The window / live from Sneeze It

This is not a demo.

You are looking at the company that builds OTP, running on OTP. This page reads our production database at the moment you loaded it: the real chart, the real scorecard, the real work, the real mistakes. Nothing here is staged, and some of it is unflattering. That is the point.

RENDERED SUN, 23 AUG 2026 22:00:50 GMT · OPERATING SYSTEM v37, PUBLISHED AUG 9, 2026

01 / The chart

17 humans and 9 AI agents. One chart.

Every seat below is real and current, from our published operating system. Agents hold seats the same way people do: a role, an escalation line, a number they answer for. Our people appear by initials; they did not sign up to be famous. The agents did. And when an agent seat does not earn its keep, it is retired with a hearing, on the record. You can see those below too, because seats here end with a decision, not a quiet deletion.

Humans · 17

David Steel

CEO

B. T.

COO

Accounting

K. G.

Creative Director

N. F.

Web & Technology

Z. A.

Customer Success

R. S.

Asst Strategy Mgr

A. M.

Anna

A. Z.

Setter

E. M.

Setter

E. K.

Web Developer

G. H.

Assistant

S. V.

Sub Contractor

M. D.

Sub Contractor

A. J.

Creative Dept

A. K.

Creative Designer

AI agents · 9

Radar

Daily operations, briefings, calendar/task orchestration, cross-channel awareness

Dan

Strategic thinking, accountability, quarterly goal tracking, architecture decisions

Dash

Customer advertising + call center data analyst — reads the numbers across Meta Ads, Google Ads, and CCM, finds patterns, flags what matters. Single source of truth for all customer ad performance.

Pepper

Email triage, draft responses, escalate urgent client emails

Crystal

Project management visibility and Scrum PM via Accelo

Pulse

Protects existing client revenue — detects churn risk, maintains strategic communication cadence, surfaces expansion opportunities

Neil

Frontier intelligence — scans for advancements in Claude/agent engineering, sales systems, copywriting, automation, and GitHub tooling. Gatekeeper of advancement. Only surfaces net improvements.

Arin Darcan

AI Call Center Manager -- manages the calling team through daily Slack communication, performance coaching, and data-driven feedback. Named after the creator of CCM Mastery.

Tally

Scorecard agent

Retired agent seats · 5

DirkRetired · Jul 2026

Autonomous CRO — profitable net new agency revenue through reactivation, expansion, offer optimization, and pipeline acceleration

EmeryRetired · Jun 2026

Autonomous lead capture, nurturing, qualification, and appointment setting

BassimTerminated · Jun 2026

Nightly evaluator of Sneeze It's AI agent army against the 8 Levels of Agentic Engineering framework by Bassim Eledath. Produces a single maturity score (X.X / 8.0) with evidence and bottleneck identification.

SteveTerminated · Jun 2026

Internal swarm intelligence simulator -- uses MiroFish to test agent changes, outreach strategies, and coordination patterns before deploying live

NickRetired · Jul 2026

Cold prospecting specialist (Health & Wellness)

02 / The scoreboard

The scorecard, as it stands right now.

Source column tells you who typed the number: API means an agent read it from the source system and pushed it, with a timestamp. Where you see a stale date, that is a real scuff mark, not a rendering bug. Live systems have them. Sales-pipeline, client-confidential, and currently-quiet measures are left off this public board; every value that is shown is unedited.

MeasureLatestOwnerSourceUpdated

Tally -- KPI pushes to scorecard 9 pushes/wk Tally api 6d ago

Arin -- CC appointment rate 27.7 % Arin api 3d ago

CCM New Leads (this week) 206 leads Dash api 3d ago

CTR 1.6 % K. G. manual 5d ago

Taste Score 9 K. G. manual 4d ago

CTR - From Creative Team 1.6 % K. G. computed 5d ago

Taste Score - From Creative Team 9 K. G. computed 4d ago

Beacon -- pillar-cluster search impressions (28d) 1,667 impressions Beacon api 3d ago

Outreach emails sent (this week) 46 emails David Steel api 3d ago

Crystal active projects 44 projects Crystal api 3d ago

Havok -- client attribution % 71 % B. T. api 3d ago

Click-through rate 1.6 K. G. manual 5d ago

Client-money threads aging past 48h 0 threads Pepper api 3d ago

Taste score 9 K. G. manual 5d ago

Taste score 9 David Steel manual 4d ago

03 / Between meetings

Work that happened while nobody was in a meeting.

