L

Leo

Autonomous Outbound Sales Agent · Stella Module

Researches. Drafts.
Follows up. Escalates.

Leo is an AI sales agent that runs inside any Stella deployment. He finds the right contacts, writes emails in your organisation's voice, sends them only when you approve, and never drops a lead — even if it goes quiet for months.

Researches carefully Drafts with warmth Follows up patiently Escalates decisively Knows when to wait
leo · campaign: erasmus-coordinators
leo › run scan --campaign erasmus-eu Scanning 14 university networks... ✓ Found 38 new Erasmus coordinators Cross-referencing existing contacts... ✓ 31 net-new contacts identified   leo › draft --contact "Dr. K. Müller, TU Berlin" Drafting using Stella RAG context... ✓ Draft ready · Sending to Zulip for approval   leo › status --campaign erasmus-eu Sent: 18 · Opened: 11 · Replied: 3 Hot lead: Univ. of Valencia · Escalating → Next follow-up: 47 contacts · in 6 days   leo ›

What Leo does

An outbound agent who never forgets a lead

Leo is not a mail merge tool. He is not a CRM plug-in. He is an autonomous agent that runs inside your Stella deployment, inherits your organisation's identity and knowledge from Stella's RAG, and conducts outbound sales campaigns the way a skilled human sales professional would — carefully, patiently, and with genuine relevance.

Leo scans the web for contacts, drafts emails in your voice and your context, sends them through an approval gate, monitors responses, follows up at the right intervals, and escalates decisively when a contact shows genuine interest.

When a lead goes cold, Leo remembers. After an appropriate interval, he returns — with a different angle, a different piece of value, a different reason to reconnect. Dead leads are just sleeping leads.

The Leo personality
R

Researches carefully

Before drafting anything, Leo reads the contact's organisation, their role, their stated priorities. The email he sends is relevant to that specific person — not a mass-market pitch with a name swapped in.

W

Drafts with warmth

Leo writes like a human colleague who has done their homework. Not corporate. Not pushy. Warm, relevant, and built from Stella's RAG — so it sounds like it comes from your organisation, because it does.

P

Follows up patiently

Most sales happen on the third or fourth contact. Leo knows this. He follows up at intelligent intervals, never too soon, never forgotten. Each follow-up adds value rather than simply repeating the previous message.

D

Escalates decisively

When a contact replies, clicks, or shows meaningful engagement signals, Leo escalates immediately — routing the hot lead to the right human in your team via Zulip, with full context included.


How Leo runs a campaign

Six steps. One agent. Always running.

Each Leo campaign runs autonomously within these six stages — with a human approval gate before anything is sent.

01 🔍
Contact discovery
Leo scans the web for contacts matching the campaign criteria — job titles, organisations, geographies — and cross-references against existing contacts to find net-new leads.
02 📖
Research & context
For each contact, Leo reads their organisation's website, recent activity, and stated priorities. He queries Stella's RAG for relevant organisational context. He knows why this person might care.
03 ✍️
Draft & approve
Leo drafts the email in your organisation's voice. The draft is sent to the designated manager via Zulip. They approve, edit, or reject. Nothing is sent without human sign-off.
04 📤
Send & monitor
On approval, Leo sends the email and monitors: opens, clicks, replies, bounces. Every signal is logged to the campaign's SQLite contact database.
05 🔄
Follow-up sequence
Non-respondents receive intelligent follow-ups at campaign-defined intervals. Each follow-up is a fresh draft — different angle, same warmth. Cold contacts are revisited after a defined dormancy period.
06
Escalation
When a contact engages — replies, clicks repeatedly, or meets escalation criteria — Leo routes them immediately to the right team member via Zulip with full campaign context attached.

Human in the loop

Leo drafts. You decide. Always.

Leo never sends anything without explicit human approval. Every draft arrives in your team's Zulip channel with the contact's research summary, the draft email, and two simple options: approve or edit.

Edits flow back to Leo. He incorporates feedback, learns the preferred style for that campaign, and applies it to future drafts — getting better the longer he runs.

Approvals can be delegated per campaign. An Erasmus campaign might route to the Managing Director. A junior programme campaign routes to the junior team lead. Each campaign defines its own approval path in the YAML job file.

L
Leo · Draft approval request

#leo-erasmus · campaign: erasmus-coordinators-eu

Leo
New draft ready for approval

Contact: Dr. Katrin Müller, International Coordinator, TU Berlin
Research: TU Berlin runs 340+ Erasmus partnerships. Dr. Müller posted last month about expanding English-language pre-departure programmes.

Draft subject: English preparation for your TU Berlin Erasmus students — from Dublin
Leo · Draft email
Dear Dr. Müller,

I came across your recent post about expanding English preparation for Erasmus students — we do exactly that at Travelling Languages in Dublin, and I thought it might be worth a brief conversation.

We work with coordinators across Europe to provide intensive programmes specifically timed around departure windows...

