AI agents for sales, marketing and admin · Barcelona · Milan

We build robots
that sell.

Autonomous AI agents that run your revenue playbook end to end, with your people in the lead. Your context, your rules, your infrastructure.

Book a human Talk with Glen
A Selling at Scale robot, line art
Built by sales engineers, around your history

Designed by people who ran revenue at Google and Meta. Your context, your data and what has already worked for you sit at the core.

Writes and speaks like your team

Every email, message and call in your own tone of voice. No AI tells, nothing a buyer recognises as generated.

WhatsApp ready

Reaches your reps and field agents where they already are. A brief every morning, a report every night, nothing to install.

Robots at work for
myPOSAssogiocattolidcaCareerOSSovrano AImyPOSAssogiocattolidcaCareerOSSovrano AImyPOSAssogiocattolidcaCareerOSSovrano AI
Why robots, why us

One team of robots. The whole revenue funnel.

Built by people who ran demand, sales and expansion at Google and Meta, and now encode that playbook into robots that work inside your business, on your terms.

Cover the whole revenue funnel

One robot per stage, from the first campaign to the renewal. No stage left to a spreadsheet, no handoff that quietly dies.

  • demand·Alex reads the market and builds the marketing campaigns that fill the top of the funnel
  • meetings·Glen prospects, qualifies and books
  • expansion·Sophie renews, upsells and catches churn early
  • admin·Rebecca keeps calls, inbox and calendar off your plate

Customised build

No two deployments are the same. Your robots are built on what your best people know, then keep learning from what your team decides.

  • context·your ICP, segments, intent signals, competitors, win patterns and objections
  • playbook·your process, encoded as rules the robot follows
  • database·leads found and enriched from the source that fits your target. Your CRM stays current without anyone typing
  • tools·the stack you already run, nothing to migrate
  • tone·each rep's own voice, learned from what they send

Total sovereignty

Your context layer belongs to you and travels with you. Change model, change host, keep everything the robots know.

  • deploy·on-prem, your cloud tenancy, or ours
  • model·agnostic by design, swapped when a better one ships
  • portable·your context and playbook, never locked to a vendor
  • regulated·built with banks, insurers and law firms in mind
The roster · AI agents

Meet your first four hires

Four AI agents. They know your context, run your playbook and report to your people. Your team talks to them through our web UI, Slack, or WhatsApp for the people on the road.

Glen, the meetings engine robot, wearing a headset

Glen AI SDR, meetings engine

Finds your best prospects, qualifies every inbound lead, and runs multi-touch sequences across email, LinkedIn, WhatsApp and voice until there is a qualified meeting on the calendar. Works 24/7. Never misses a lead.

At myPOS Italia: 2.8x more contacts a day and 30 hours a week back for every SDR. AEs spend their time closing, not prospecting.
Full specs
Sophie, the revenue expansion robot, holding a coffee

Sophie Revenue expansion

Monitors every customer for adoption signals and expansion triggers. Spots cross-sell and upsell, flags churn risk early, and triggers renewals at the right time. Your NRR growth engine.

Churn risk flagged while there is still time to act. Renewals triggered on time. Account managers spend their week on relationships instead of spreadsheets.
Full specs
Alex, the demand generation robot, holding a megaphone

Alex Demand generation

Analyses your market, competitors and trends. Generates campaign options, scores them, and builds the winners end to end: strategy, creatives, distribution. Your marketing department in a box, with your marketer holding the pen.

More campaigns tested every quarter, no agency fees, and every qualified lead flows straight into Glen.
Full specs
Rebecca, the personal assistant robot, with a desk phone

Rebecca Personal assistant

The personal assistant for lawyers, accountants and notaries. Email, PEC and WhatsApp in one place, agenda and reminders, notes taken on your premises, and one AI receptionist connected to all your tools.

Every call answered, day and night. Hours of admin handed back to every partner each week. Runs on your own hardware if you want it to.
Full specs
Glen for the people on the road

Your agents already live on WhatsApp.
Now so does your pipeline.

Designed for the field rep and the independent, multi-mandate agents who sell for you without being on your payroll. You can't manage them through your CRM, and they won't install your app. Glen briefs every rep at 07:30, preps every meeting, and writes the CRM at night. Two taps per item. Nothing to install.

