AI that skilled people can rely on as readily as their own tools.
Aurifex LLC · AI enablement for skilled work
Gold-standard
AI
for
skilled work.
Built so far for healthcare, learning, medical education, careers, books, music and jewelry. Every product solves a real problem in skilled work, is proven on that work before it ships, and is effortless to use.
--days
--hours
--min
--sec
to Veritas, Q1 2027
Aurifex LLC · 2024
[ 01 ] // Why a gold standard
People use AI.
Few trust it yet.
Most people already use AI at work or at home. Fewer than half are willing to trust it, and most want more say in how it’s used. For skilled work, where an answer has to be right, that gap is the whole problem.
- 66%
- of people use AI regularly
- 46%
- are willing to trust it
- ~6 in 10
- Americans want more control over how AI is used in their lives
- 66%
- of AI users rely on its output without checking it
University of Melbourne & KPMG, Trust, attitudes and use of AI: a global study, 2025 (47 countries) · Pew Research Center, How Americans view AI, Sept 2025
Build AI that solves real problems in skilled work, prove it on that work, make it effortless to use, and hold every product to the same gold standard.
Aurifex is Latin for goldsmith, and our mark is gold itself: Au, element 79. In medicine and science, the gold standard is the best available test, the one everything else is measured against. It doesn’t mean perfect. It means measured, compared and shown.
That is the standard we hold every product to, in any industry. It has to solve a real problem for people who are skilled at what they do. It has to be proven on that work before it ships, with every number saying how it was measured. And it has to be effortless enough that the expertise, not the software, gets the attention.
Gold-standard AI isn’t a claim. It’s a test every Aurifex product has to pass.
- →Real problems in skilled work, chosen before the technology
- →Proven on the real work it was built for, before release
- →Effortless fits the way skilled people already work
[ 02 ] // Our products
Gold-standard AI,
at work
Every Aurifex product applies the same standard to a different kind of skilled work. One is live, and six launch through 2027. Each has its own page with the product, the evidence and where it stands.
Aurifex HealthcareLive
VeritasLaunching Q1 2027
NightdriveLaunching Q2 2027
BookloftLaunching Q3 2027
RanvierLaunching Q3 2027
TruefillLaunching Q3 2027
GoldsmithLaunching Q4 2027
UW HealthProduct · 2023–2025
CODAmarketProduct · 2021–2023
GallifyProduct · 2016–2021
Live
01
HealthcareEnterprise
Aurifex Healthcare
Gets hospital patients to the right next place, sooner.
Predicts each patient’s discharge date and next place of care, flags likely inpatient rehab candidates in the Epic worklist, and drafts the referral with a cited source for every line. A case manager approves every referral before it’s sent.
- 85%+
- discharge dates within one day
- 0.91
- AUROC, rehab-candidate flag
- 45→10
- minutes per rehab referral
Gold standard hereEvery referral cites its source note, and a case manager approves it before it’s sent.
Live
Launching in 2027In order of launch
02
LearningConsumer
Veritas
Master what you read.
Reads your textbook, notes or any document to you in your own voice, every word highlighted. Then it checks what you understood and brings back what you missed, until you’ve mastered it.
- Q1 2027
- public launch
- 2,500+
- on the waitlist
- 97%
- source-grounded answers
Gold standard hereEvery answer comes from your own book, and every question type passed 400+ real chapters before launch.
Launching Q1 2027
03
MusicConsumer
Nightdrive
Any song, slowed + reverb, in seconds.
An app that turns any song file you own into slowed + reverb, sped up, nightcore, phonk, lo-fi or 8D versions. Describe the sound you want in Highbeam and it builds the effect chain, or take every control in the Advanced studio.
- Q2 2027
- public launch
- 63/63
- end-to-end checks passing on the current build
- 26
- researched presets built in
Gold standard hereEvery change can be undone, originals are never touched, and files never leave your device.