[View full draft]
✓ Approve & send
✏ Edit draft
✕ Skip

You
Approved. Good research on the post — that's a strong hook.
Leo
Sent ✓ · Monitoring for response · Follow-up scheduled in 7 days if no reply.

Campaign examples

Any outbound goal. One configuration file.

Leo runs any outbound campaign that can be described in a YAML job file. Each campaign has its own contact criteria, messaging strategy, follow-up cadence, and escalation routing. Here are three examples from real deployments.

Campaign type · Institutional partnerships

Erasmus Coordinators

University international offices across Europe

Leo targets Erasmus and international mobility coordinators at European universities — identifying contacts who manage pre-departure language requirements and Erasmus+ programme preparation for outbound students.

Scans university websites and Erasmus network directories
Researches each coordinator's current programme structure
Drafts outreach in English, German, French, Italian, Spanish as appropriate
Follows up twice at 7-day intervals, then dormancy review at 90 days
⚡ Hot leads escalate to: Managing Director
Campaign type · Corporate partnerships

Job Placement Companies

Recruitment and relocation agencies in Dublin

Leo identifies job placement and international recruitment companies operating in Dublin — particularly those placing non-native English speakers who may need pre-employment English support or ongoing workplace language training.

Targets HR and candidate services leads at Dublin recruitment firms
Research focuses on international hiring volume and candidate origin countries
Messaging frames language support as a placement quality differentiator
Follows up at 10-day intervals; re-engages dormant contacts at 60 days
⚡ Hot leads escalate to: Operations & Groups team
Campaign type · Junior programmes

Junior Programme Contacts

Schools, tour operators & youth organisations

Leo targets secondary school language departments, educational tour operators, and youth exchange organisations across Europe — promoting junior English immersion programmes including summer camps and school group packages.

Identifies language teachers and department heads at secondary schools
Targets educational travel agencies with European school group specialisms
Seasonal cadence — campaign intensity peaks in October and February
Re-engagement of previous-year contacts prioritised each autumn
⚡ Hot leads escalate to: Junior programmes team

Configuration

A campaign in fifteen lines.

Each Leo campaign is defined in a simple YAML file. Non-technical managers can read and understand it. The file defines who Leo should contact, how he should approach them, how often to follow up, and who receives hot leads. No code required.

Leo's campaign logic, contact database, and draft history are all stored locally — inside the Stella deployment. No contact data, draft content, or campaign intelligence is ever sent to external services.

New campaigns can be added at any time by dropping a new YAML file into the campaigns folder. Leo picks it up on the next scheduled run.

campaigns/erasmus-coordinators.yaml
campaign: "erasmus-coordinators-eu"
description: "Erasmus coordinators across Europe"

target:
  roles:
    - "Erasmus coordinator"
    - "International mobility officer"
  regions: "EU"
  exclude_existing: true

messaging:
  tone: "warm, professional, relevant"
  rag_context: true # pulls from Stella RAG
  languages: "auto" # matches contact language

cadence:
  initial_wait_days: 0
  follow_ups: 2
  follow_up_interval_days: 7
  dormancy_review_days: 90

approval:
  channel: "#leo-erasmus"
  approver: "@managing-director"

escalation:
  triggers: "reply OR 3x_click"
  route_to: "@managing-director"
  channel: "#hot-leads"
What makes Leo different

Not a mass-mailer. An agent.

The difference between Leo and a standard email automation tool is the difference between a skilled sales professional and a bulk mail list.

Stella RAG inheritance

Leo queries Stella's RAG — the same knowledge base that CLARA and the other modules use — to understand your organisation before drafting. Every email sounds like it comes from someone who knows your school deeply, because it does.

Multilingual by default

Leo detects the contact's likely language from their organisation and country, and drafts in that language where appropriate. An Erasmus coordinator in Valencia receives a different email from one in Munich — not just different names, different languages.

Contact memory across campaigns

Leo's contact database persists across campaigns. A contact who was approached in an Erasmus campaign and went cold will not receive a duplicate outreach from a junior programmes campaign six months later — Leo knows the history.

Dormancy & re-engagement

After the initial sequence, dead leads are not deleted. Leo revisits them after the configured dormancy period with a fresh approach — a different angle, a different value proposition, a different reason to care. Persistence, not spam.

Approval learning

Every edit a manager makes to a Leo draft is a training signal. Over time, Leo learns the preferred style, tone, and framing for each campaign — reducing the editing burden with each cycle.

Full audit trail

Every draft, every approval, every send, every response is logged in the local SQLite contact database. Full campaign history is available for review at any time — and never stored outside your infrastructure.

🔒

Leo runs inside your Stella deployment — contact data never leaves your infrastructure

All contact discovery, draft generation, campaign data, and email history is processed and stored on your own NVIDIA hardware. No contact list, no draft, and no campaign intelligence is ever sent to external AI providers or third-party services. Leo uses Stella's local AI models for all drafting — the same privacy architecture as every other Buongiorno module.

Ready to give Leo a campaign?

Leo is available as a Stella module for any Buongiorno deployment. Get in touch to discuss your first campaign.

leo@buongiorno.ie See the full Stella platform