New ATLAS for SMEs — up and running in hours, without losing the human touch. See it
AI-powered omnichannel platform

Your customers write across 6 channels.
Your team replies from just one.

Your customers speak. ATLAS listens, responds and resolves.

Unify WhatsApp, Instagram, Email, Telegram, SMS and Web Chat. The AI replies in seconds; your team steps in when it matters.

Native interface in English, Spanish and Catalan.
María García

via WhatsApp · Mortgages Campaign

Hi, I'm interested in the youth mortgages I saw on your website 🏠
Hi María! We have several mortgage options with special conditions for under-35s. Are you already a customer or is this your first enquiry?
It's my first enquiry. I'd like to buy a flat for around €180,000
Perfect. For a property at that price we finance up to 90%. If you'd like, I can run the numbers and check feasibility. Shall we?
✓✓ Read · 14:32
Type a message…
Customer service operator smiling with headset — ATLAS omnichannel AI platform
Built in Barcelona
Data in the EU
AES-256-GCM encryption
Live in hours, not weeks
GDPR · Native consent gate
💬
WhatsApp
📸
Instagram
📧
Email
✈️
Telegram
✉️
SMS
🌐
Web Chat
💬
WhatsApp
📸
Instagram
📧
Email
✈️
Telegram
✉️
SMS
🌐
Web Chat
💬
WhatsApp
📸
Instagram
📧
Email
✈️
Telegram
✉️
SMS
🌐
Web Chat
💬
WhatsApp
📸
Instagram
📧
Email
✈️
Telegram
✉️
SMS
🌐
Web Chat

When your team loses time,
your customer loses patience.

If you run communications across multiple tools, this happens to you. ATLAS solves it from day one.

Customer service agent operating ATLAS with WhatsApp and Instagram conversations in their CRM
🔀

Scattered channels and slow replies

Your team jumps between tools, messages get lost and every minute without a reply is a customer cooling off. Without centralization or automation, you can't scale.

→ Unified inbox + AI that replies in seconds
Read more →
🧩

No pipeline at all

You run leads on spreadsheets. No stages, no Kanban, no real follow-up.

→ Native pipeline + Kanban
Read more →
📊

No real visibility

You don't know who handled what, when, or with what outcome. In other tools auditing is a paid add-on.

→ Full audit log included
Read more →
🔒

Vendor lock-in

You pay per agent, per resolution, per proprietary AI. If you want to switch, migration takes months.

→ Dedicated tenant, your data, your keys
🤖

Single-vendor AI

Tied to a proprietary AI you don't control. You can't switch models or run a local one.

→ AI-agnostic, you pick the model
🧩

No context across channels

The same customer writes via WhatsApp and Email, and no one on your team notices. Every channel is a silo, every conversation starts from zero.

→ 360° contact view with multichannel auto-merge

Everything your team needs.
Nothing it doesn't.

Every ATLAS module is designed so your teams work faster, with better context and less friction.

01

Several tools open at once? One unified inbox.

WhatsApp (self-hosted and official), Instagram, Email, Telegram, SMS and Web Chat converge in a unified inbox. No tab-switching. No context lost. With real-time delivery ticks.

  • WhatsApp official (WABA) and self-hosted (WAHA) — your choice
  • WhatsApp templates approved by Meta, with AI optimizer
  • Bidirectional AI ↔ human handoff without losing context
  • Per-channel health monitor with circuit breaker and retries
  • Multiple accounts per channel · Smart bulk sending

Inbox

All AI Human
MG
María García 14:32
It's my first enquiry. I'd like to buy a flat for…
💬
AI
CR
Carlos Ruiz 14:28
I need to speak to someone from the department of…
📸
Waiting
LP
Laura Pérez 14:15
Perfect, we're all set then. Thanks so much!
📧
Human
AT
Ana Torres 13:45
Do you have any special renewal offers?
✈️
AI
02

Slow replies? The AI answers before your team even reads the message.

Because not all AI models think alike

Tell us your business and we'll tell you which model best understands your needs.

Pick OpenAI, Anthropic, Gemini, xAI, Microsoft Copilot, Perplexity, local models with Ollama or any OpenAI-compatible API. The AI replies, transcribes audio, analyses images and automatically escalates to a human when it detects it can't resolve.

