We got tired of AI work that doesn't stick.
So we built a firm that ships systems instead of strategies.
Who We Are
Perpetual Quest is an AI consultancy built for operators, by operators. We work with mid-market companies — typically $10M to $30M in annual revenue — and, in our home service division, with smaller and small-market companies that still have real inbound, a crew to dispatch, and jobs leaking between the phone and the truck. In both cases we deploy AI systems that run real operational functions on bounded workflows. Not demos. Not prototypes. Production systems with defined authority, humans owning the exceptions, and a team that uses them every day.
Our work is organized around three programs: Signal (diagnose and prioritize), Forge (build, evaluate, and deploy — with training embedded), and Run (managed governance and operational assurance, with Optimize as a phase inside Run). The programs are designed to connect — most of our clients move through more than one of them over time, each building on what came before. After they own the system, many continue into Run for managed assurance — evaluation, governance, optimization, and expansion — by choice, not lock-in.
We build on OpenClaw and Hermes and use the simplest architecture that meets each workflow's reliability, control, and performance requirements. You own the deployed system, configuration, runbooks, and documentation at handoff, and you can operate it without us. These are the same systems we run inside our own portfolio of businesses. We haven't built something we wouldn't use ourselves.
Who We Work With
Mid-market is still the core of Signal, Forge, and Run: operators who have enough complexity that a bounded workflow pays for the engagement, and enough leadership bandwidth to own what we deploy.
Home services is a distinct division. We provide the same discipline — diagnose, deploy, govern — as seven growth systems for HVAC, plumbing, electrical, roofing, restoration, and similar trades. That includes small-market companies: owner-operated shops and multi-crew businesses in smaller cities and towns, not only metro groups at the top of the revenue band. If you have a service area, a CRM or job board, and leads that die unanswered, you are in the market we built that work for.
Mid-market operators
Typically $10M–$30M. Strategy, implementation, and managed governance on a named workflow — Signal, Forge, and Run.
Home service division
Small-market and growing trades companies. Reactivation, reviews, ads, lead nurture, receptionist, and coaching — sized to how you actually sell and dispatch.
Home service growth systems →Why We Exist
Here's what kept happening before Perpetual Quest existed: a mid-market company would read about AI, feel the pressure to act, and hire a consultant. The consultant was smart. The consultant knew the space. The consultant delivered a strategy document that accurately described what the company should do. Then the engagement ended.
Six months later: the document was somewhere in a shared drive, the team had gone back to their old workflows, and the company was no further along. The failure wasn't the consultant's intelligence. It was the model. A document is not an outcome. A presentation is not a deployed system. And a consultant who leaves after delivery has no accountability for whether any of it works.
We built Perpetual Quest because the operators we talked to weren't asking for more advice. They were asking for something that worked.
The other thing we noticed: most AI consulting was built around enterprise buyers — Fortune 500 mandates, 18-month engagements, committees to navigate. Mid-market operators — the COO at a $15M services firm, the CTO at a $25M SaaS company — were being sold the same products at a smaller scale, without acknowledgment that their constraints were fundamentally different. The same gap shows up in home services, including small-market shops: the agency pitch assumes a marketing department you do not have, and the enterprise pitch assumes a year you cannot spend. They don't have an AI team to hand things off to. They need something that runs a bounded workflow with defined authority — that a lean internal team can operate, control, and own the exceptions for — without being forced into a vendor they can't exit.
We exist to serve those operators specifically.
Our Model: Deploy, Don't Describe
The difference between Perpetual Quest and most AI firms is what the engagement produces. Most firms produce a deliverable — a roadmap, a report, a set of recommendations — and measure success by client satisfaction with that deliverable. We measure success by whether something is running in production.
Scope is defined before we start.
For Forge, we don't do open-ended time-and-materials projects. Before a Forge engagement begins, you know what we're building, what done looks like, and what the 90-day structure is. Vague scope is how projects fail. We don't take it. (Run is a separate, defined monthly partnership after you own the system.)
We stay through the defined stabilization period.
Forge includes a stabilize-and-transition period in Days 76–90. We don't disappear the day the system goes live. The first weeks of production are where unexpected things surface, so we stay through the defined stabilization window and transition with named owners, runbooks, and operating cadence in place.
