About
Why I built this
Fromenance started with a phone call my mother made two months ago. Communication provenance is the answer I could not find for her that day: when a customer asks a bank "did you send this?", the bank should be able to say yes or no, in seconds, and mean it.
The call
What happened to my mother
In July my mother got an email that looked exactly like her bank. Same logo, same tone, a warning that her account had been hacked and a link to secure it. She was careful. She did not type anything into the email. She clicked through to see what was going on, landed on a page that looked exactly like the bank, and that page gave her a phone number to call. So she called it.
A calm, professional voice answered with the bank's name and walked her through "securing" the account. Her account had not been hacked. It was being hacked, right then, by the people on the phone, and the warning that brought her to them was the crime itself. By the time she hung up, thousands of dollars were gone.
Every step she took was reasonable. She was skeptical enough to want to speak to a person, and she called the number she was given. What she never had was a way to ask the only question that mattered and get a real answer: did my bank actually send this? The bank had DMARC, but the email did not use the bank's domain. The spam filter had guessed and guessed wrong. The copycat site came down later, which helped the next person and did nothing for her. And if she had reached a real agent, the agent would have looked at the same email and guessed too.
I started building Fromenance within days. The bank is the only party that knows for certain what it sent, so the record has to live on the bank's side, and the answer has to reach the customer before they pick up the phone. I wrote the whole story in The phone call my mother made.
What we believe
Three things we are sure about
01
Only the sender can answer
Nobody else has the record. A classifier can say a message looks wrong. Only the institution can say "we sent this to you at 10:42" or "we have no record of this". So we built the registry on the sender's side and left the guessing to everyone else.
02
The answer should be definitive, and take seconds
Not a score. Not a ticket that comes back on Thursday. Verified, Not verified, or Known fraud, with wording that is locked so it never claims more than the registry supports, delivered before the customer has put the phone down.
03
The fraud team should get the lure
Every message that fails verification was convincing enough that a real customer paused and asked. That is a phishing sample delivered by its target. It should land on your fraud team's desk in seconds, with the domains and phone numbers already pulled out, not in an abuse mailbox someone reads on Friday.
Who we are
A small team, on purpose
Fromenance is built by George Rios and a small team at Many Software Enterprises. George spent his career in enterprise software and insurtech, which is where he watched institutions buy control after control and still have to tell a customer "we can't be sure". Then it happened in his own family. We stay small because the product is infrastructure and the buyer is an institution. The sale is a conversation with the person who owns impersonation, and we would rather have that conversation ourselves.
We write the way we build. If a page on this site makes a claim, the product makes the same claim in its tests. If we do not know a number, we do not print one.
How we work
What a pilot with us is like
- We add to your stack, we do not replace it. Keep DMARC at reject. Keep your gateway and your takedown vendor. We sit above them, and we feed them what your customers send us.
- Your customers never see our name. The address is on your domain, the page is on your domain, the reply is in your branding. Fromenance shows up in your admin app and our docs, and nowhere else.
- Security first, then pilots. We finished the data handling story, the DPA, and the subprocessor list before the first pitch, because a bank's vendor review was always going to ask.
- We build in order. The intelligence tier is built after a tenant is paying for verification. Cross institution warnings are built after an analyst has confirmed a campaign from real submissions. No calendar dates, just gates.
Start with one stream and a 60 to 90 day pilot.
Pick your fraud alerts or your transaction alerts, the messages your customers already squint at. At the end you get a written report: how many people asked, what they were told, which campaigns turned up, and what your existing feeds missed.