VANI: outbound email writer

VANI writes to people who have never replied

VANI is the Sales team's outbound email writer, and is marked Beta. Its name is Sanskrit for voice. VANI is a writer, not a voice or calling agent: it drafts the first-touch, follow-up and breakup emails NORA sends to prospects who do not know you yet.

Its sibling, LIPI, writes to people you have already spoken to. The two are split on purpose. A cold email has to earn thirty seconds of attention from a stranger, so VANI works from a research dossier; a warm one continues a conversation, so LIPI works from the CRM record.

What VANI reads

For each prospect, VANI is handed what NORA's research produced:

• the prospect dossier: the company, the person, the signals and the pitch hooks NORA found, each with its source
• the cadence step it is writing for: first touch, follow-up or breakup, and what was sent before

Around that it assembles the same workspace grounding LIPI uses: its own instructions (including your workspace's edits), your Knowledge Hub, your brand voice, your Offerings, and the sender's name. It may name a product or outcome only if it appears in your Offerings. A workspace with neither Offerings nor knowledge gets drafts written on the brand name alone, with no product claims.

What VANI produces

One email per call, in a fixed shape: a subject of 60 characters or fewer, a preview line, a body of three to five short paragraphs, exactly one low-friction ask, and a list of every specific claim it made about the prospect.

That claim list is the point. Every claim about the prospect must appear in the dossier, signals, hooks or contact record, and a validator checks it. VANI does not invent funding, hires, headcount, customer names or numbers, does not name-drop customers, and never opens with a stock phrase or a greeting with an unfilled name in it.

For Prompt World, a first touch to a head of marketing would open on something NORA actually found, such as a posted role for a sales enablement lead, and close on one question, such as whether a 15 minute call next week is worth it.

Where the decision to send sits

VANI declares no actions of its own. The send is NORA's, and NORA's actions carry their own levels:

A Sales team installed at sign-up is seated at L4, which is enough for drafting. Sending is gated separately from the level on NORA's seat, and on a new workspace every send of VANI's drafts waits for a person's approval. A workspace admin can raise or lower NORA's level on her Configuration tab. See the Sales team for the whole picture.

You can read and edit every draft on NORA's screen before it goes, whatever the level.

Briefing VANI

You do not address VANI in the room, and you may not see it in the @ list. Brief NORA instead; she calls on VANI whenever outreach is drafted. To change how VANI writes, open its agent page and edit its instructions. Your edit is kept as your workspace's own copy, and a newer version from Nagent is offered for review rather than applied over it.

Our logic

A cold email is the one place a fabricated detail does the most harm, because the reader has no reason to forgive it. So VANI writes only what the research can back, lists what it claimed so it can be checked, and signs as a real person or as your team. It never presents itself as a person who is not one.

Limits today

• Beta. Read the first batch before it goes, especially while your Knowledge Hub and Offerings are thin.
• Email only. VANI writes email. It does not write LinkedIn messages, place calls or speak.
• No memory of its own. VANI does not keep Smriti or learn through Karmic; it writes from what NORA hands it each time.