Private beta · onboarding now

Operate
like ten.

Ondari is an AI proxy that lives in your company Slack — a parallel you that fields and answers the repeated questions found in your documentation systems and conversation histories. It only taps you in when there's a genuine need for a human.

Join waitlist → Contact

Private beta · setup in under 30 min

#ask-alex — Ondari

#ask-alex Proxy active

Jordan M. 2:14 PM

When are OKR scores due this quarter?

Ondari answering on your behalf

Scoring is due 5 business days before the new quarter starts. Post your team update in #okrs by the 15th.

CONFIDENCE

91% · OKR policy v3

Riley T. 2:31 PM

Can you approve the $8k design agency invoice?

Ondari answering on your behalf

Escalating — this needs VP sign-off per the expense policy. Tagging @alex with the context.

CONFIDENCE

24% · needs a human

# Message #ask-alex…

How it works

Set up once.
Then stay out of your own way.

Your proxy handles the rest — answering from your knowledge, learning from every escalation, and interrupting you less over time.

Open your channel

Install Ondari in Slack and name it (#ask-charlie, #hey-parker) — whoever your team already reaches for.

Feed it what you know

Drop in your SOPs, connect Google Docs, pull Slack history. The more context, the smarter the proxy.

It answers. You don't.

Every reply lands with a confidence score and the source it came from.

It learns the edges

Every escalation you resolve becomes an answer next time. The handle rate climbs; your distractions fall; your focus time is protected.

The line

It knows what it
doesn't know.

Most chatbots guess. Ondari draws a line — and only crosses it when a question genuinely needs you.

"What's the vendor onboarding process?"

HANDLED · 94%

"When are OKRs due this quarter?"

HANDLED · 90%

"Who approves contractor invoices over $10k?"

HANDLED · 86%

your threshold — 65%

"Should we counter on the Meridian contract?"

→ YOU · 22%

You set the threshold. It respects it.

Every response carries a confidence score from 0–100%. Above your line, Ondari answers and cites its source. Below it, it tags you once — in the thread, with the reason it couldn't answer — and never guesses in your name.

Pin certain topics to always route to you, regardless of confidence: budget calls, personnel, anything that should never be automatic.

confidence < 65% → escalate to @you

The dashboard

The inside view.

Beyond the channel, Ondari gives you an operator's dashboard — what's getting answered, what pulls you in, who keeps asking, and what to teach your proxy next.

Ondari — Dashboard

Questions answered

1,284

+18% this month

Hours saved

47.5h

≈ 6 workdays

Pulled you in

112

8.7% of questions

Avg confidence

84%

+3 pts

Answered vs. escalatedlast 8 weeks

AnsweredEscalated

Top askers

JMJordan M.214
RTRiley T.168
PNPriya N.142
DKDevin K.97

Not just a chatbot.

A complete proxy system — knowledge, escalation, analytics, and integrations, built in.

Live Slack proxy

A real channel your team can message. Answers come with confidence scores, source citations, and a clean escalation trail.

Knowledge base

Add docs, SOPs, or Slack messages as context — paste in a single email thread if that is where a policy lives. See which sources get cited most — and which are going stale.

Confidence-based escalation

Set your threshold. Ondari escalates when it's unsure — tagging you once, in-thread, with the reason it stopped.

Analytics

See exactly how much time your proxy saves — questions answered, handle rate, confidence trends, and the topics coming up most.

Escalation review queue

Every escalation lands in a queue. Resolve it, or "add answer to knowledge" to teach the proxy so it handles it next time.

Integrations

Connected today, with more on the roadmap. New integrations ship to every plan as they land.

Google Drive
Slack history
Upload & paste
Notion — soon
Linear — soon
HubSpot — soon
Confluence — soon
Zoom — soon
Calendar — soon

Security

Your knowledge stays your knowledge.

You are pointing Ondari at your SOPs, your team docs, and your Slack history. Here is exactly what happens to them.

Scoped to your workspace

Your data is isolated per workspace. It is never pooled with another customer’s and never used to train models — ours or anyone else’s.

Permission-aware answers

The proxy answers only from sources the channel is allowed to see. Beta scopes at the channel level; source-level permission mirroring is next.

SOC 2 in progress

Type I is our launch target and Type II is on the roadmap. Not certified yet — ask us where it stands and we’ll tell you straight.

Built for you, specifically

The person everyone
comes to.

