An introduction · Not a pitch

I build integrity infrastructure
for AI in regulated industries.

My name is Jeremiah Hearne, founder of Ethereal Connections Co. This page is the short version of who I am, what I build, and which industries I'm currently set up to serve. Every claim on this page is something you can click and verify in the next thirty seconds.

Sovereign hardware stack Five-layer integrity model Walk-away-complete engagements
01 · Who I Am

A hybrid systems architect, not a marketer.

I came to AI infrastructure through buildings, not through software. Most of the people selling AI compliance services have never been on a job site, never been in a NICU, never managed a ninety-million-dollar construction project. I have. That background shapes how I think about what "infrastructure" actually means and why most of what's marketed as compliance is theater.

Background

Where the work comes from

  • NICU and Fire/EMT — first careers, where the cost of a bad audit trail is measured in lives
  • B.S. Architecture, Texas Tech — trained to think in load paths, structural redundancy, and lifecycle
  • NYC Tier-1 PM on $85M SAYA and $45M SKYHOUSE — large-scale construction project management
  • Ardezen — fabrication founder, hands-on with the gap between design intent and built reality
  • OSHA 30 · CDL-A HAZMAT
What ECCO is

Integrity infrastructure, built by the operator, kept close to the artifact.

ECCO is the Inspiration Engine — a transmission mechanism for a doctrine the world needs. The company is the carrier wave. The doctrine is the signal. Every output ECCO ships — the AI Scanner, the AI Library, the Integrity Blueprint, the Patron Toolkit, the Sanctioned Pipeline, the Autonomous Bridge, the Provenance Architecture — is a domain-specific output of that engine.

The whole operation runs on a sovereign five-node hardware stack I built and maintain myself. That's not a limitation — it's the architecture. Integrity infrastructure has to be cheap, public, and copyable, and the only way to keep it that way is to keep the operator close to the artifact. When you talk to ECCO, you talk to me. When something gets built, I built it. When something breaks, I fix it.

02 · What I Build

Four commissioned tiers. Each one is complete on its own.

Every engagement is walk-away-complete. You're not entering a funnel. You're commissioning a discrete piece of work with a defined scope, a defined deliverable, and a defined end. If we work together once and you never call me again, that's a successful engagement.

Tier 01 · Wedge
The Sanctioned Pipeline
$750 · Flat · Two weeks
Substance →
Tier 02 · Assessment
The Patron Toolkit
$7,500 · Flat · One deliverable
Substance →
Tier 03 · Integration
The Autonomous Bridge
From $3,500 setup · $500/mo
Substance →
Tier 04 · Architecture
The Provenance Architecture
Custom · Scoped per engagement
Substance →
03 · Industries I Serve

Forty regulated verticals. Three worked examples.

Every commission begins with the regulatory homework for your specific vertical — statute review, enforcement posture, compliance gap map, built artifacts that survive an audit. The three panels below are worked examples of that depth, fully verifiable — click any citation. ECCO is set up to commission across all 40 regulated-industry verticals the AI Scanner maps. If yours isn't among the three shown, that changes nothing about what you receive: same forensic depth, authored to your industry. Tell me your vertical in the contact form below.

Insurance
AEC · Construction
Real Estate

Independent insurance agencies and carriers in Colorado

Active regulatory pressure · Dated deadline

The pressure

Colorado's Division of Insurance amended Regulation 10-1-1 in October 2025. The amendment expanded the regulation's scope from life insurers only to also include private passenger auto insurers and health benefit plan insurers. The compliance deadline for the new lines is July 1, 2026, and a compliance report is due to the Division by that date and annually thereafter.

Source: Colorado DOI Notice of Adoption · Faegre Drinker analysis

What the regulation actually requires

A risk-based governance framework for ECDIS (external consumer data and information sources), board oversight of the framework, senior management accountability, a cross-functional governance group, written policies for the design and ongoing monitoring of algorithms and predictive models, an inventory of every system in use with version control, documented testing methodology for unfair discrimination, model drift monitoring, third-party vendor selection process, and a documented annual review. The compliance report itself is short. The work behind it is not.

