure are the ones that act inside the tools the team was already using. Here's the architecture we now default to — and why we delete the UI in our first project meeting.

Read essay
Latest

Recent writing.

EVAL · v 1.0 → v 1.14
PASS RATE96%
Evals7 MIN

Write the eval before the agent.

If you can't grade the output, you can't ship the system. A small ritual that's saved us six months of arguing about quality.

P Pranay N.· FOUNDER
DAY 06 — DAY 21
SHADOW
LIVE
DAYS OF EACH7 / 9
AGENTS RUSHED0
Ops11 MIN

Shadow mode is non-negotiable.

What we run, why we run it, and why no agent of ours has ever gone straight to auto-action.

P Pranay N.· FOUNDER
SDR · DRAFT · v 14
"Saw you just hired three SEs — usually means y'all are pushing into mid-market…"
VOICE MATCH93%
TIME TO DRAFT2.3s
Sales6 MIN

The SDR agent won't write better than your founder.

It will write more of it, faster. Here's why that's actually the win — and how we tune voice without losing it.

P Pranay N.· FOUNDER
GRAPH · vs · CHAIN
DEFAULTchain (loop)
Architecture14 MIN

When to use a graph, when to use a chain.

Our heuristic for picking between LangGraph and a flat tool-calling loop. (We use the loop more than you'd think.)

P Pranay N.· FOUNDER
CSAT · 12 MO
CSAT START72
CSAT END80
Support9 MIN

CSAT goes up when the bot is honest.

What we learned from a year of triage agents: people forgive an AI that admits what it can't do.

P Pranay N.· FOUNDER
HEADCOUNT · CAP
FILLED11
CAPPED AT12
Hiring8 MIN

Why we cap the studio at twelve.

An honest accounting of what we lose by scaling slowly. And what we'd lose more of by scaling fast.

P Pranay N.· FOUNDER
LLM-AS-JUDGE · RUBRIC
FACTUAL9.4
TONE9.1
RELEVANCE8.7
SAFETY10.0
AGGREGATE9.3 / 10
Evals12 MIN

A field guide to LLM-as-judge.

The prompts, the rubrics, the false-positive rates, and the moment we throw it all out and grade by hand.

P Pranay N.· FOUNDER
9 AM REPORT · YESTERDAY
98%
RUNS218
AUTO-RESOLVED203
HOURS SAVED38
Ops6 MIN

The 9am report is the product.

Why we send a one-page daily digest to every client — and why it does more for retention than any feature.

P Pranay N.· FOUNDER
3 LAYERS · GUARDRAILS
1INPUT VALIDATION
2CONFIDENCE GATE
3DEAD-MAN'S SWITCH
PRODUCTION-DEFAULTALL 3
Architecture10 MIN

Three guardrails we ship by default.

Input validation, confidence scoring, and the dead-man's-switch we put in every agent we own.

P Pranay N.· FOUNDER

"The agents that endure are not the smartest. They're the ones whose authors stayed in the room long enough to learn what 'good' meant for that business."

K
Pranay Navoth
Founder, MES — from "The dashboard is a graveyard"

One essay,
every other Tuesday.

No promo, no roundups. Just the writing — sent to the operators building with AI.