If you're the connective tissue of your org — the one who knows the answer to everything — Ondari is for you.

Chief of Staff

You know where everything is. Now everyone else can find it too.

Ondari fields the process questions, policy lookups, and "who do I talk to" redirects — so you stay in strategic mode.

"What's the vendor onboarding process?"
"When are OKRs due?"
"How do I prep for the board meeting?"

Executive Assistant

Your job is to protect the exec's time. Ondari protects yours.

Answer the repetitive requests automatically. Escalate the ones that actually need a human. Get your focus back.

"Is Thursday afternoon free for a call?"
"What's the travel policy for conferences?"
"Who approves contractor invoices?"

Founder / Head of Ops

You're the team's Wikipedia. Time to make that automatic.

At 20 people you still answer every question. Ondari lets you delegate institutional knowledge without another hire.

"What's our hiring process for engineers?"
"How do I submit a purchase request?"
"What's the deal with the new equity plan?"

What changes when the pings stop

The math of a team of one.

These are the standards we hold the beta to — and the numbers your dashboard reports back, so you can hold us to them.

80%

handle-rate target within 30 days of setup

target time saved per owner, per month

answers sent in your name below your threshold

±10pts

calibration target — a 90% should be right about 90% of the time

Pricing

Simple and honest.

Pricing for when the beta opens up. Anything marked Coming is on the roadmap, not shipped — you will never be billed for a feature that is not live.

Starter

$29 /mo · per proxy

1 proxy channel
5 knowledge sources
500 answers / month
Slack escalation tagging
Basic analytics

Join waitlist

Questions, answered.

What is Ondari?

Ondari is an AI proxy that lives in your company Slack. It sits in a channel your team already messages you in, answers the repeated questions it can find in your documentation and conversation history, and tags you in only when a question genuinely needs a human.

How does Ondari decide when to escalate to a human?

Every Ondari response carries a confidence score from 0 to 100%. You set the threshold — 65% by default. Above your line, Ondari answers and cites the source it used. Below it, Ondari stays quiet, tags you once in the thread with the reason it stopped, and never guesses in your name.

You can also pin specific topics to always route to you regardless of confidence, such as budget decisions, personnel matters, or anything that should never be automatic.

Where does Ondari get its answers?

Ondari answers from the knowledge you give it: SOPs and internal policies, connected Google Docs, and your existing Slack history. Every answer cites the source it came from, so your team can check the original document rather than taking the reply on faith.

What tools does Ondari integrate with?

Today Ondari connects to Google Drive and your Slack history, plus anything you paste or upload directly. Notion, Linear, HubSpot, Confluence, Zoom transcripts, and Google Calendar are on the roadmap and will ship to every plan as they land. Starter includes five knowledge sources; Pro and Team include unlimited sources and every integration that is live.

How long does setup take?

Setup is designed to take under 30 minutes. You install the app, name the channel your team already reaches for, then point Ondari at your documents. The proxy starts answering immediately and gets more accurate as you resolve escalations.

Does Ondari get better over time?

Yes. Every escalation you resolve becomes an answer Ondari can give on its own next time. Unanswered questions collect in a review queue alongside a list of knowledge gaps, so you can see exactly what to teach the proxy next. The handle rate climbs and the interruptions fall.

Who is Ondari for?

Ondari is built for the person everyone comes to: Chiefs of Staff, Executive Assistants, founders, and Heads of Operations. If you are the connective tissue of your organization and spend your day answering the same process, policy, and "who do I talk to" questions, Ondari handles that traffic for you.

What happens to my data?

Your knowledge is scoped to your workspace. It is not pooled with other customers and it is never used to train models. The proxy answers only from sources the channel is permitted to see — during beta that is channel-level scoping, with source-level permission mirroring next on the roadmap.

SOC 2 Type I is our launch target and Type II is on the roadmap. We are not certified yet, and we would rather tell you that here than let you find out in a security review.

How much does Ondari cost?

Ondari starts at $29 per month per proxy for Starter and $79 per month for Pro. Team is $199 per month for five proxies and is currently early access, since the shared knowledge base, admin dashboard, and SSO are still being built — get in touch and we will onboard you as they land. Ondari is in private beta, so pricing applies once your workspace is onboarded.

Your team has questions. Your proxy has answers.

Setup takes under 30 minutes once you are in.

Join waitlist →Contact