GEMProspect IntelligenceGlobal Epicenter of Mobility · Detroit Regional Partnership
Demonstration system seeded with public-source sample data. Every company, signal and figure shown is drawn from the cited public source on its profile — no fabricated data. Prepared by LIFEOFMORR LLC for the Detroit Regional Partnership (DRP) / Global Epicenter of Mobility (GEM).

How it works

The delivery method behind this demonstration — the same method LIFEOFMORR LLC proposes to operate for the Detroit Regional Partnership / Global Epicenter of Mobility.

1

Scheduled ingestion — filings first

The system ingests on a fixed schedule from primary public sources: SEC EDGAR filings (8-K, 10-K, S-1 and exhibit press releases), USAspending.gov federal award records, state economic-development announcements, and company press releases and public announcements. Filings lead because they are dated, attributable and auditable — the same standard this demonstration's sample data was built under.

2

Signal extraction

Each document is screened for expansion-class events: new plants, expansions, production starts, capex programs, market entries, major funding rounds, leadership/site-selection hires and ownership changes. Every extracted signal keeps a pointer to the exact source document, its date and its URL.

3

Enrichment — firmographics and contacts

Matched companies are enriched with firmographics (HQ, footprint, headcount band, parent company) and business-development contacts from licensed and public sources, so a profile is actionable the day it is verified — not a research project.

4

AI-drafted profiles, mandatory human verification

AI drafts the profile, the summary and a first-pass fit score. Nothing publishes on the model's word: every profile enters a human verification queue, where an analyst checks each claim against its cited source, corrects the score, and only then releases it to the dashboard. This demonstration shows the post-verification state — including deliberately discounted scores where a signal is old (LG Energy Solution) or carries risk (Wolfspeed).

5

Source citation on every claim

No citation, no claim. If a jobs figure, an investment amount or a date is not in a source an analyst actually read, it does not appear in the product. Profiles carry source name, source date and URL, and alerts reproduce the full citation trail.

6

Scoring — published model, visible rationale

Fit score = signal strength (/25) + North America presence & move (/25) + scale (/25) + recency (/25). The model is fixed and published here; each company's analyst rationale is shown on its profile and on the prospect list, so DRP staff can defend or override any rank.

7

Alerts and DRP-controlled outreach

Verified high-fit signals become alerts in the format shown on the Alerts page: company, signal, Detroit Region relevance, recommended action, sources. Outreach drafts are generated for DRP's convenience — and nothing sends without DRP approval. DRP controls the send, the message and the relationship; the system only prepares the work.

8

Cadence and accountability

Ingestion runs on schedule, the verification queue is worked daily, and the ranked prospect list is refreshed as new signals verify. Every number in the system can be traced to a source document and a verification event — the audit trail is the product.

What this demonstration is — and is not

This demo is a working instance of the product, seeded by hand with 19 real, source-cited company profiles compiled from public sources in November 2026. It is not connected to DRP systems, it holds no private data, and it sends nothing. In production, the same screens are fed by the pipeline above. See the sample alerts and the ranked prospect list for the output standard.