07:30 · the morning brief, confirmed in one tap
G
Glen · Aurum Pay
Business account · online
Tuesday 22 September
Good morning Luca. Napoli and Caserta today, 7 items.
Meetings
10:45 Gioielleria Esposito, Via Toledo 118. Goal: signature on 1 terminal. Open point: Amex fee.
15:30 Ottica Russo, Caserta. Goal: demo. Decides: Sig.ra Russo, not the son.
Calls
09:00 Bar Centrale, quote sent 15 Sep, silent since
09:20 Farmacia Sant'Anna, renewal, contract ends 30 Sep
12:00 Panificio De Luca, new lead from the September campaign
Before 10:00
Send Esposito the Amex fee sheet (I drafted it, one tap below)
Chase Ottica Bianchi for the signed form
Drive Napoli → Caserta 38 km, leave by 14:40.
07:30
Confirm plan 07:41
Fee sheet sent to Esposito from the company line. I'll ping you 30 min before each meeting. Write or send a voice note any time. 07:41
11:58 · a voice note becomes a CRM write the rep approves
G
Glen · Aurum Pay
Business account · online
L
0:07 / 0:19 11:58
"Firmato, un terminale, installazione lunedì prossimo. Ne vuole un secondo per il negozio di Chiaia a ottobre, stesso prezzo."
Here's what I'll write:
Deal Esposito Toledo
Won, 1 terminal, €[value]
Installation
Mon 29 Sep, ticket to ops
New deal
Esposito Chiaia, 1 terminal, Oct
Next step
Call 6 Oct, fix the Chiaia date
Meeting note
Amex 1.6% accepted, no discount asked
11:58
Save it 12:01
Saved. Ops has ticket #4471 and will confirm the slot with Esposito. Next up: 12:10 call with Panificio De Luca, they asked about split payments on the campaign form. 12:01
18:49 · day closed, CRM updated, tomorrow queued
G
Glen · Aurum Pay
Business account · online
Follow up later, I'll say when 18:35
When? A day is enough, I'll pick the slot. 18:35
friday morning 18:35
Fri 25 Sep 09:30, Bar Centrale, it's in Friday's brief.
2 of 6 · 09:20 call, Farmacia Sant'Anna. Renewal, contract ends 30 Sep. 18:35
4 items later
Day closed. 6 of 6 confirmed in 5 minutes.
Done 4 · Moved 1 · Skipped 1
Won: Esposito Toledo, 1 terminal, install 29 Sep
New: Esposito Chiaia, October
Booked: Bar Centrale Fri 09:30, Farmacia Sant'Anna Thu 11:00
Moved: Ottica Russo demo to Thu 15:30, Sig.ra Russo away
Skipped: Ottica Bianchi chase, back on tomorrow's list
CRM updated: 7 activities, 2 deals, nothing left for you to type. Giulia gets the same summary in her weekly roll-up, not this chat.
Tomorrow: 1 meeting, 4 calls. Brief at 07:30.
18:49
grazie, a domani 18:50

One Glen per mandate

Each company deploys its own Glen on its own WhatsApp Business number, inside its own deployment. Your Glen never sees another company's pipeline. The rep's phone is the only place the mandates meet.

The rep leads, Glen types

Every CRM write is proposed first and saved on the rep's tap. Discounts and pricing authority never sit with Glen: it names the human who holds them and opens the request.

What you see

Not the chat. CRM activities and deals updated the same evening, a weekly roll-up per agent, and exceptions: items open two days running, meetings pushed twice, discounts requested.

Business-initiated messages are approved WhatsApp templates; everything else runs in the 24-hour conversation window. Reps write and speak in their own language, Glen answers in it. See the full day, seven screens.

How we work with your people

Human in the lead

Most AI vendors sell human-in-the-loop, and turn your best people into a queue of approvals. We build for human-in-the-lead: the robot does the work at machine volume, your person sets the direction and decides what ships.

the robot proposesthe robot executes the human directsthe system learns Humanin the lead

The robot proposes

Researches, drafts, sequences, scores, forecasts, follows up. Actions completed, not just suggested.

The human directs

Approves, edits, kills or changes the rules in one tap, from our UI, Slack or WhatsApp.

The robot executes

Machine speed, 24/7, every channel, no fatigue, no forgotten follow-up.

The system learns

Every decision your team makes teaches the Sales OS what success means in your context.

Nobody gets replaced. Your SDR stops copy-pasting and starts running a pipeline engine. Your AE stops preparing and starts closing. Your marketer stops producing and starts directing.
The Sales OS moat

Context beats model

Models are becoming a commodity. What the robot knows about your business, and the method it runs, are not. Three layers, autonomous, proactive, and sharper after every interaction.

SALES OSCONTEXTYOUR TOOLS

Sales OS

Closes the rules engine gap. Twenty years of running revenue at Google, Meta and papernest, codified into an execution layer. Not a playbook we read. A playbook we ran.

