About Glasir
Small on purpose. Senior on purpose.
Glasir is an AI and automation consultancy in Northern Virginia, founded by Nate Daniels. The person who scopes your work is the person who builds it. There’s no bench, and no junior team you meet after you sign.
The name
Glasir is the golden-leaved tree that stands before the doors of Valhalla, the most beautiful tree among gods and men.
It stands at the threshold. That felt about right for the work: the thing you walk past on the way in, doing its job, not asking for attention.
- Based in
- Northern Virginia
- Founded by
- Nate Daniels
- Discipline
- Engineering, not slideware
Point of view
Most AI projects fail because nobody counted the hours.
The pattern is always the same. A tool gets bought because it was impressive in a demo, a pilot starts, and six months later there’s a dashboard nobody opens and an invoice nobody can defend. Nothing was measured going in, so nothing can be judged coming out.
Glasir writes a business case before it writes code. We shadow the actual workflow, count the hours it burns, and put a number next to the fix. Sometimes that number says don’t bother, and we say so. That’s the part that makes the rest of the advice worth having.
The background here is hands-on engineering and homelab infrastructure (building, running, and fixing systems), not management consulting, and it shows up in the work. We talk about where the data lives and who can reach it before we talk about strategy, because the second conversation is worthless without the first.
Glasir is deliberately small, and that’s the point. You get the thing large firms sell hardest and deliver least: the senior person, on your project, for its whole life. No bench, no handoff to a junior team you meet after you’ve signed.
How we work
Three steps, in this order, every time.
The order matters as much as the steps. Look, then price, then build, and never skip straight to the build because the software looked good.
- 01
We look at the week, not the wish list
Thirty minutes to start, then a proper walk through the real workflow: the pad by the phone, the spreadsheet holding it all together, the thing everyone works around. Not what the org chart says happens.
- 02
We write the business case
A ranked backlog with hours, cost, and return next to each item, plus the items we recommend against. You can take that document to another firm, and some clients do. That’s a fair outcome.
- 03
We build it, then we prove it
The first workflow goes to production in weeks. We measure it against the number we wrote down, hand over the documentation, and only then talk about what’s next.
Principles
What we argue about, and why.
These aren’t values on a wall. Each one has cost us a project at some point, which is roughly how you tell a principle from a preference.
Count the hours before you buy the software
Most AI projects fail because nobody did the arithmetic first. If we can’t point at a specific task, the people doing it, and the hours it eats each week, there’s nothing to automate yet. We’d rather find that out in week one than after you’ve signed a year of licences.
We’ll tell you when not to automate
Some work is too rare, too varied, or too cheap to be worth wiring up. A shop that quotes twice a month doesn’t need a quoting engine. It needs an afternoon and a template. Saying so costs us a project and buys us a client who believes the next thing we say.
Build on what you already pay for
Your stack isn’t the problem. n8n, QuickBooks, Jobber, ServiceTitan, Twilio, Stripe, Claude. These already do the work, and what’s missing is the wiring between them. Rip-and-replace is the most expensive way to solve a plumbing problem.
You own it, in your accounts
Workflows run under your credentials, on your subscriptions, with the documentation handed over. If you fire us, everything keeps running. A consultancy that holds your automation hostage isn’t a partner, it’s a subscription with extra steps.
Ship the first thing early
Assessment phases are where projects go to die. We build the highest-return workflow while the process map is still warm, against your real data, so you can judge us on something that runs instead of something we presented.
Report the number we promised, including the misses
Every engagement ends with a measurement against the business case we wrote at the start. When we overestimated a saving, that goes in the report too. The alternative is a slide that says “success” and a client who never checks.
What we build on
Your tools. Your accounts. Your keys.
We don’t sell a platform, so we have no reason to talk you onto one. Almost every build is the tools below, wired together properly and handed over.
- n8n
- Workflow automation: the connective tissue between everything else.
- QuickBooks
- Invoicing, payment reconciliation, and the reporting you already trust.
- Jobber & ServiceTitan
- Field service scheduling, dispatch, and job records.
- Twilio
- Texting and calling that routes leads to a human in seconds.
- Stripe
- Payments, links, and the reminders that chase them.
- Claude
- Grounded assistants and document work, answering from your files with a citation.
Where Glasir CRM came from
We kept stitching the same integrations together (lead to quote to invoice to payment) for one client after another. At some point the honest thing was to build it once and keep it. That’s Glasir CRM. It isn’t a requirement for working with us, and we’ll tell you when switching isn’t worth the disruption.
See Glasir CRMWho we work with
Companies where the manual work is obvious once you look.
If your office manager re-types job sheets into QuickBooks, or the owner builds quotes on a Friday night from memory, we already know the shape of your week.
- HVAC, plumbing, electrical, and roofing shops between 10 and 50 people
- Construction and specialty trade contractors drowning in job paperwork
- Government contractors who need AI that survives a security review
- Professional services firms where the owner still builds the invoices
Who we’re wrong for
If the plan is to cut headcount, we’re the wrong firm. Every shop we meet is short-handed, not overstaffed. If you want a large team, a procurement portal, and a partner logo on the deck, we’re the wrong firm for that too. Small is the offer, not an apology for it.
Talk to the person who’ll build it.
Thirty minutes with Nate. We map where your time actually goes, tell you what’s worth automating and what isn’t, and you leave with a written summary either way.