The other pressure most agencies haven't connected

Since February 2025, AT&T, Verizon, and T-Mobile have been blocking — not throttling, blocking — 100% of unregistered A2P 10DLC text traffic. Most agencies have noticed deliverability degradation and blamed their CRM, their phone provider, or their leads. The actual cause is unregistered or misregistered traffic getting silently dropped at the carrier level. Fines for non-compliance can reach $10,000 per violation.

Source: Apten 2026 A2P 10DLC compliance guide

What I can do

For Reg 10-1-1: Tier 3 or Tier 4 — I build the documentation framework, the ECDIS inventory, the testing methodology, the vendor selection process. Not a SaaS. Not a checklist tool. Actual built artifacts that survive an audit. For A2P 10DLC: Tier 1 — flat $750, brand and campaign registration cleanup, usually delivered in under two weeks once I have the EIN and platform credentials.

One important clarification: Colorado SB24-205 (the broader Colorado AI Act) does not apply to insurers. Insurers are explicitly exempt because they're already covered by Regulation 10-1-1. If a vendor is pitching you SB24-205 compliance, they haven't read the carve-outs.

Architecture, engineering, construction, and fabrication firms

Emerging contractual and litigation exposure · Where my background runs deepest

Why this vertical

I came up in this industry. Texas Tech architecture degree. NYC Tier-1 project management on real $85M and $45M construction projects. Founded a fabrication shop. I've been on the job sites where document chaos costs real money and bad audit trails turn into RFI nightmares, change order disputes, and litigation exposure. Most people selling AI tools to AEC firms have never opened a set of construction documents in their lives.

The pressure

AEC firms are adopting AI tools fast — automated takeoff (Togal.ai), photo documentation and progress tracking (OpenSpace, DroneDeploy), contract review (Document Crunch), generative design tools, BIM-integrated copilots — and almost none of these deployments have governance infrastructure underneath them. When a discovery request comes in during litigation, when a client asks "how was this estimate generated," when an insurer wants to know what software touched a critical decision, the answer is usually a shrug.

What the construction legal community is now saying

This is not an abstract concern. ConsensusDocs — the construction-industry standard contract publisher — has published two articles in the past year flagging the exact exposure. The July 2025 piece, The Hidden Risks of Using AI on Construction Projects, walks through the liability, IP, and documentation risks of AI adoption without governance. The February 2026 follow-up, When Your Scheduler Hallucinates: Managing AI Risk on the Job Site, by Jones Walker LLP partner Jason Loring, introduces the concept of AI as a "silent witness" in construction disputes: courts and arbitrators are beginning to recognize that if your predictive tool flagged a problem in real time and a supervisor ignored it, that may support a "superior knowledge" argument under the Spearin Doctrine and erode a contractor's disclosure-duty defense. Loring is explicit that standard ConsensusDocs 200 and AIA forms "weren't drafted with the assumption that AI might be continuously monitoring conditions and generating alerts that may or may not be reviewed by qualified personnel."

Source: ConsensusDocs · The Hidden Risks of Using AI on Construction Projects (July 10, 2025) · ConsensusDocs · When Your Scheduler Hallucinates (February 16, 2026)

The state-law backstop most AEC firms haven't connected

Colorado's SB24-205 — the Colorado Artificial Intelligence Act — takes effect June 30, 2026 after a five-month delay enacted by SB25B-004 in August 2025. The statute applies to any "deployer" doing business in Colorado that uses a high-risk AI system as a substantial factor in a "consequential decision" affecting a consumer. Employment is on the list of consequential decisions. For an AEC firm, that means AI-assisted hiring, subcontractor prequalification scoring that affects who gets work, safety-monitoring systems that influence employment action, and algorithmic tools that touch worker-scoring workflows are all potentially in scope. A small-deployer exemption exists for firms under 50 full-time employees, but only under three conditions: no training the AI on the firm's own data, limiting use to the developer's disclosed purposes, and passing through the developer's impact assessment to affected consumers. Customizing an off-the-shelf model with your own project history kills the exemption.