Launching Q2 2027
04
BooksConsumer
Bookloft
Your personal librarian.
A private home for the books you own. It catalogs your shelves, answers “do I own this?” in a second, and Octavia, its AI librarian, suggests what to read and buy next. Whole-bookcase scanning arrives at launch.
- Q3 2027
- public launch
- <10 min
- target to catalog a 300-book bookcase from one video
- <1 s
- target to answer “do I own this?”, offline
Gold standard hereEvery match carries a confidence label, anything unsure goes to you to check, and your catalog stays in files you control.
Launching Q3 2027
05
LearningMedical education
Ranvier
A clear daily path to your MCAT score.
An MCAT prep platform that builds a path to your target score and coaches you along it: a 45-minute adaptive diagnostic, a daily path that leaps what you already know, and Node, an AI tutor that builds sessions around your weak spots.
- Q3 2027
- public launch
- +10 pts
- target average score gain
- ±3 pts
- target projection accuracy
Gold standard hereEvery explanation points to the passage, and projected scores will be checked against real results.
Launching Q3 2027
06
CareersConsumer
Truefill
Apply with confidence.
Truefill finds the roles that fit you, tailors your resume for each, fills every application from your real background and reads every answer back. Anything it can’t confirm is held for you, and nothing is sent until you approve.
- Q3 2027
- public launch
- 96%
- of 5,121 reviewed fields needed no change
- 0.8%
- of fields answered wrong; the rest of the edits were blanks, never guesses
Gold standard hereEvery answer is read back against the page, and nothing is sent without your approval.
Launching Q3 2027
07
JewelryMarketplace
Goldsmith
Custom jewelry, without the markup.
Describe a ring in plain words and get manufacturable 3D designs built from licensed parametric CAD templates, then request comparable quotes from verified manufacturers on one design file.
- Q4 2027
- public launch
- ~1 s
- to the first manufacturable design
- 98%+
- of designs manufacturable after automatic fixes
Gold standard hereEvery design passes manufacturing checks before a quote, and every quote is on the same design file.
Launching Q4 2027
Shipped and ScaledEarlier work
Products Edward Profio helped build before today’s line-up, each shipped to real users and scaled. None is counted in the outlook.
08
ProductHealthcare
UW Health
GenAI inside Epic, from pilot to system-wide.
At Wisconsin’s largest health system, Edward led the AI/ML roadmap and launched Epic-integrated GenAI agents that draft patient-message replies, shift summaries and routine documentation for clinicians to review.
- 42,000+
- estimated clinician hours returned in a 500-clinician pilot’s first year
- 93%
- clinician-judged task success
- Org-wide
- scaled after the pilot year
Product · 2023–2025
09
ProductPublic art
CODAmarket
A marketplace for public-art artists, with CODAworx.
Edward was recruited into the founding team and became a co-founder of CODAmarket, which launched in 2021 to CODAworx’s network of 17,000+ artists. He owned the product, its pricing and artist support.
- +40%
- click-through on recommended artworks
- +30%
- repeat purchases
- 2×
- 90-day user retention
Product · 2021–2023
10
ProductDigital art
Gallify
A gallery for digital art, on your own wall in AR.
An AI-powered iOS marketplace for digital art, with AR viewing. Edward joined early and became Chief Product Officer and a co-founder; in 2021 the team took what it had built into public art with CODAworx.
- 7,500+
- users: 2,500+ artists and 5,000+ buyers
- ~15%
- growth month over month
- +30%
- click-through from recommendations
Product · 2016–2021
[ 03 ] // Our approach
How a product
earns the standard
The same three steps behind every Aurifex product, adapted to what each kind of work needs.
- 01
Start with a real problem
Find a problem people live with every day in their work, and decide how success will be measured before building anything. The problem comes first; the technology follows.
- 02
Test it on real work
Build the test from real examples of the work, set the bar before building starts, and hold every release to it. A change that misses the bar doesn’t ship, and results are shared with how they were measured.
