About

A casino dealer started this company.

Candor AI is a Pomona-based AI company. The founder is Alex Escobar — born in Guatemala, raised in San Bernardino, sixteen years in Pomona, working full-time at a casino while he builds this business with his own savings. That biography is not marketing copy. It is the reason this company exists, and the reason it works the way it does.

Why it exists

This started with the people
at his table.

Alex deals cards for a living. The customers at his table are hair salon owners, nail salon owners, tax preparers, mechanics, accountants, restaurant owners — the small business owners of Pomona and the surrounding Inland Empire communities. He hears what they deal with on a slow Tuesday: missed phone calls, spam bookings, no-shows, walk-ins they couldn't quite communicate with, hours wasted answering the same five questions.

AI tools could help with that. The problem is that almost none of the AI tools being built today are built for them. They are built for tech companies, for venture-backed startups, for enterprise customers — and then resold downmarket with the rough edges still attached. The pricing is wrong, the language is wrong, the assumptions are wrong, and the support is impossible to reach when something breaks at six in the morning before the salon opens.

Candor AI exists to fix that. Not at scale. On purpose.

I don't want to be the biggest AI company. I want to be the one forty businesses in Southern California can call and trust. That's the whole thing.

The forty-client cap

Why we say no
to client forty-one.

Every AI agent Candor AI builds is bilingual, legally compliant with California's specific rules around recording disclosure (CIPA), data privacy (CCPA), and consumer protection — and it gets tested before it goes live with real customers. That work is meticulous. It does not scale linearly with another sales hire.

Capping the client list at forty is how we promise the forty-first thing about us — that the people we already serve are not going to be pushed to the side because someone with a bigger budget walked in. The cap is the feature.

When the list fills, we open a waitlist. We do not negotiate the cap. The clients who joined first keep their rates. That is the agreement, and we honor it.

See how onboarding works
Equally important

What Candor AI
won't do.

A lot of what makes a small business AI company trustworthy is what it refuses to do. Here is our list — written down on purpose, in public, so we can be held to it.

  1. We don't promise outcomes.

    We don't tell you the agent will book twenty appointments a week, or recover a specific percentage of missed calls. Anyone making that promise is guessing — or worse. We tell you what the agent does, what it doesn't do, and what we've seen with other clients in similar situations.

  2. We don't replace your people.

    Candor AI agents are built to support the people who run a small business, not to replace them. The relationships, judgment calls, and customer trust at the heart of a working business stay with the humans who built them.

  3. We don't hide what the agent can't do.

    Before you sign, we will tell you what the agent will not handle — and what we are recommending you keep human. We will also tell you about recording disclosure requirements, response time expectations, and any limitations of the platform we're deploying on.

  4. We don't sell your data.

    Your client data, your conversations, your business intelligence — it stays yours. Our written agreement spells this out plainly. Specific terms are in the Terms of Service (coming soon); the principle does not change.

  5. We don't pretend to be lawyers, accountants, or insurance experts.

    The AI doesn't either. When a question gets into territory that needs a CPA, an attorney, or an insurance broker, the agent says so — and routes the conversation accordingly. Pretending otherwise is how small businesses get hurt.

Next

Want to see
how onboarding works?

The path from a first conversation to a working AI agent is not complicated, but it has steps — and we ask things of you at each one. The page next to this lays out what the process actually looks like.

See how it works