An AI SDR by OnePgr · Currently in supervised mode

Outbound, built
on research.
Not guesses.

AIxSDR researches every account, scores prospects against your ICP and Anti-ICP, and earns the right to send on your behalf — starting supervised, graduating to autonomous as it learns your judgment.

Filed under — Research-first outbound
Stack — OnePgr · Kampaign · Ush3r
Status — Live · accepting design partners
For — Founders, GTM teams, sellers
Figure 01 — Your AI SDR's Trust Ladder
L1
Supervised
L2
Reviewed
L3
Trusted
L4
Delegated
L5
Autonomous
← Every draft reviewed Sends on your behalf →

The opposite of spam isn't volume control.
It's research. Deep, opinionated, specific-to-your-buyer research.

The AIxSDR brand premise · Filed 05/2026

Three steps. In order. Skipping any one is how outbound got broken.

STEP 01
It researches first.

Before a single email is drafted, AIxSDR's research engine reads everything public about your prospect — job posts, 10-Ks, recent funding, leadership changes, tech stack signals, podcast appearances, recent LinkedIn posts.

You ask: "Research Stripe and give me their top pain points." It comes back with a sourced answer. That research becomes the foundation for everything downstream.

→ Powered by the OnePgr research engine
STEP 02
It builds your ICP — and your Anti-ICP.

Through a guided 5-round conversation, AIxSDR builds a weighted ICP: industries, company size, funding stage, target titles with P1/P2 priority, tech signals with confidence scores, and per-persona pain points and motivations.

Then it does what no other AI SDR does: it builds your Anti-ICP — specific, opinionated signals that mark a prospect as a bad fit and deprioritize them. Who NOT to email is as important as who to.

→ See exhibit below for sample Anti-ICP flags
STEP 03
It earns the right to send.

Day one, every draft hits your queue for review. Approve it, edit it, or skip it — every action trains the model. Each approval is +1. Each edit, -2. Each skip, -1.

As your AI SDR's score climbs, it earns autonomy: from Supervised to Reviewed to Trusted to Delegated to fully Autonomous. It only sends on your behalf after it's earned it.

→ Trust Ladder explained in detail below
Exhibit A · Sample Anti-ICP flags from a live profile

Most AI SDRs tell you who to email.
We tell you who not to.

Every flag below is a real Anti-ICP signal from one of our customer's ICP profiles. Each one represents a hard-earned lesson about who not to waste outbound on. The AI deprioritizes any prospect matching these signals — automatically.

Flag 01
Under 30 employees — still in product-finding mode
→ No revenue urgency yet
Flag 02
Solo founders or 2-person teams — no one to run a POC
→ Pilot would stall
Flag 03
Solo technical founder who explicitly hates sales
→ Cultural mismatch
Flag 04
Founders publicly hiring a CRO/VP Sales — cycle stalls 4–6 months
→ Re-engage post-hire
Flag 05
CRO/VP Sales hired in last 30 days — re-engage at day 75
→ Honeymoon period
Flag 06
Ex-FAANG infra founder with 'build everything' mindset
→ Will build in-house
Flag 07
Raised from 'build GTM in-house' VC doctrine firms
→ Vendor-resistant board
Flag 08
5+ SDRs/AEs already in seat — build already done
→ Wrong stage
Flag 09
Repeat founder with existing vendor relationships from prior company
→ Switching cost too high
Flag 10
Already using Outreach or Salesloft in production
→ Replacement, not net-new
Flag 11
Existing in-house outbound team that's working
→ Nothing to fix
Flag 12
Still in beta / closed access — no outbound urgency
→ Pre-revenue

An AI SDR
that earns
autonomy.

Every other AI SDR starts on day one with the keys to your domain, your brand, and your reputation. We think that's backwards.

AIxSDR starts at Level 1, Supervised. Every draft hits your queue. Every approval, edit, and skip trains the model. As trust accrues — quantified, scored, visible — autonomy unlocks one level at a time.

When AIxSDR finally sends on your behalf, it's because it's earned the right.

Figure 02 — Five levels of autonomy
L1 · 2/50
Supervised
100% reviewed
L2
Reviewed
80% reviewed
L3
Trusted
50% reviewed
L4
Delegated
20% reviewed
L5
Autonomous
Spot-check only

The other AI SDRs
send on day one.
We earn the right.

Other AI SDRs
AIxSDR
Research
List enrichment, light intent data
Deep, sourced, per-account pain research
ICP definition
Filters: industry, size, title
Weighted multi-dimensional model + Anti-ICP
First send
Day one. Volume from minute one.
Only after the AI has earned the right.
Autonomy
Binary: on or off
5 levels. Earned, not granted.
Who NOT to email
Exclude list (optional)
Anti-ICP as first-class concept
Default mode
Volume. More sends = more meetings.
Quality. Right account + right message.

What your AI SDR
is actually doing right now.

Exhibit B · ICP scoring
Every prospect scored honestly.

Including the ones we recommend against. Real signals from a live profile.

Robert Waye · Head of Growth48 · Grey
Curtis Lee · VP Sales42 · Grey
Jason J. Lee · Head of Sales40 · Grey
⚠ Industry mismatch: Financial ServicesANTI-ICP
⚠ No revenue stage data — likely Series C+ANTI-ICP
Exhibit C · Trust log
Every action measured.

Real trust history from a Level 1 AI SDR. +1 unedited send, -2 edit, -1 skip.

21m ago · Send unedited+1
6d ago · Send unedited+1
6d ago · Edit & send-2
6d ago · Edit & send-2
Current score2 / 50
Exhibit D · A real draft from a real AI SDR
Drafts that earned their send.

Hi Robert,

I noticed you recently took on the CRO role at Credivera — congrats on the promotion. Given your background scaling revenue at organizations like Vecima and Aware360, I thought you might find this relevant.

We work with B2B SaaS companies in the IT services space to accelerate their sales pipeline and improve deal velocity. Many of your peers are using our platform to reduce sales cycle length by 30–40% while increasing conversion rates...

From: rajiv@getkampaign.com → To: robert@credivera.com · Status: Awaiting your approval

Start at Level 1.
Earn your way up.

Run your first ICP test, build your Anti-ICP, and watch your AI SDR earn autonomy one approval at a time. Free during open beta.

Start at Level 1 — Supervised
No credit card · Currently accepting design partners