- 03
Make it effortless
Put it where skilled people already work, cut every step that isn’t needed, and design it so the expertise gets the attention, not the software. If it’s hard to use, it isn’t finished.
[ 04 ] // The Aurifex standard
What gold standard
means at Aurifex
Purpose
It solves a real problem in skilled work.
Skilled people lose hours to the slow, repetitive part of their craft, the part that keeps them from the work only they can do. Every Aurifex product starts from one of those problems, chosen before the technology, and earns its place only by making the work faster, clearer or better.
Proof
It’s proven before it ships.
Every release is tested against real examples of the skilled work it is for, with the bar set before building starts. Nothing ships on a hunch, and every result we share says how it was measured and what it doesn’t show.
Experience
It’s effortless to use.
A gold standard includes the experience. Our products fit into the way skilled people already work, ask for as little as possible and stay out of the way, so their expertise gets the attention and AI simply makes it go further.
[ 05 ] // What it takes
What it takes
to hold the standard
Built once at Aurifex and shared by every product, so each one meets the same standard without starting from scratch.
Product and experience design
Product leadership, research, interaction design and review panels for every build, so each product feels simple from the first screen.
AI engineering and evaluation
Shared evaluation sets, prompt-injection and leakage tests, subgroup checks, and a release gate that blocks a change that misses the bar.
Security and compliance
Least-privilege access, audit logs on every AI action, a pause switch, and the compliance each industry requires.
Research and strategy
Market sizing from the bottom up, competitive reviews and the published evidence, written down before the first screen is drawn.
Brand and go-to-market
Naming, identity, websites, launch plans and pricing work, so each product reaches its market with its own voice.
Capital and operations
Finance, legal and release discipline: separate development and production builds, and nothing promoted without approval.
| Product | Core problem | What AI does | Held to the standard by | Stage |
|---|---|---|---|---|
| Aurifex Healthcare | Patients wait for the next place of care | Predicts discharge dates, drafts rehab referrals | Cited sources, clinician approval | Live |
| Veritas | We forget most of what we read | Reads in your voice, checks, schedules review | Answers grounded in your own book | Launching Q1 2027 |
| Nightdrive | Audio tools are hard to use | Builds the effect chain from a description | Undo everything; files stay on your device | Launching Q2 2027 |
| Bookloft | Nobody remembers every book they own | Reads spines from video and catalogs them | Confidence labels; you check the unsure ones | Launching Q3 2027 |
| Ranvier | MCAT prep without a clear path | Plans each day; an AI tutor on every passage | Explanations tied to the passage; scores checked | Launching Q3 2027 |
| Truefill | Job applications are slow, and shortcuts guess | Finds roles, tailors resumes, fills forms | Every answer read back; you approve each send | Launching Q3 2027 |
| Goldsmith | Custom jewelry is opaque and overpriced | Turns words into manufacturable 3D designs | Manufacturing checks on every design | Launching Q4 2027 |
[ 06 ] // Markets
Where skilled work
needs better AI
Different industries, the same pattern: a problem skilled people live with every day, and room for a product that makes it effortless.
Healthcare · Aurifex Healthcare
The days between medically ready and the next place
A patient is ready to leave, but the rehab bed, the authorization or the equipment isn’t arranged. The bed stays full while admitted patients wait in the emergency department. Hospitals already measure length of stay and excess days, so the result can be checked against each unit’s own baseline.
Learning · Veritas
People read less, and forget most of what they read
The share of Americans reading for pleasure on a given day fell from 28% in 2004 to 16% in 2023. Students who practiced recall remembered 61% a week later; students who reread remembered 40%. The tools people use today each do one step, and none knows what you missed.
Bone et al., iScience, 2025 · Roediger & Karpicke, 2006
Music · Nightdrive
Slowed, sped-up and 8D edits, without the learning curve
Slowed + reverb, nightcore and 8D versions are a lasting listening trend, but most tools are built around a single speed slider or need real audio skills. Nightdrive leans on researched presets, effect combinations and proper file export.