Source: Colorado General Assembly SB24-205 · Baker Botts on the SB25B-004 delay to June 30, 2026

What I build for AEC

The Enterprise Integrity Stack (Tier 4) is purpose-built for this. It's the layer underneath OpenSpace, Togal.ai, and Document Crunch — the audit trail, the version control, the access logging, the vendor governance, the documentation that ties an AI-generated artifact back to the human decision it informed. The kind of infrastructure a litigator can't tear apart in deposition because every action has a timestamp, an actor, and a reason.

What I can offer that pure-software vendors can't

Operational fluency. I can walk a job site. I can read a set of contract documents. I can sit in an OAC meeting and not need translation. The AI infrastructure I build is shaped by the actual workflows of the people who'll use it, not by a software engineer's assumption about what a project manager does.

Brokerages, property management companies, and real estate technology firms

Emerging regulatory pressure · Wide addressable market · Low literacy

The pressure

Real estate is in the middle of a fast, unregulated AI adoption curve. Tenant screening tools that use algorithmic scoring. AI-assisted lead routing and lead scoring. Automated valuation models. Generative listing copy. Chatbot-driven prospect intake. Almost none of these deployments have audit infrastructure, and almost all of them touch consumer data in ways that have fair housing, fair lending, and consumer protection implications.

What's actually coming

Federal fair housing enforcement is increasingly looking at algorithmic decision-making in tenant screening and mortgage underwriting. State-level AI regulations (Colorado's SB24-205, effective June 30, 2026, and similar laws in other states) treat housing as a "high-risk" use case requiring documented impact assessments, consumer disclosures, and governance frameworks for AI systems that influence consequential decisions about housing access. Colorado SB24-205 in particular applies to deployers regardless of where they're headquartered, as long as the AI touches Colorado consumers.

Source: Colorado General Assembly SB24-205 · Baker Botts on the deadline delay to June 30, 2026

The small-deployer exemption trap

SB24-205 contains a conditional exemption for deployers with fewer than 50 full-time employees — which covers most independent brokerages and small property management companies. But the exemption is narrower than most operators realize. To qualify, a firm must meet all three of these conditions at the same time: (1) it does not use its own data to train or fine-tune the high-risk AI system; (2) it uses the AI system only for the uses the developer disclosed; and (3) it passes through the developer's impact assessment to consumers. Most brokerages that have configured their lead-scoring tools on their own CRM history, customized tenant-screening rules with their own applicant data, or fine-tuned a valuation model on their own transaction records have already killed the exemption and don't know it. The under-50 number is not a safe harbor by itself.

Source: statute text at C.R.S. § 6-1-1703(6) · confirmed by IAPP's Colorado AI Act summary and TrustArc's compliance guide

What I build for real estate

The infrastructure that proves you exercised reasonable care. The impact assessment template, the consumer disclosure language, the governance framework, the bias testing methodology, the documented vendor selection process. Same five-layer architecture as the insurance and AEC builds, scoped to the specific consequential decisions your operation makes. Tier 3 for most brokerages and property managers; Tier 4 for proptech companies building their own algorithms.

The honest framing

This is the vertical with the lowest current literacy and the largest addressable market. Most real estate operators don't yet know they have exposure. The opportunity is to be the firm that gets ahead of the regulatory wave instead of the one that scrambles after the first enforcement action. If you're reading this and thinking "we should probably be doing this," you're probably right.

Your vertical, diagnosed first

The AI Scanner maps regulatory pressure across all 40 verticals in sixty seconds.

The three panels above are worked examples. ECCO commissions across every vertical the Scanner maps — from reinsurance, capital markets, and pharma to legal, healthcare, education, government, defense, proptech, and the rest of the forty. Run the Scanner first. Bring the result to the contact form. Commission scoping starts from evidence, not self-assessment.

Take the AI Scanner →
04 · How I Work

No funnel. No pressure. No surprises.

The first conversation is always discovery, not selling. Twenty minutes. I ask four questions about your operation, your stack, and what you're actually trying to solve. If we're not a fit, I'll tell you in those twenty minutes and recommend who to call instead. If we are a fit, you'll get a written scope before any money moves.

Every engagement is walk-away-complete. Tier 1 is a discrete two-week fix with a defined deliverable. Tier 2 is a tailored assessment that stands on its own or serves as the on-ramp to deeper work. Tier 3 is a defined buildout plus an optional ongoing layer you can cancel at any time. Tier 4 is custom-scoped with a written milestone schedule. There is no point in any engagement where the only way out is to spend more money.