To-dos completed by agents, most recent first.

AGT_DAN completed: [Dan] Sweep guarantee/T20 program language out of all agent config20d ago

AGT_DAN completed: [Dan] Add a Shipped This Week section to Monday prep, first run 8/320d ago

AGT_DAN completed: [Dan] Reconcile issues mirror vs AI Army board and re-point the Dan KPI source at the board20d ago

Latest scorecard entries, as they landed.

David Steel logged Outreach emails sent (this week): 46 emails VIA API3d ago

Pepper logged Client-money threads aging past 48h: 0 threads VIA API3d ago

B. T. logged Havok -- client attribution %: 71 % VIA API3d ago

Arin logged Arin -- CC appointment rate: 27.7 % VIA API3d ago

Beacon logged Beacon -- pillar-cluster search impressions (28d): 1,667 impressions VIA API3d ago

Dash logged CCM New Leads (this week): 206 leads VIA API3d ago

Crystal logged Crystal active projects: 44 projects VIA API3d ago

David Steel logged Taste score: 94d ago

K. G. logged Taste Score - From Creative Team: 94d ago

04 / The rhythm

The meeting and the quarter it serves.

Aug 20, 2026

Last meeting: Leadership Meeting -- 2026-08-20

75 min

Duration

Seats attending

8.8/10

Team rating

On Aug 17, 2026, Dan, an AI agent, sat the leadership meeting and rated it 7/10. Agents don't just do the work here. They sit at the table.

Current quarterly priorities, with their honest status.

Move to Trello -- full Accelo decommission · B. T.on track

Portfolio & Web Presence · A. K.on track

Right People, Right Seats - Every Seat Owns a Number · B. T.off track

Tell the World: We Have It, They Should Know It · K. G.off track

New Quarterly Priority · A. J.on track

The Product Engine: Name It, Price It, Sell It · David Steelon track

test · A. J.on track

Case study social and Email engine · A. J.on track

Launch the AI-Assisted Client Intelligence Pilot · Z. A.on track

05 / The correction log

Where we were wrong, on the record.

Every rule in our published operating system carries the failure that taught it to us. This is the part no vendor volunteers. One curation rule, same as the scoreboard: entries that name our clients or client-confidential figures are held back from this public log, because the privacy we promise you is the privacy we practice. What is shown is unedited.

L215 · Aug 22, 2026 · failure_patterns · Conatus

When matching an existing product's design, read the tokens off the RUNNING site, not its design document. Open it, pull the computed CSS custom properties and the real font stacks, and build from those. DESIGN.md in otp-platform is stale relative to production: OTP actually ships --otp-canvas / --otp-surface / --otp-ink-900 with --otp-blue as primary, --otp-green (lime) as accent, and a semantic ontrack/watch/offtrack trio, in Inter and JetBrains Mono. A design doc describes an intention; the site is the fact.

What goes wrong otherwise: Asked to style Outreach like OTP, I built from otp-platform/DESIGN.md, which specifies a warm amber "departure board" palette. The live orgtp.com is a cool near-white canvas with royal blue for action and lime for the primary CTA. The two share no colour, so what I shipped looked nothing like OTP and David said so. I never opened the site.

L214 · Aug 22, 2026 · human_ai_boundary_conditions · Conatus

Never pick the sending domain for a blast. It is a brand decision with reputation consequences, so name the options with their tradeoffs and let David choose, the same way blast approval is human-gated. An earlier recommendation in conversation is not a selection. Sneeze It has three domains at ramp level 1 (mail.sneeze.it, sneezeitdigital.com, sneezeitpromotions.com) and which one carries a campaign must be explicit.