Books · Bookloft
Home libraries nobody has cataloged
Reading apps track what you read, but almost nothing tracks the physical books you own: what you have, where it is and what it’s worth. Typing in hundreds of titles is the reason most collections never get cataloged.
Medical education · Ranvier
About 86,000 people sit the MCAT each year
They prepare for months with a patchwork of question banks, flashcards and practice exams, and the coaching that helps most is billed by the hour. Ranvier starts with the MCAT, built to make the daily plan and the coaching part of the product.
Examinee volume from AAMC data, as used in Ranvier’s plan
Careers · Truefill
Hundreds of applications, the same questions every time
A serious job search means filling the same history, work authorization answers and resume into form after form. Auto-apply tools trade accuracy for volume and guess what they don’t know. Truefill is built for the opposite: fewer, better applications, every answer checked.
Jewelry · Goldsmith
A $367B–$371B category with an opaque custom slice
Custom work is commonly marked up 200% to 500%, and a buyer sees only the final number. Estimates for custom jewelry services alone run from about $1.09B to $3.63B depending on the publisher, so Goldsmith sizes its market from the manufacturers and buyers it can reach.
2024 aggregator estimates, as cited in the Goldsmith plan
Outlook · projected net revenue, 2027–2031
$8.7M expected net revenue in 2031
- $8.7M
- expected net revenue in 2031
- Up to $13.5M
- net revenue in 2031
- 20,000
- simulated outcomes
View as a table
[ 07 ] // Milestones
Where Aurifex
is going
Aurifex is founded
Aurifex LLC is formed in Wisconsin.
Goldsmith plan presented to investors
Product plan, research plan, business model and financial model, followed by a working platform in 12 weeks.
Aurifex Healthcare evaluated on 25,000 admissions
Discharge-date predictions, rehab-candidate flags and drafted referrals tested on held-out historical data and simulated cases.
Aurifex Healthcare goes live
Aurifex Healthcare is live, and six more products are in development for 2027.
Veritas launches
Veritas opens to the 2,500+ people on its waitlist and to the public, starting with students who have an exam ahead.
Nightdrive launches
Nightdrive opens to the public.
Bookloft, Ranvier and Truefill launch
Bookloft opens on the web with whole-bookcase scanning. Ranvier opens for MCAT students. Truefill arrives in the Chrome Web Store.
Goldsmith launches
AI jewelry design and a verified-manufacturer marketplace open to buyers, jewelers and makers.
First full year with all seven live
Every product is in market for a full year. Expected net revenue: about $1.0M.
Hyperscale
Expected net revenue of about $2.8M in 2029 and $5.6M in 2030, led by Veritas.
$8.7M expected net revenue
Across all seven products on the current trajectory, with an expected company value of about $31M.
[ 08 ] // Leadership
Edward Profio, MBA, MSc, PMP
Serial Entrepreneur · Chief Executive Officer & Founder, AI/ML · Harvard University ’27 · Carnegie Mellon University ’25
Edward is the founder and Chief Executive Officer of Aurifex LLC. He is co-founder and Chief Product Officer of Aurifex Healthcare, and built Veritas, Goldsmith, Nightdrive, Bookloft, Ranvier and Truefill from research to working product. At UW Health he launched Epic-integrated GenAI assistants that returned an estimated 42,000+ clinician hours in the first year of a 500-clinician pilot, then scaled organization-wide. Earlier he held founding product roles at Gallify and CODAmarket.
Raw material in.
Gold out.
If your work deserves gold-standard AI, or one of our products touches it, write to Aurifex LLC.
Partners
Health systems, schools, manufacturers and distribution partners.
Investors
Conversations about Aurifex, its mission and its products.
Builders
Teams and founders who want to bring AI into skilled work the same way.
Careers
Product, AI engineering and design roles across Aurifex.
info@aurifex.com