System Architecture Map

The Daily Cycle

24 hours of autonomous execution. Every agent fires on schedule.

5:00 AM

👑
Sovereign Agent — Daily Directive
Reads yesterday's order_execution_log.json. Scores execution 0-100. If <60%, fires EXECUTION FAILURE alert and reissues missed orders with escalation. Issues today's 5 binding orders for Jeff.

            Command
            Accountability
          

8:00 AM

📞

Transcript Pipeline — Drive Ingestion

Pulls new calls from Google Drive. Ingests, classifies, routes to 6 Advisory agents (CEO, CMO, CFO, COO, Coach, Session Recap). New transcripts processed within 1 hour of call ending.

Data 6 Agents

9:00 AM

📡

Market Intelligence — 12 Feed Scan

Scans 12 market feeds. Items scoring 8+ flagged directly to Sovereign context. Competitors, industry signals, enterprise opportunity triggers.

Intelligence Score ≥8 → Sovereign

11:30 AM

🤝

BNI Agent — Chapter Sweep

Tracks Power of One metrics, visitor pipeline, and referral matches. Posts binding orders to Slack. Manages chapter registry and meeting prep.

BNI Slack Orders

12:00 PM

📬

Email Triage #2 — PRISM Inbox

Classifies inbox, creates drafts, flags stale threads. Cross-checks sent folder before surfacing follow-ups. No duplicate recommendations.

2:00 PM

🎤

Speaking Agent — Sweep #1

Reads Sovereign dispatch. Pipeline scan, CFP windows, conference calendar. Builds Jamie's 5 binding orders, posts to Slack. Scores execution from yesterday's log.

Speaking → Jamie

5:00 PM

🔬

Self-Eval Agent — Yesterday's Review

Reviews all yesterday's agent outputs. Logs lessons, calibration notes, quality drift signals. Feeds back to Sovereign context for tomorrow's directive.

Intelligence

8:00 PM

📊

Daily Pipeline Refresh

Prospect stages updated across 248 contacts. GHL CRM queue populated. Client pipeline dashboard updated at clients.prismaiconsultants.com.

CRM 248 Prospects

9:00 PM

🎯
BizDev Agent — Full 6-Step Sweep
Step 1: Gmail scan for big rocks ($5K+). Step 2: Prospect intel build from email + transcripts + calendar. Step 3: Reactivation drafts for expired holds. Step 4: Going-dark prospect outreach drafts. Step 5: 248 prospects → GHL sync. Step 6: Read Sovereign dispatch, calculate pressure mode, build Aubrey's 5 binding orders, post to Slack.

            BizDev
            → Aubrey
            6 Steps
          

11:00 PM

🎤

Speaking Agent — Sweep #2

Outreach drafts for contacts due for follow-up. Jamie brief for tomorrow. Conference intel updated. Pipeline stage transitions logged for execution grading.

Speaking Jamie Brief

34 Active Agents

8 categories. Every agent has a specific role, schedule, and output target.

Command

2 agents

👑

Sovereign PRISM

Math-enforced daily orders. Reads execution logs, grades scores, issues 5 binding orders for Jeff. EXECUTION FAILURE alerts if score <60%.

Runs: 5AM daily · Output: sovereign_dispatch.json

🏢

Sovereign VersAssist

Delivery health monitoring. Capacity utilization tracking for Saleem and the VA team. Surfaces delivery risk before it hits clients.

Runs: Daily · Output: vs_dispatch.json

Advisory PRISM

6 agents

💼

CEO Advisor

Strategic lens on transcripts. Revenue strategy, market position, resource allocation signals.

Triggered: New transcript · Output: ceo_insights.json

📢

CMO Advisor

Marketing and positioning signals from calls. Identifies messaging opportunities and brand gaps.

Triggered: New transcript

💰

CFO Advisor

Financial signals: pricing conversations, budget objections, revenue timing, cost concerns.

Triggered: New transcript

⚙️

COO Advisor

Operational signals: process bottlenecks, team capacity, delivery risk, workflow inefficiencies.

Triggered: New transcript

🧠

Coach Advisor

Client coaching insights. Behavioral patterns, progress signals, breakthrough moments.

