How Vanta’s reps call the accounts they own while an AI SDR emails the ones nobody owns

What Vanta did

Pipeline was almost all inbound

When Stevie Case joined Vanta as its first CRO, she has said, it sold almost only to startups, all inbound, without much use of its CRM or a forecast. As late as about 2023, almost all pipeline still came from search marketing.^2,3,4

At the end of 2025 her leadership team set a 2026 pipeline target more than 70% above the year before, which Case said would take many more SDRs and AEs who build pipeline themselves.⁵

By August 2026, how much came from outbound depended on who was buying.²

Security buyers don’t answer cold email

Upmarket, the buyer Vanta most needs is usually a chief information security officer (CISO), and in September 2025 Case said CISOs don’t respond to cold email: most assume it is phishing and filter it out of their inbox.⁴

“Human SDR is picking up a phone and having a meaningful conversation with someone actually cuts through when AI-generated email does not.”

Stevie Case, CRO, to First Round Review, February 2026⁶

Where Vanta draws the line

In September 2025 Case said she was not a fan of AI-written email or automated outreach; Vanta was instead hiring dozens of SDRs to sit in an office and pick up the phone.⁴ She restated it in February 2026: “jobs like SDR for the foreseeable future should still be human first.”⁶

On owned accounts the work happens on the phone. By Case’s September 2025 account, roughly 70 to 75% of outbound meetings come from calls.⁴

Vanta’s April 2026 posting for a business development rep says “Cold calling is at the heart of this role” and sets “daily activity minimums of 120 activities.” In August 2026 Case said cold calling was still Vanta’s primary way of generating pipeline.^2,7

What an SDR finds each morning

Every SDR opens Salesforce to a ranked list someone else built.

The list comes from a propensity model Ryan Levy’s revenue analytics team built from closed-won accounts, reverse-engineered down to the size of the security team, its titles and its tech stack.⁴

The signals are ones only a compliance vendor would look for, such as a SOC 2 badge on a startup’s website, and they differ by buyer.⁴

Software then finds the people in each account who look like Vanta’s buyers, and RevOps feeds them into Salesforce as a stack-ranked list, Case said in October 2025.^1,4

The prep got shorter

Every name comes with prep, and by September 2025 Vanta had cut the first, long-form version down.⁴

Case calls the prep raw materials, and Vanta discourages SDRs from taking a template and hitting send.^1,4

A mockup of the prospect’s trust center

Some outreach carries a mockup of the prospect’s own trust center, the public security page Vanta hosts more than 6,000 of, with the prospect’s logo on it.^4,8

Vanta Trust Center · Vanta’s product page×

An AI SDR, fenced off from owned accounts

By April 2026 Vanta also ran an AI SDR that sends email on its own, limited to accounts “not in Salesforce. No owned accounts, no existing opportunities,” The Signal reported from its interview with Shashank Khanna, whose GTM engineers built it.⁹

It sorts the people it finds into personas Vanta calls Suzy Security, Emilia Engineer and Forest Founder, and the signal picks the sequence: “A competitor customer gets a competitive rip sequence.”⁹

A red light, such as a company that just bought a competitor, is suppressed and re-queued instead of emailed. “You can’t write copy good enough to make a red light green,” Khanna wrote in May 2026.¹⁰

The result

The AI SDR, in Khanna’s own numbers

At a public event in May 2026 he showed a live dashboard: 2,447 emails to 585 contacts at 410 accounts, and 9 open opportunities. His slide put the open rate at 46%, 1,059 of 2,290 delivered, and the meeting-booked rate at 4.7%, 9 open opportunities from 192 completed prospects, with “No rep intervention.”¹¹

Shown live by Shashank Khanna, Vanta, May 12, 2026

Khanna’s AI SDR dashboard×

Both rates cover one period and only accounts nobody owns.¹¹

What it took

Three teams built it. Levy’s team built the model and shows its scores in Salesforce and Tableau. RevOps writes the prompts and the list feed, and Khanna’s GTM engineers built the AI SDR.^4,9 By January 2026 Khanna was hiring a new GTM Engineering team to build agents for pipeline and admin work.^12b

Willie Pierson, who leads sales development, wrote in January 2026 that more than 100 people had joined the SDR team in a year.¹³

What to borrow

1. Let your buyer set where AI stops. If buyers filter cold email, use AI for the list and the prep, and leave the message and the call to a rep.
2. A few sentences and two or three angles are enough prep. Vanta dropped its long research document after Case found it took reps too much analysis.
3. Test full automation on accounts with no owner and no open opportunity, and measure it apart from your human reps.

How to build a version of this in Unify

Build it in Unify

From here on, the page describes our suggested version in Unify.

Most of what Vanta had to build was the data underneath: a revenue analytics team trained a model on closed-won accounts, RevOps ran an enrichment tool and the list feed into Salesforce, and Tableau showed the scores. Unify combines your CRM, warehouse and product usage data with enrichment from 40+ sources, with warehouse tables synced in through a sync tool or the API. The ranking comes from Audience criteria and Signals you can read, in place of a trained model, and direct numbers come from waterfall enrichment across 11+ phone and email vendors.

To start, build one Audience of accounts that look like your wins, with two or three Signals you can explain. When an account enters it, a Play assigns the owner, finds people across up to five personas with direct numbers and starts a phone-first Sequence the rep approves, with AI Research on each task. Accounts with no owner and no open opportunity can go to a second, more automated Play, compared with the first in Insights.

outbound motion

Do it in

Closed-won profileWho Vanta wins, reverse-engineered

✓

AudienceAccounts that look like the ones you win, from CRM fields and firmographics

Category signalsSOC 2 badge, incorporation date, a breach

✓

AI Agent qualification and SignalsAn Agent question only your category asks, such as “Does the website show a SOC 2 badge?”, plus website visits, new hires and champion job changes

Daily ranked listFrom an in-house propensity model

✓

Audience and Signals, worked in orderOwned accounts in the Audience, each with the Signals that matched, for the rep to work through; with criteria anyone on the team can read

People inside the accountContacts added for the SDR

✓

Prospect action and B2B Data1 to 10 people per company across up to five personas, with direct phone numbers found by waterfall across 11+ vendors

Short prepA few sentences and 2 or 3 angles

✓

AI ResearchA short summary and angles on each task, built from the Observations you choose

The SDR writes and callsMost meetings come from calls

✓

SequencesCall steps first, then action items and manual emails the rep approves, with delays in days

Owner and AE pairingAssigned in Salesforce

✓

Owner assigned by the PlayThe Play runs when an account enters the Audience and assigns the owner, synced both ways with Salesforce or HubSpot

The AI SDR fenceOnly accounts not in Salesforce

✓

A second Audience and Play, compared in InsightsAccounts with no owner and no open opportunity, with automatic email steps. Insights > Pipeline generated shows the two Plays side by side

Closed won and lostFed back into the system

✓

Two-way CRM sync and InsightsClosed won and lost come back from your CRM; Insights > Pipeline generated, with its Closed Won filter, shows which Plays the won deals came from, and you change the Audience criteria by hand