Roderic Andrews — Case Study Roderic Andrews rodericandrews.com ↗
Case Study · Political Intelligence Platform

War Room AI.

Six agents. Every channel. Always watching.

A production intelligence platform for political campaigns — social, news, ads and public filings monitored around the clock, turned into a living picture of strengths and threats, by AI agents that already know the campaign.

live at app.thewarroom.ai
01 · The BriefWR·CS

The problem, the build, the now.

The problem Signal arrives faster than any team can read.

A campaign's mentions, news cycle, ad spend and opposition moves land across a dozen channels at once. Small teams miss things — and in politics, what you miss at 2am is on the morning shows at 9.

What I built A staff that never sleeps.

War Room AI: six specialist AI agents running continuously over a live knowledge index of the campaign — monitoring every channel, scoring it into strengths and threats, alerting the right person, and drafting the brief before anyone asks.

How it looks now Live, multi-tenant, in production.

Running at app.thewarroom.ai with isolated tenants per campaign. Below: every major surface, exactly as a client sees it. Want it walked through live? That's the call.

02 · The Platform12 surfaces

What it does.

This is not ChatGPT with a skin on. ChatGPT and Perplexity start every conversation from zero — they know the internet, they don't know you. War Room is the opposite: your data — every mention, story, ad, filing and document you feed it — coming in daily, adapting, re-analyzed over months. What accumulates is a wall of intelligence about your race that can't be bought, scraped or prompted into existence anywhere else. That wall is the moat.

The shape is simple: input, analysis, output. Intelligence ingests the firehose — social, news, prediction markets, your own documents. SWOT turns it into a living picture of strengths, weaknesses, opportunities and threats, scored and re-scored continuously. The Workspace turns that picture back into work product: briefs, memos, talking points, a full branded presentation — in minutes, not mornings.

And it isn't a concept. War Room is in production for lobbyists, lobby groups, trade associations, charities and companies — including big ones — pulling solid MRR today. The surfaces below are what those customers use every day.

War Room Command Dashboard — SWOT radar with live strengths, weaknesses, opportunities and threats
Command Dashboard · SWOT radarWR·01
WR·01

Command Dashboard

The whole race on one screen.

Real-time metrics with a SWOT radar — strengths, weaknesses, opportunities and threats surfaced the moment they emerge, scored and prioritized instead of merely listed.

Use case "Where do we stand?" — answered before coffee.

Monday, nine days out, the candidate asks the question every campaign lives by. The dashboard answers in one glance: what's working, what's threatening, what moved overnight. No fourteen-tab briefing, no analyst scramble — the picture is already on the wall.

War Room Daily Clips — scored quick-insert rail beside the day's sourced, ranked link digest
Daily Clips · the day's most relevant linksWR·02
WR·02

Daily Clips

The morning's firehose, pre-read.

The daily SWOT brief says where the race stands; directly beneath it, Daily Clips assembles the evidence — every news story, prediction-market shift and social mention from the last 24 hours, scored, ranked and linked in one forwardable document.

Use case The whole day, one document.

Wednesday, 7am: two source-backed items made the cut — a permitting story and a water-rule shift — each with its link and a "why this matters" line. The client reads it before the staff meeting; the staff meeting starts informed.

War Room Intelligence Hub — live news and media feed with sentiment, platform mix and agent actions
Intelligence HubWR·03
WR·03

Intelligence Hub

Every channel, watched around the clock.

Live social monitoring plus your own documents — the AI reads everything against a knowledge index built from the campaign itself, not from generic training data.

Use case The story starts on a blog. It's already tracked.

A policy narrative breaks on a niche blog at lunch and jumps to X by dinner. The Hub has read both since minute one — alongside your polling memos and debate transcripts, uploaded weeks ago. The analysis cites your documents, not the internet's guess.

War Room Publication Bias Atlas — 3,039 outlets plotted by lean and factual reporting, coverage-mix bar and rated roster
Publish · Publication Bias AtlasWR·04
WR·04

Publication Bias Atlas

Know the messenger.

A built-in directory of the press itself — every outlet in your coverage rated for lean and reliability, Ground-News-style, without leaving the platform. When a story lands in the Hub, you already know who is telling it and where they're telling it from.

Use case "Is this news, or is this their news?"

A hostile-leaning outlet breaks a story about the client. The Atlas tags its lean instantly, so the response is calibrated to the messenger — not a panic, not a shrug, a decision.

War Room X.com dashboard — narrative momentum, reach and engagement over a 60-day window
X.com DashboardWR·05
WR·05

X.com Dashboard

The room's mood, quantified.

A dedicated social surface — narrative momentum across any window, reach and engagement at a glance, wins logged, and an AI narrative review of the posts that moved the room.

Use case Sixty days of noise, compressed to one number.

"+8 net sentiment, 72% confidence, 102.4M reach." The candidate sees the trend before the trackers do — and the weekly report writes itself from the same screen.

War Room Ad Insights — public intel view: opponent spend, new ads this week, FEC raised, and what changed
Ad Insights · public intelWR·06
WR·06

Ad Insights

See the money.

Your own paid delivery and your opponents' public ad spend side by side — Meta and Google — without logging into every platform at midnight.

