Velzyx engineers custom AI and operations systems across multiple categories — from voice agents to route calculation engines to complete front-office platforms.

02 — What "portfolio" means here

Not case studies. Not screenshots. Systems that run.

Most AI companies have a portfolio that is, in practice, a marketing surface — logos collected from pilots, screenshots from product demonstrations, and a vague claim of impact attached to each one. Velzyx is structured differently, and the portfolio reflects that. Each entry on this page is an AI system Velzyx engineered, owns, and operates as production infrastructure for the customers who use it. There is no system on this page that Velzyx is not currently responsible for in the field.

That structural choice matters because it changes what the portfolio is for. The portfolio is not a sales artifact. It is the body of work that the engineering practice can be evaluated against. When a new operator asks Velzyx to build something, the most honest answer to "what have you done before?" is the list below, with each entry written in language that matches what the system actually does. The portfolio is also how the engineering practice itself stays sharp: every new system in production raises the floor of what the underlying primitives can do for the next one.

The five systems on this page sit on top of the same engineering layer. The platform page describes that layer in detail. Each system is configured per operator on top of those primitives. The engagement process behind each one is described on the how it works page.

03 — The thread that runs through all five

Engineered around a workflow, owned end-to-end, deployed for real.

The five systems serve different industries and different kinds of operators. A dental front office, a medical spa, and a commercial real estate underwriter do not share a vocabulary, let alone a software stack. But the engineering posture behind each Velzyx system is consistent: the system is engineered around an operational workflow that already exists, it is owned end-to-end by the team that built it, and it ships into production from the first real interaction rather than living indefinitely in a pilot environment.

That consistency is what makes the portfolio readable as a single body of work. The dental engagements taught the engineering team how to handle live voice traffic, structured intake, and the operational realities of healthcare-adjacent data. The med spa build taught the same runtime how to hold deposits, run medical pre-qual, and pair Voice AI with the website and ads that create the consult. The real estate engagements taught the team how to handle the long lifecycle of an inquiry through a deal. The commercial underwriting engagement taught the team how to handle dense financial workflows where the system's outputs feed an analyst's judgment. Each system added something to the primitives that the next one inherits.

Operators evaluating Velzyx for a new engagement should read the product pages below as a way of triangulating the engineering posture, even if their industry is not on the list. The pattern of work is the same. The configuration is what changes.

05 — Dental voice deployment, in context

The system that taught the platform how to listen.

Our dental voice deployment is the oldest of the Velzyx voice systems and the one that anchors the platform's voice runtime. The deployment is, on the surface, an AI receptionist for dental practices — the system that picks up the phone when the front desk cannot, verifies insurance, books appointments across providers, runs emergency triage, and follows up on overdue hygiene recall. Underneath the surface, it is the canonical configuration of the Velzyx platform on top of the workflow engine. Every dental practice on the deployment is a separately-configured instance, with the provider names, insurance acceptance, recall language, and emergency protocols of that specific operator.

The portfolio extends beyond dental. The same voice runtime is configured for appointment-driven verticals, each one a distinct deployment with its own intake flow, its own scheduling logic, and its own escalation rules. The dental engagements remain a large live segment; Aria Medspa is the aesthetics build on the same platform. Read the full product page on the dental deployment page or med spa.

A published case study on a specific dental deployment is in preparation and will follow once the operator approves the write-up.

06 — Aria Medspa, in context

The system that taught the platform how to hold a consult.

Aria Medspa is the Velzyx sister brand for medical spas. On the surface it is Voice AI on the practice line: after-hours injectables, laser, and body consults, deposits, medical pre-qual, memberships, and write-back to Boulevard, Zenoti, or a calendar if the spa has no PMS yet. Underneath it is the same voice runtime as Aria Dental, configured for aesthetic intake, cancellation policy, and the 12–16 week tox curve. Marketing sits on the same team — website, local SEO, ads, map pack — so ranking does not donate the consult to voicemail.

This is not a patient clinic. Visitors to the product site are spa owners and operators. The industry page on Velzyx is the parent-company view; the product lives on Aria Medspa. Read the med spa deployment page or open the product site.

07 — AgentCentric, in context

The system that taught the platform how to host a brand.

AgentCentric is the Velzyx web platform for luxury real estate. The deployment is, in shape, a set of engineered websites with custom IDX surfaces for individual agents and small teams, but the engineering underneath is heavier than the surface suggests. Each AgentCentric brand has its own editorial template, its own listing flow, its own inquiry lifecycle, and its own integration set against the systems the agent or team already runs. The platform supports brands without forcing them into a uniform shape, which is what makes it usable for operators whose brand identity is part of their business model.

Active brands on AgentCentric today include Solis Muscolino, Home with Miriam, Templer First Homes, and PropNation, with additional agent deployments shipping on a regular cadence. The engineering work on each one is meaningful — AgentCentric is not a template marketplace where agents pick from a dropdown. It is an engineered platform that Velzyx configures per agent, with the same discipline that runs through our dental AI receptionist. Read the full product page on the AgentCentric site.

The reason AgentCentric belongs in the portfolio alongside an AI voice product is that the same engineering primitives sit underneath both. The workflow engine that handles a call for our dental AI receptionist product is the same workflow engine that handles a listing inquiry for AgentCentric. The integration layer that writes to a practice management system is the same integration layer that writes to a real estate CRM. The platform is one body of work.

08 — AnalytixCRE, in context

The system that taught the platform how to underwrite.

AnalytixCRE is the Velzyx underwriting engine for commercial real estate. The deployment takes a deal — hotel, multifamily, office, retail — and returns a full underwriting analysis in minutes. The system is live with commercial real estate investors who run it against real deals before committee. The engagement is scoped per client.

The engineering work behind AnalytixCRE is in the workflow engine and the observability stack. Underwriting is a dense workflow with a long sequence of decisions, each one feeding the next, and the system has to be legible to an analyst who is using the output to inform a real capital allocation decision. The observability layer makes every decision the system makes visible to the analyst, so the engine functions as an assistant to the underwriter rather than a black box that hands back a number. The engagement is structured so the analyst always remains the decision-maker. Read the full product page on the AnalytixCRE site.

AnalytixCRE is among the younger systems and the one whose surface area is growing fastest as new deal types are added. The same engagement pattern that shaped Aria Dental, Aria Medspa, and AgentCentric runs through it: a specific operational problem in a specific industry, engineered around an existing workflow, deployed in production, and owned end-to-end by the team that built it.

09 — What comes next

More industries, more configurations, more work in the field.

The portfolio will grow in two directions. The first is more configurations of the existing systems. our dental AI receptionist's twenty-vertical roadmap is well underway, with new appointment-driven industries being added on a regular cadence. AgentCentric continues to bring on agent brands. AnalytixCRE is expanding the deal types it underwrites. The platform absorbs each new configuration, which means the engineering team can take on additional operators without diluting the depth of the work on any given engagement.

The second direction is new systems in adjacent industries where the same engineering posture works. The criterion is consistent: a focused operational problem, a real workflow to design around, and an operator willing to run the system in production from the first day. Where those conditions hold, the next Velzyx system can be built. Where they do not, the engagement is the wrong shape and the engineering team will say so during discovery.

Operators considering an engagement should start with the how it works page, which walks through the engagement process in detail, and continue to the case studies for closer reads on specific deployments. From there, the right next step is a discovery conversation. The system that follows from it is what eventually shows up on this page.