Context-aware

Closes the context and live data gap. We reverse-engineer your data to extract your winning patterns. The robot reasons over your ICP, your discovery framework and your win patterns, and reads your systems live.

Integration and orchestration

Closes the integration gap. Your existing tools, consolidated under one robot. If you are missing any, they come included.

The self-learning loop, why the moat compounds. Layer one took twenty years to build and it is not static. Every deployment generates data that makes the Sales OS smarter: which emails convert, which qualification questions predict a close, which deal patterns signal a stall. That flows back into the method.
Deployment

Sovereignty is a buying criterion

Enterprises want the efficiency without handing their institutional knowledge to a frontier lab through a general-purpose tool. Our answer is structural, not a privacy policy: the context layer is yours and portable, and you choose where the robots run. The reasoning layer is model-agnostic, so no provider can hold your revenue infrastructure hostage.

On-premInside your network, on your hardware. The strongest option for banks, insurers, law firms and anyone with a regulator to answer to.strongest for regulated clients
Your cloud tenancyYour account, your region, your access controls, your audit trail. We deploy into it and never hold your keys.
Our cloudThe fastest start. Same robots, same context layer, hosted by us in the EU. Move to your own tenancy later without rebuilding.
AI CRM

Pay only for the CRM seats you actually need.

Your CRM stays your system of record. Your reps work in Glen's console or on WhatsApp, and Glen writes every activity, contact and deal to your CRM through the API, with ownership and attribution intact. Managers keep the seats that need reporting and forecasting. Reps don't need one to work the pipeline.

Your repsGlen's console on the web, or WhatsApp on the road. Queues, approvals, replies, one screen per rep.
→
GlenEnriches, scores, drafts, sequences, logs. Applies your rules: territories, directives, duplicate checks, ownership.
→
Your CRMThe system of record. Every write attributed, every sequence enrolled after the write is verified. Your managers report from here.
Glen's SDR console, as deployed at myPOS Italia Live productOpen full screen
This console is built for a desktop screen. Open it in a new tab.

Sequences, replies, records to verify, directives per SDR, statistics. Italian, because the team is Italian. Every approval writes to the CRM first, then enrolls the sequence, so nothing goes out with a stale field.

Sequences

Every channel in one sequence. One approval.

Glen builds the sequence per prospect: email, LinkedIn InMail, WhatsApp or SMS, and a voice call, each written from your context layer in your rep's own tone. The rep approves once. Channels that belong to a person stay with that person. InMails and AI calls included

Imola Gioielli · Marco Grossi, ownerhot · Villaricca (NA) · 5 steps · 11 days
Day 009:10
Email · fees and instant settlementFrom Maddalena's mailbox. Opens tracked, reply routed to her queue.
Glen sends
Day 2morning
LinkedIn InMail · same argument, 40 wordsDrafted by Glen, sent by Maddalena from her own account in one tap.
Rep sends
Day 411:00
WhatsApp · approved template, opt-in on fileFalls back to SMS if the number is not on WhatsApp.
Glen sends
Day 616:30
Call · Glen, 90 seconds, books the visitIntroduced as Maddalena's assistant. Transcript and outcome to the CRM.
Glen calls
Day 909:10
Email · break-up, one lineStops the moment any channel gets a reply.
Glen sends
step 2 · LinkedIn InMail · preview
Buongiorno Marco, le ho scritto martedì sulle commissioni. In una parola: con myPOS incassa il giorno stesso e paga meno su ogni transazione. Quindici minuti questa settimana per farle vedere i numeri sul suo volume? Maddalena
why this step

In this campaign, retail jewellers in Campania answered on LinkedIn more often than to a second email. Glen moved the InMail to step 2 for the segment.

Approve sequenceEdit stepSkip LinkedIn
EmailSent from the rep's own mailbox, warmed and monitored. Replies land in the rep's queue with Glen's reading of them.
LinkedInGlen drafts, the rep sends from their own account. No automation on LinkedIn's side, no risk to the rep's profile.
WhatsApp and SMSApproved templates for the first touch, opt-in recorded, free-form replies inside the conversation window.
VoiceA call from Glen, introduced as the rep's assistant, disclosed as an AI where the law requires, with the transcript written back to the CRM.