Use case "They've gone dark." No — they haven't.

The opponent claims silence. Ad Insights shows their Meta spend tripled across three swing counties last week, plotted next to your own delivery. The media buy answers reality, not the press release.

War Room AI Workspace — brief editor with export to Google Docs, DOCX, PDF and email
AI Workspace · briefs & reportsWR·07
WR·07

AI Workspace

From live intelligence to a finished brief.

The output side of the wall. The Workspace turns live intelligence into finished work product — structured briefs and reports you can edit, download as Google Docs, DOCX or PDF, or send straight to a client's inbox by email.

Use case The Friday report writes itself.

Every Friday the account director exports the week's brief as a branded PDF and emails it to the client before the client thinks to ask. Twenty minutes of review, zero hours of assembly.

War Room Presentations — generated branded deck built from clipped intelligence
Presentations · clip → deckWR·08
WR·08

Presentations

A board-ready deck in ninety seconds.

Paste clipped intelligence into the generator and War Room builds a full twelve-page presentation — your logo, your design, every chart sourced from the live index. Not a template with gaps: a finished, editable deck.

Use case Four hours of analyst work. Ten minutes, start to finish.

Before the board meeting, the director clips the week's sharpest intelligence, pastes it in, and walks out with the deck — the work that used to eat an afternoon, done before the coffee's cold.

War Room Talk & Chat — voice orb with a grounded conversation thread: SWOT intelligence, prediction-market shifts
Talk & Chat · ask by voice or text, any pageWR·09
WR·09

Talk & Chat

Ask out loud — or in the box.

An assistant that already knows the campaign. Ask by voice or text on any page — grounded in your workspace, no pasting context, no re-explaining. Talk reads the answer back for operators whose hands are full; Chat threads it, cites it, and turns it into a deliverable.

Use case Hands on the wheel — or on the keyboard.

Driving between canvass launches, the candidate asks out loud: "What's the latest on the factory closure?" The orb answers from the live index — sources, sentiment and all — before the next exit. Back at the desk, the analyst asks the same brain in Chat and walks out with the answer cited, threaded and exported.

War Room switchboard — every campaign in its own sealed tenant
War Room control plane — MCP connections, encrypted tokens, two-factor authentication
Switchboard + control plane · sealed, encrypted, 2FAWR·10
WR·10

Multi-Tenant

Your campaign, sealed.

Every client and campaign runs in its own isolated tenant. Nothing crosses — not data, not memory, not alerts. MCP connections pull from the tools your team already uses — tokens encrypted, never visible to model context — and 2FA on every seat that touches client data. Trust is the architecture, not a promise.

Use case Five campaigns. Five sealed rooms.

An agency runs a senate race, a mayor's race and three ballot measures. Each gets its own memory, its own alerts, its own agents. The senate race never sees the mayor's data — that's not a setting, it's the architecture.

War Room Operations Center — daily LLM cost trend with spike-day detection
War Room admin — campaign configuration with perspective lock, keyword groups and trackers
Operations + admin · costs, health, provisioningWR·11
WR·11

Operations Center

Every token accounted for.

Platform-level monitoring: daily LLM spend with spike detection, health, security, usage, growth, churn — the same discipline a fund runs on its money, run on its intelligence. And the admin plane sits right beside it: provision tenants, lock the campaign's perspective, wire keyword groups and trackers. One panel between a signed contract and a live war room.

Use case The $356 day is flagged before the invoice lands.

A daily cost window exceeds 1.5× the trailing average and turns red on the chart. Nobody discovers spend on the invoice — the platform watches its own economics the way it watches the race.

War Room AI design system — metric tiles, SWOT quadrants, alert rows
War Room AI design system — chat input states and top navigation specs
Design system · data, inputs, navigationWR·12
WR·12

Design System

Designed, not defaulted.

Before a line of code: a full design system — metric tiles, SWOT quadrants, alert rows, chat input, navigation — authored by the same operator who wrote the backend. The product looks like it was designed because it was.

Use case Six hats. One operator.

Engineering, design, marketing — War Room was built by one person wearing all of them. No hand-offs, no telephone game between the person who designs it and the person who ships it.

03 · The Proof4 months

One operator. Production grade.

580K lines of code in 4 months, part-time
≈4× the output of a two-dev AI agency
700+ shipped improvements
6 AI agents running continuously

Built like production software, not a demo: a live knowledge index the agents read against, long-term memory that carries across sessions, a hardened action layer, and three environments with staged promotion. The boring disciplines are the product.

Designed, engineered & shipped by Roderic Andrews — Product Designer & Design Engineer

04 · The Pitchhire the operator

Hire the operator.

War Room is the evidence. Your product is the point.

Everything on this page — the engineering, the design, the positioning — was built by one operator. That's the offer: a product designer who writes the code and wires the growth. No hand-offs, no telephone game, no agency margin.

Design · Engineering · Marketing · Intelligence · Video · Growth

War Room public product page — positioning and marketing copy by the same operator
The marketing hat · public product page05

Book the walkthrough.

Thirty minutes. Every agent. A live tenant.

I'll walk you through every agent — live, with your questions.