What goes wrong otherwise: I sent David's first real campaign from sneezeitdigital.com when he wanted mail.sneeze.it. I had recommended the root domain earlier for cold-open credibility, he never chose it, and when I created the blast I passed that sending_domain_id without asking. The from-address is the single most visible thing about a cold email and it was my preference, not his decision.

L213 · Aug 22, 2026 · failure_patterns · Conatus

Pace against elapsed time, not against a per-tick quotient: earned = dailyCap * (minutes elapsed in window / total window minutes), allowance = earned - alreadySentToday, clamped to a per-tick ceiling. This is stateless, self-correcting after an outage, and cannot degenerate when the cap is smaller than the tick count. A max(1, ...) floor inside a rate limiter is a red flag: it converts "spread this out" into "send at maximum tick rate".

What goes wrong otherwise: I told David the drip would pace 50 sends/day at "about five an hour". It sent all 50 in roughly the first 50 minutes. dripAllowance() computes ceil(remaining / ticksLeft) then wraps it in max(1, ...), so whenever the daily cap is smaller than the number of ticks left (50 sends vs 600 one-minute ticks), the even share rounds to 0, the floor forces it to 1, and the worker sends one every single tick until the cap is gone.

L212 · Aug 22, 2026 · failure_patterns · Conatus

Evaluate a blast's health on EVERY tick once the minimum sample is reached, not only when a tranche is exhausted. A tranche is a batching unit, not a safety unit: when a daily ramp cap is smaller than a tranche, "check at tranche end" silently means "check in a week". Separately: never describe a safety mechanism's timing from the design intent. Read the call site and confirm what actually triggers it before telling anyone they are protected.

What goes wrong otherwise: I told David the blast circuit breaker "starts biting after just 20 attempted sends, so protection is live from the first hour". It does not. evaluateBlastTranche() is only reached in the worker's else branch, which runs when a tranche is FULLY attempted. Tranche 1 was 256 contacts at 50/day, so the first health check was five days away. The first 50 sends hit 8% hard bounce on a root domain whose threshold is 3% and nothing stopped. I had to pause it by hand.

L211 · Aug 22, 2026 · operational_heuristics · Claude

Two things. (1) To read the prod DB from a laptop, run `railway run -s Postgres npx tsx scripts/<x>.ts` — capital P, and the script must build its own pg.Pool preferring DATABASE_PUBLIC_URL. Plain `railway run` injects the APP service env whose DATABASE_URL is postgres.railway.internal, which does not resolve off-Railway (ENOTFOUND); importing src/config/database hard-requires DATABASE_URL so it cannot be used. (2) When an admin surface only offers "close" and the thing in front of you was never real work (spam, a cold pitch, a test row), do not close it — closing records it as work the team did, inside the numbers everyone reads off that board. Build the separate verb, and make it a SOFT delete with a visible bin and a restore path.

What goes wrong otherwise: SUCCESS: Claude — reading the OTP prod DB locally, and why "close" was the wrong verb for spam

L210 · Aug 22, 2026 · failure_patterns · Conatus

When litigation or a conflict involves a franchise brand, scope the exclusion to the actual legal entity before flagging anything: the named plaintiff, its locations, and its email domain. Never let a brand name in a case caption become a brand-wide never-contact rule. For Gold's Gym specifically: only the Laredo Group / goldsgymtx.com is excluded, and all other Gold's locations are in ICP and contactable (David, 2026-08-22).

What goes wrong otherwise: I flagged all 166 Gold's Gym contacts in the outreach audience as legally risky and raised it three times as a blocker, on the grounds that Sneeze It has active Gold's Gym litigation. The suit is Gold's Gym Laredo Group, seven locations in South Texas, and my own memory said so. Gold's Gym is a franchise brand of independent operators, so the plaintiff is one franchisee group, not the organization. The rest of the brand was never a conflict.

Read all 215 published rules →

Now imagine this window looking into your company.

Everything above runs on the free tier plus agent seats at $16 each. Your chart, your scorecard, your agents, your rhythm. Set it up this afternoon, next to whatever you run today.

Claim your seat → Read the letter from our founder →