REX Product Brief
General-purpose AI turns sellers into researchers. Citylitics REX does it for them.
ChatGPT, Claude, and Copilot can help answer questions — but only after your team knows what to ask, supplies the right context, verifies the output, and decides whether it matters. Citylitics REX continuously monitors, validates, and connects public infrastructure signals so sellers can focus on the right accounts, projects, and next-best actions.
The DIY reality
$8–10M
Estimated AI project cost including engineering, token usage, data verification, maintenance, and support.
67%
Failure rate for internal AI projects.
2M/week
Documents processed by Citylitics REX AI every week.
The core objection
Why “we’ll just use AI” still leaves the work with your sales team
General-purpose AI can help with research, but it does not continuously monitor your market, know your territory priorities, validate official records, or connect signals to live pursuit opportunities.
Using ChatGPT / Claude / Copilot directly
- Seller identifies what to investigate
- Seller searches for missing information
- Seller explains the market, account, and need
- Seller fact-checks the result
- Seller decides whether the signal matters
Using Citylitics REX
- REX monitors the market continuously
- REX filters signals through the Infrastructure Market Signals Graph
- REX applies customer priorities, products, territories, and accounts
- REX validates records and marks what is not confirmed
- REX connects signals to opportunities and next-best action
Who does the work?
REX shifts the research burden away from sellers
Research task
General-purpose AI
Citylitics REX
Decide what to investigate
Seller identifies the question and initiates the research
REX continuously monitors the entire market every day across 50,000+ cities, utilities, and agencies to deliver actionable insights
Find and organize source material
Seller searches for missing information and sorts incomplete results
REX uses the Infrastructure Market Signals Graph to extract early-stage insights regarding future projects and filter out the noise
Provide context and refine prompts
Seller explains the market, account, and need, then iterates
Customer priorities, products, territories, and accounts are built into the platform
Verify accuracy
Seller fact-checks sources and determines what is reliable
REX applies validation and quality controls, cites official records, and marks what is not confirmed
Determine relevance and when to act
Seller decides whether the signal matters and what to do next
REX connects signals to opportunities and shows where to focus, why it matters, and when to act
2MDocuments discovered weekly
2000+Human-verified pre-RFP leads that users action every week
1–5 yearsEarly visibility into future RFPs
The real cost of doing it yourself
The cost is not just the model call
To replicate procurement intelligence at market scale, teams need token capacity, engineering, data infrastructure, source access, entity resolution, human QA, and ongoing maintenance.
~$1.2M/yr
Token costs alone to have AI agents read the market. Tokens buy comprehension of one document at a time — not a cross-referenced graph, history, or verification.
~$5M/yr
The sustained replication stack: crawlers, proxies, parsers, archival storage, entity resolution, human analysts, and contact enrichment.
2x
Specialist tools succeed twice as often as internal builds. The same research found most enterprise AI pilots do not deliver measurable P&L impact.
MIT — State of AI in Business 2025
What the $5M DIY stack actually includes
A working internal build has to do more than prompt an LLM. It needs the operating stack to collect, process, verify, and maintain procurement intelligence continuously.
~$1.2MTokens
40k docs/month × 100k tokens × 2.5x agentic loop × $10 per 1M blended tokens.
~$1.55MEngineering
Specialized data/scraping, AI/RAG, DevOps, and product management capacity.
~$1.05MHuman QA
Analysts reviewing ambiguous municipal, infrastructure, and compliance records.
~$1.25MData & Infra
B2B registries, proxies, cloud storage, vector clustering, and layout-aware PDF processing.
~$5M/yrTotal cost
You cannot crawl the past.
Pre-RFP signals accumulate one to five years before a solicitation, and many source documents are no longer online. Citylitics holds archival copies past link-death. While a two-year internal build runs, the engine discovers another 200 million documents and ships another 130,000 verified leads. The gap at the end is wider than the gap at the start.
This is not manual research through an AI interface. It is ready-to-use sales intelligence.
The Citylitics REX Intelligence Engine
Working for you, every day
1
Unmatched Deep Crawl
Deep-web crawler captures ~20x more relevant insights from 40,000+ cities, utilities, and agencies that other platforms miss.
2
Proprietary Signal Graph
Purpose-built AI models and signal graph track 31,000+ daily market signals that are out of reach for other platforms.
3
Analyst-Verified Data
40+ in-house data analysts review thousands of records daily so your sales team can act on insights with confidence.
4
Integrated + Personalized
Guided user adoption, workflow development, and CRM integration help maximize ROI. The platform adapts with each user’s territory and opportunity preferences.
Avoid DIY Research Loops | Capture Earlier Signals | Turn Intelligence Into Action
Citylitics REX replaces manual AI research with continuously monitored, source-backed public infrastructure intelligence. Instead of asking sellers to prompt, verify, and interpret results, REX identifies relevant market activity, validates what matters, connects signals to accounts and projects, and enables the team to take the next-best sales action.