Sounds like your rep. Reads like a person wrote it.

what the sender wrotethe playbook WriterEditor draftAI tells flagged · rewrite round 1 · 2 · 3 passes sounds like the senderno AI tells
No AI tellsTwo agents work every message. A writer drafts it to the rules of the best playbooks in the business. An editor hunts for anything a buyer would spot as generated: filler, stock phrases, template rhythm. The draft goes back and forth until the editor finds nothing.
One tone per repGlen learns each rep's tone from what they actually send: length, greeting, formality, the words they use and the ones they never would. Maddalena's messages read like Maddalena's, Riccardo's like Riccardo's.
Your rules, every timeWhat Glen may say comes from your context layer: pricing, claims, positioning, what won last quarter. Approve once, and every message on every channel follows the same rules.

Test every angle at once. Keep what wins.

Glen Ashort & informalBlonger, with bulletsCopens with a questionDcase study attachedlive scorereplies · meetingsABCleadingDyour companyby productby audience
Many tests, in parallelShort and informal. Longer, with bullets. A question up front. A case study attached. Glen runs every variant at the same time, on every channel, instead of one A/B test a quarter.
A live scoreEvery reply and every meeting booked counts the moment it happens. You see which message is winning today, not in a report next month.
What works for youScores split by product and by audience, because what lands with a CFO won't land with a store owner. You pick the winner; Glen makes it the default and keeps testing.
Getting started

Two paths to deployment

Forward deployed GTM engineers

We come to you

A GTM engineer embeds with your revenue team for two to three weeks. They interview your top performers, extract the winning playbook, map your ICP and objection handling, and codify all of it into the context layer. They wire up your CRM, enrichment, calendar and comms. When they leave, the robots know your business the way your best sales leader does, and they never forget.

  • Playbook extraction: your best rep's knowledge, codified
  • Live integration: your CRM, enrichment, calendar, Slack
  • Handoff in 90 days: robots operational, engineer moves on

Self-serve onboarding

No engineer needed, no CRM change

A guided flow that collects your business context in under an hour: positioning, ideal customer, pricing, competitors, sales process. The answers are codified into the context layer automatically, and the robots connect to your existing tools with the keys you provide.

  • 60-minute setup: answer questions, the robots learn your business
  • Zero migration: works with whatever tools you already use
  • Scales down: from a two-person firm to a Series A startup
Pricing

Priced like headcount. Paid like a result.

Tiered monthly pricing for every robot, plus a fee per outcome the robot delivers.

RobotWhat it doesSetupMonthlyPer outcome
Glen
Books qualified meetings across email, LinkedIn, WhatsApp and voice€3,000€1,500up to 5 repsDefined per meeting and sale
Sophie
Renewals, upsell and churn signals across your customer base€3,000€1,500up to 5 account managersDefined per renewal and expansion
Alex
Campaigns built end to end, qualified leads handed to Glen€3,000€1,500up to 5 usersDefined per marketing-qualified lead
Rebecca
Communication, agenda, notes and reception for the whole practice€3,000€1,500up to 5 lawyersDefined per complex answer

Setup covers discovery, consulting, build, onboarding and training, so the robot arrives knowing your business. Monitoring, support, training and updates are included in the monthly price.

The maths: five SDRs, one Glen

The question is not what a robot costs against a salary. It is what a qualified meeting costs you today, and what it costs when the same five people get Glen.

MetricYour 5 SDRs todayYour 5 SDRs with Glen
Fully loaded cost€21,000 / month€21,000 + €1,500, plus the per-meeting fee
Leads processed per month750 - 1,0002,500 - 7,500
Response time2 - 24 hoursUnder 5 minutes
CoverageBusiness hours, five people168 hours a week, every rep
Qualified meetings per month50 - 75175 - 250
Time each SDR spends calling, meeting in person and closingAbout 30% of the weekClose to all of it, 70% more
Cost per qualified opportunity€280 - 420€90 - 130
Same five people, three to four times the meetings, each at a fraction of today's cost. And the part that compounds: Glen takes the research, the writing and the CRM off each SDR, so every one of them gets about 70% of the week back to call, meet in person and close.
Case studies

Results, told by the people who run them.

Glen, AI SDR

myPOS Italia

Retail merchant acquisition, Campania
Glen gives our SDRs superpowers, and keeps them at the center of the sales motion. It hands them back 30 hours a week of segmentation, prospecting and analysis, time they now spend on the phone, building a human, consultative relationship with every merchant.
Luca CortiHead of Partnership & Alliances, myPOS
Alessandro BoccaCountry Manager, myPOS
Read the full case study →
73.2%
of SDR time saved on manual work
researchpreparationoutreach
2.1x
leads processed per day
2.8x
contacts made per day
30h
back per SDR, every week, for calls and meetings
About us

Sales engineers, not AI consultants

Built by the people who scaled Google, Meta and papernest, now encoding twenty years of playbook into robots.