I take on a small number of engagements at a time. The hardware stack is sovereign and the operator is one person — me. That's the constraint and it's also the value. You're not getting a junior consultant assigned to your account.

The engagements I seek are the ones that expand what's possible. The Sanctioned Pipeline exists because A2P 10DLC failures silently kill legitimate business communications — it's necessary grind work and I'll do it well. But the work I most want to do is Tier 3 and Tier 4: building autonomous bridges and provenance architectures with operators trying to stand up something new that needs an integrity layer to do it honestly. If you're a kindred spirit with a hard problem, open the conversation.

I publish my methodology. The AI Library, the AI Scanner, the Integrity Blueprint, the Patron Toolkit — every one lives at the subdomains linked in the nav. If you want to know how I think before you talk to me, the materials are there. Read them. They are not gated. They are not lead magnets. They are the work.

05 · What Already Exists

Click anything below. Verify in real time.

I don't have public client case studies yet — I'm in the early phase of insurance-vertical work and the doctrine I work under (see Section 06 below) means I don't fabricate testimonials or anonymize references in ways that obscure the truth. Instead, here's what ECCO has actually built and shipped. Every link is live. Every link is something you can verify in the next thirty seconds.

Free diagnostic tool
AI Scanner
A 60-second, 5-question integrity diagnostic spanning 40 regulated-industry verticals. No email gate. Results delivered immediately. Built and live since April 2026.
scan.etherealconnectionsco.com →
Public knowledge surface
AI Library
152+ AI tools indexed across the categories that matter for regulated-industry deployments. Reference-grade. Updated continuously. Free, public, no gate.
library.etherealconnectionsco.com →
10-page document
The Integrity Blueprint
The framework I use when I sit down with an organization to figure out what's actually broken. Compliance gap map, provenance architecture, the five-layer stack model, the shadow audit protocol, the regulatory calendar. PDF, downloadable, no fluff.
Request the Blueprint with your context →
Mid-tier paid offering
The Patron Toolkit
A discrete one-time engagement for organizations that need a tailored audit and remediation plan but aren't ready for a full Tier 4 buildout. $7,500 flat. Walk-away-complete.
toolkit.etherealconnectionsco.com →
06 · How We Work Together

Mutual visibility. Mutual advocacy. The doctrine.

Most consulting firms hide their clients. Confidentiality clauses. NDAs. "We can't say who." That posture is built around the assumption that your competitive advantage comes from secrecy. ECCO is built on the opposite assumption: transparency is immunity. The more publicly the work is named, the harder it is to undermine.

So here's the offer, plainly stated. Part of working with ECCO is mutual visibility. When we engage, I ask for permission to publicly name your organization as a working client. In exchange, ECCO advocates for your work — references it in the Transparency Project, includes it in the public portfolio, and treats your success as part of the doctrine's spread. Your visibility increases. Mine increases. The work we did together becomes evidence that integrity infrastructure is real and buildable. Win-win, in the actual sense of the term.

There are honest carve-outs. If your industry has legitimate regulatory or competitive constraints on being publicly named, we can scope the visibility differently — named privately to other prospects only, anonymized case study, or fully confidential with a written agreement. The point isn't to force exposure. The point is to make the default *visibility* instead of the default *secrecy*, and to be transparent about that from the first conversation rather than negotiating it after the engagement starts.

Whatever shape the visibility takes, it's named in the discovery conversation, named in the scope document, and named in the contract. There are no hidden costs. The transparency starts at the door.

07 · Contact

Two paths. Whichever is easier.

No calendar widget. No fifteen-field form. Just enough to know who you are, what industry you're in, and what you're actually trying to solve. I'll respond personally within one business day.

Send a direct note

The form below pipes directly to me. I read every submission myself.

Connect on LinkedIn

If you'd rather start the conversation on LinkedIn, my profile is the right place. I respond to DMs personally. A short note about what you're trying to solve is all I need.

If neither of these channels works for you, the main site at etherealconnectionsco.com has additional contact paths.

Ethereal Connections Co.