Training is embedded, not bolted on.
We don't hand off a system and then schedule a training day. Training is embedded in Forge during rollout and continues as operator enablement inside Run — at a cadence that fits working hours. Adoption is built in, not hoped for.
We validate in our own businesses first.
Perpetual Quest maintains a portfolio of operating businesses where we run the same production agent systems we deploy for clients. When we tell you a system works, we've run it.
How We Validated the Approach
We validated agent ops in our own product business, then shipped the same discipline — sites, lead paths, and follow-up — for home-service operators. Not projections — what is live.
EchoTexting
SMS automation platform · Own portfolio
Agent workflows manage campaign sequencing, response classification, and escalation routing at the volume a human ops team couldn't sustain. This is where our reliability work got stress-tested. A system that holds at 100 messages is different from one that holds at 100,000.
What it produced
Absorbed a workload that would otherwise have required two to three additional ops hires.
What we learned the hard way
Volume reveals edge cases that staging never will. Stabilization after go-live is not optional.
GoulenCo Roofing
Greg Goulenko · South Bay roofing
A reroof-first public site with estimate and inspection paths, Google review discipline, and coastal service-area pages — so homeowners get a clear next step instead of a generic contractor template.
What it produced
Live conversion site at goulenco.com with lead capture and local trust.
What we learned the hard way
Offer clarity beats feature lists — reroof-first copy converts better than a laundry list of every service.
Villareal & Son Roofing
Abel Contreras · High Desert since 1965
Situation-based intake and an instant estimate path for a small-market family shop — leak, storm, replacement, and commercial routes that match how High Desert homeowners actually call.
What it produced
Live funnel at villarealroofing.com — estimate without waiting on a callback.
What we learned the hard way
Small-market shops need intake that works without a marketing department — not a mid-market org chart.
Schulze Plumbing Inc.
South Bay plumbing · Residential & commercial
An estimate-first plumbing site for South Bay homes and businesses — service-organized offers, a three-step path that confirms scope before a number, and license, hours, and reviews stated carefully next to the ask.
What it produced
Live conversion site at schulzeplumbinginc.com with estimate capture and local trust.
What we learned the hard way
Straight estimates beat published bait prices — confirm the real job before you send a number.
Screenshots and full write-ups on Testimonials.
What We Believe
We tell you what won't work.
If your scope isn't achievable, we say so before you sign. If your team isn't ready for a deployment, we say so before you start. We'd rather lose a deal than start an engagement with a misaligned expectation.
We define done before we start.
Not “we'll make progress on AI” — but “a 90-day path to this specific workflow in production, with your team certified to operate it — go-live only when the evidence gates are met.” Vague scope is how projects fail.
We build for your team to own it.
We build every system so your team retains control and can operate it without us. Clients retain Perpetual Quest when they want ongoing evaluation, governance, optimization, incident support, and carefully managed expansion — not because the system is intentionally dependent on us. Lock-in is our failure. Managed assurance is a choice.
We work with companies ready to execute.
We're not the right partner for a company that wants to understand AI in general. We're for companies that have identified a problem, have a team with capacity to participate, and are ready to commit to an outcome.
What we won't promise
ROI in the first four weeks.
A production deployment takes 8–12 weeks. Anyone promising measurable ROI before go-live is selling you the idea of the work, not the work itself.
Zero oversight required.
AI systems need monitoring, drift detection, and periodic recalibration. We'll build you something that runs lean — but “set it and forget it” is not an honest description of any production AI system.
Results without your team's involvement.
The fastest engagements happen when a client-side owner has decision authority and time to participate. If you can't give us that, we can't give you the outcome we're scoping.
Identical results to what we saw in our own portfolio.
Our numbers come from our businesses. Your starting point, team, and workflow are different. The ranges are honest — the specific outcome depends on what we find when we look at your operations.
EXPLORE • EVOLVE • ENDURE
Ready to talk?
The discovery call is 30 minutes. We'll ask about your operations, where you've tried AI before, and what's in the way.
30 minutes. No sales pressure. We'll tell you whether we're the right fit — and if we're not, we'll tell you that too.
Book a Discovery Call