GoogleMetapapernestColgate
Marco Grossi, line-art illustration

Marco Grossi

CEO, revenue architect
  • Scaled Google and Meta sales teams to $1B+ in managed revenue, including five years leading Meta's SMB team for EMEA
  • As CBO at papernest, drove 3x revenue in two and a half years, the company's first positive EBITDA in ten years, and managed 200+ people
  • Previously VP Sales at Stocard, exited to Klarna
Federico Corradini, line-art illustration

Federico Corradini

CBO, commercial strategy
  • Ex Colgate-Palmolive, Bolton and Google, where he led the Italian team to first for growth in EMEA and second globally
  • Founder and CEO of XChannel, a €3M+ ARR marketing agency serving 100+ SMB and mid-market clients
  • Professor at Politecnico di Milano, marketing and AI
Javier Solorzano Zavala, line-art illustration

Javier Solorzano Zavala

CTO
  • 25+ years across software, hardware, IoT and cloud, with CTO seats at Risk Factor, AVUXI and OneMind Technologies
  • Designed on-premises and cloud architectures for regulated and security-sensitive clients, with strict data sovereignty requirements
  • Deployed AI and ML systems for SMBs and enterprise

Deep SMB and mid-market expertise from the commercial side, deep regulated-infrastructure expertise from the technical side. We built these systems at scale. Now we are encoding them into robots.

FAQ

Questions buyers ask us

What does Selling at Scale do?

Selling at Scale, S.L. is a Barcelona company, with a team in Milan, that builds and deploys AI agents for revenue teams. We call them robots. Glen books qualified meetings, Sophie runs renewals and upsell, Alex builds marketing campaigns, and Rebecca is the assistant for law firms, accountants and notaries. Your people stay in the lead: the robots do the volume work, your team sets direction and approves what ships.

What is an AI SDR, and what does Glen do?

An AI SDR is software that does the work of a sales development rep: finding prospects, qualifying leads, writing outreach and booking meetings. Glen does it across email, LinkedIn, WhatsApp and voice, in each rep's own tone, and writes every activity to your CRM. At myPOS Italia, Glen gives every SDR about 30 hours a week back for calls and meetings.

How much does it cost?

Every robot has a one-off setup fee of €3,000 and a monthly fee of €1,500 for up to five users, plus a fee per outcome agreed per deal: per meeting and sale for Glen, per renewal and expansion for Sophie, per marketing-qualified lead for Alex, per complex answer for Rebecca. Monitoring, support, training and updates are included in the monthly fee.

Where do the robots run, and who owns the data?

You choose: on your own hardware on-premises, in your cloud tenancy, or in our EU cloud. The context layer, meaning your playbook, your rules and what the robots learn about your business, belongs to you and moves with you. The reasoning layer is model-agnostic, so you are never tied to one AI provider.

Does it work with our CRM and tools?

Yes. The robots connect to the stack you already run through APIs, starting with your CRM, calendar, email and messaging. Your CRM stays the system of record: every contact, activity and deal the robots touch is written back to it with ownership intact, so reps do not need a CRM seat to work the pipeline.

How long does it take to go live?

There are two paths. A forward deployed GTM engineer embeds with your team for two to three weeks, extracts your playbook and wires up your tools, and the robots are fully operational within 90 days. The self-serve onboarding collects your business context in about an hour and connects to your tools with the keys you provide.

Will the robots replace our sales team?

No. They take research, writing, data entry and follow-up off your people, so the team spends its week on calls, meetings and closing. A person always sets direction and approves what goes out.

Is it GDPR compliant?

Deployments are designed around GDPR from the first workshop. You choose where the robots run and where data is stored, the context layer stays yours, and a data processing agreement is part of every contract. Banks, insurers and law firms usually choose the on-premises deployment, which keeps the robots inside their own network.

Which languages and markets do you cover?

We are live in Spain and Italy, with France and Germany next. The robots write in the language of each prospect and each rep, and our team works in Spanish, Italian, English and Portuguese.

Next step

Book a human

Glen will happily talk to you first. If you want one of us instead, book an AI Revenue Check-Up: one hour, no pitch, clarity on which robots would work for you and what they would return.

  1. We map your current stack and team structure (15 min)
  2. We identify integration gaps, manual work and unrecovered capacity (15 min)
  3. We show you which robots fit, with the ROI maths on your numbers (30 min)
Book a humanTalk with Glen

Direct contact

Email: hello@sellingatscale.com

LinkedIn: Selling at Scale

Spanish, Italian, English, Portuguese. We reply the same working day.