Triggered: New transcript · Weekly eval

📝

Session Recap

Produces structured session recaps. Loads prospect_intelligence before running. Plain-text email format, approved by Jeff before send.

Triggered: Post-session · Approval required

Sales Pipeline

3 agents

🎯

BizDev — The Hunter

9PM nightly 6-step sweep. Big rocks, reactivations, going-dark outreach, GHL sync, Aubrey's orders. Reads Sovereign dispatch for pressure mode.

Runs: 9PM nightly · Output: Slack orders + drafts

🎤

Speaking Agent

2PM + 11PM. CFP windows, conference calendar, outreach pipeline. Builds Jamie's 5 binding orders. Now posts orders to Slack with execution tracking.

Runs: 2PM + 11PM · Output: Slack orders + drafts

🤝

BNI — The Connector

11:30AM daily. Power of One tracking, visitor pipeline, referral match engine. Chapter registry + meeting prep. Slack reports.

Runs: 11:30AM · Output: BNI Slack channel

Intelligence

4 agents

📡

Market Intelligence

Scans 12 market feeds daily. Score 8+ items flagged to Sovereign context. Competitors, enterprise triggers, industry signals.

Runs: 9AM · 12 feeds · Score threshold: 8/10

🔬

Self-Eval Agent

Reviews yesterday's outputs across all agents. Logs lessons and quality drift signals. Calibration data feeds Sovereign.

Runs: 5PM · Output: eval_log.json

💸

API Usage Reporter

Tracks all LLM API costs. All bots use Gemini Flash only. Zero Claude API calls in automation. Cost per agent, daily/monthly trending.

Runs: Weekly · Output: usage_report.json

🔍

Prospect Intelligence

Enriches 248 prospects from Gmail, transcripts, GHL, and calendar. Injected into Session Recap and analyze_transcript() pipelines.

Runs: Nightly · Output: prospect_intelligence.json

Game Studio

2 agents

🎮

Hype Man

Game portfolio marketing. Reddit community management, DistroKid music releases, PRISM Studios promotions.

Runs: Daily · Reddit + social

🩺

The Medic

Hourly uptime monitoring on 15 sites. Docker self-heal on failure. Alerts when sites go dark. Auto-restarts containers.

Runs: Hourly · 15 sites · Docker heal

Email + Content

3 agents

📬

PRISM Email Triage

3x daily. Classifies inbox, creates drafts, flags stale threads. Cross-checks sent folder. Tier 1-4 priority sorting.

Runs: 3x daily · Output: Gmail drafts

📫

VersAssist Triage

VersAssist inbox management. Client delivery threads, Saleem/James coordination, billing and project status.

Runs: 3x daily

✍️

Publisher

Sunday 7PM. Full 10/10 copy packages for Aubrey. LinkedIn posts, newsletter drafts, social content calendar. Ready for review + schedule.

Runs: Sunday 7PM · Output: content packages

Data + Sync

4 agents

🔄

GHL Push/Pull

Nightly sync: prospects.json → GHL. Pulls contacts from GHL not in local. Dedup by name + company + email domain.

Runs: Nightly · 248 contacts

💧

Nurture Sequencer

GHL campaign sequences. Session-to-Core conversion flows, reactivation drips, post-event follow-up automation.

Runs: Event-triggered

📅

Calendly Sync

Syncs Calendly bookings to prospect pipeline. New bookings trigger prospect intelligence enrichment and Sovereign context update.

Runs: Real-time webhook

💵

Revenue Attribution

Maps revenue events back to source agents. Which agents drove which conversions. Feeds $1M plan tracking dashboard.

Runs: Weekly

Client-Deployed

4 agents

🛒

SmartPantry

Client-deployed AI agent. Live production system. Custom use case, PRISM-managed infrastructure.

Client: Live

💊

Pharma Lit (Katura)

Pharmaceutical literature research agent for Katura client. Deployed and maintained by PRISM.

Client: Katura

💼

BMI Valuation (Tom Kerchner)

Business market intelligence and valuation agent for Tom Kerchner. Pipeline and market signals. Tom is also a Core retainer client ($2,500/mo) — both deployed agent and retainer.

Client: Tom Kerchner · Deployed agent

📥

Triage (Tom Kerchner)

Email/intake triage agent for Tom Kerchner. Companion to BMI Valuation agent.

Client: Tom Kerchner · Deployed agent

$1M Revenue Map

7 layers. Every layer has an agent driving it and a human closing it.

$7,500

Current MRR

3 Core clients

$10K

Next MRR Milestone

Wendy — Thursday 🔥

$1M

2026 Floor Target

2-3

New Retainers/Mo Needed

Tier 3 Session Clients

+ activations pending

Pipeline in Motion

Wendy → Core · Thursday · Strong close

If Wendy Closes

$10,000 MRR · 4 Core clients

Also Pending

Activations in queue

Layer 1

Core Retainer

$2,500/mo per client · 3 clients ($7,500 MRR) · Wendy → $10K Thu 🔥

Agents Driving

BizDev (nightly orders) Email Triage (follow-ups) Session Recap (conversion)

Active Core: Oursler · Socci · Tom Kerchner · Closing: Jeff

Layer 2

Scale Tier

$5,000/mo per client · Upsell from Core

Agents Driving

Sovereign (flags upsell candidates) Advisory (delivery signals)

Human Closing: Jeff only

Layer 3

Activations

$1,500 per activation · Workshop, sprint, one-time

Agents Driving

Email Triage Nurture Sequencer Publisher

Human Executing: Aubrey (GHL campaigns)

Layer 4

Speaking

$10-20K per engagement · Conference keynotes

Agents Driving

Speaking Agent (CFPs, pipeline) BNI (warm intros)

Human Executing: Jamie (research) + Jeff (stage)

Layer 5

Enterprise Big Rocks

$50-200K per deal · FINRA, Eli Lilly, LCCC

Agents Driving

Sovereign (flags quarterly) Market Intel (FINRA signals) VR Immersive (Eli Lilly play)

Human Closing: Jeff only

Layer 6

Momentum Sprints

$2,500 bridge product · Session clients → Core

Agents Driving

BizDev (bridge candidates) Nurture Sequencer

Drafting: Aubrey · Closing: Jeff

Layer 7

VersAssist MRR

$25K MRR milestone · Execution arm for PRISM clients + external

Agents Driving

Sovereign VersAssist (delivery health) Capacity monitoring Client Deliverables System

Operating: Saleem (capacity) · Aubrey + James (delivery)

Team Capacity Analysis

What agents expect vs. what the time data actually shows.

Aubrey

Since May 2025

1,084

total hrs · ~130/mo

Books/Manuscripts

24%

Graphics/Design/Social

13%

Email/Newsletter

13%

Research/Leads/BNI

13%

Other/Misc

30%

Agent fit: Strong. Aubrey is already doing outreach and pipeline work. BizDev formalizes it. Instead of deciding what to work on, she gets 5 orders ranked by deal value + urgency every night.

Watch: 24% books. If BizDev is in PRESSURE MODE and Aubrey is doing Atticus formatting, something is misaligned. Sovereign execution scoring will surface this within days.

Jamie

Since Aug 2025

433

total hrs · ~60/mo

Books/Manuscripts

53%

Other/Misc

20%

PRISM Shield/Cyber

Transcripts/Recaps

Research/Leads/Email

Reality check: Jamie is a book editor who does speaking support on the side. 53% books vs. 8% research/leads over 7 months. The Speaking Agent treats her as the speaking pipeline executor.

Two options: (1) Speaking orders are landing but being deprioritized by book load — Sovereign execution scoring will catch this within a week. (2) Role reframe needed — books shift to Saba or new contractor, Jamie owns speaking pipeline. The data will decide.

How Accountability Works Now

Day 0 — Order Issued

Sovereign issues order. BizDev/Speaking posts to Slack. Aubrey/Jamie acts (or doesn't). Result logged to order_execution_log.json.

Day 1 — Sovereign Grades

Reads execution log. Scores 0-100. If score <60%: EXECUTION FAILURE header + reissues missed orders at escalated urgency level.

Result — No Invisible Failures

Every missed order surfaces in the next morning's directive. Jeff sees the pattern. Role misalignment becomes visible in the data within 1 week.