  • Compatible with OpenAI, Anthropic, Gemini, xAI, Microsoft Copilot, Perplexity, Ollama and OpenAI-compatible APIs
  • Built-in prompt simulator and analyser before going live
  • Intent classifier with dozens of deterministic actions
  • Anti-hallucination: if it doesn't know, it says so. If nobody knows, it escalates
  • AI Vision for photos and Whisper for voice notes
AI-Agnostic Engine
You pick the provider. You control the data.
🟢
OpenAI
Active
🟠
Anthropic
Available
🔵
Gemini
Available
xAI (Grok)
Available
🔷
MS Copilot
Available
🟣
Perplexity
Available
🏠
Ollama
Local
⚙️
Custom
OpenAI-compatible
Your API key · Your infrastructure · Your data
No per-resolution fees. No lock-in.
03

Leads in Excel? Pipeline with Kanban that the AI moves on its own.

A pipeline with draggable stages shared across all campaigns and teams. Each team sees its own filtered Kanban view. Stages can be AI-only, Human-only or Shared.

  • AI-only, Human-only or Shared stages with drag-and-drop
  • Native triggers: combine events and actions, no code
  • Campaigns with T0–T3 tiers and configurable closing message
  • Google Calendar integration — the AI books appointments on its own
  • Tasks with 5 types, structured closure and follow-up
  • Dynamic segments that recompute on every query

Pipeline · Mortgages

Team: Sales
New 🤖 AI 3
María García
💬 WhatsApp 5m ago
Under 35
Pedro López
📧 Email 12m ago
Qualified 🤖 AI 2
Laura Sánchez
📸 Instagram 1h ago
VIP
Proposal 👤 Human 1
Carlos Ruiz
💬 WhatsApp 2h ago
€180K
Closed 5
Ana Torres
✅ Signed Yesterday
04

No visibility? Measure every conversation, every operator, every outcome.

Dashboard with key real-time indicators. Full audit log on every action. You know who did what, when, from which IP, and the values before and after each change.

  • Live KPIs: campaigns, operators, teams and pipelines
  • Multi-level escalation between teams with return-to-origin
  • 4 RBAC roles with real-time presence and auto-away
  • Audit log with dozens of actions, charts and anomaly filter
  • Global health monitor with 7 categories and red alarm
Sample data
Dashboard
Conversations today
1,284
↑ 12% vs yesterday
Resolved by AI
73%
↑ 5% vs last week
Avg. reply time
8s
↓ 3s vs yesterday
Operators online
18
of 24 total

AI that actually works.
Not a chatbot on steroids.

ATLAS doesn't just reply to messages. It reads documents, runs business logic, fills in forms and takes deterministic decisions.

📄
ATLAS Cortex
Your AI reads the PDFs your contacts send. It extracts data, classifies and triggers logic automatically.
ATLAS Functions
Define JavaScript functions the AI calls on its own. Built-in AI generator and test runner.
📊
Matrices
Structured tables the AI fills in during the conversation. With completeness tracking and Kanban widget.
🎯
Intent classifier
Dozens of deterministic intents. The AI doesn't improvise: it decides, executes and audits every step.
📚
Knowledge (RAG)
Three-pillar search — lexical, semantic and by tags — with per-product isolation. The AI answers from your catalogue, not from its imagination.
🎭
Archetypes
5 pre-loaded conversational patterns — Closer, Consultant, Concierge, Support, Onboarder. The agent adopts the right tone and momentum for each campaign.

A look inside the AI engine

AI Agent in ATLAS: advanced configuration detail (temperature, scope, RAG, transcription and vision) AI Agent in ATLAS: compact editor view
AI Agent

Configurable down to the last parameter

Base prompt, temperature, max tokens, multi-product scope, knowledge base by tags, audio transcription and image analysis — all editable from the UI. No code, no redeploy, no server touch.

Multi-LLM

OpenAI, Anthropic, Gemini, xAI, Copilot, Perplexity, Ollama… your call

Plug in any of the 8 supported providers (Ollama local and OpenAI-compatible APIs included) and assign a different model to each agent. Claude Sonnet for chat, GPT for vision, Gemini for reasoning. No lock-in: swap models from the UI as better ones land.

LLM providers table in ATLAS: OpenAI, Anthropic, Gemini, xAI, Copilot, Perplexity, Ollama, Custom
Cortex

From chat to structured document

Templates with live fields that fill in during the conversation. Reads external PDFs, extracts data and completes the final document. Example: Spanish cadastre queries in seconds, not days.

Cortex template editor in ATLAS with a live preview of a real Spanish cadastre document
JavaScript function editor for a function the AI calls itself — calculator.js Library of helpers — dates, currencies, formats, validations
Extensible

Helpers, functions and matrices: the agent as code

Three layers that extend the agent without touching SaaS code:

  • Helpers · pure reusable functions (dates, currencies, formats, validations). Available inside any expression.
  • Functions · JavaScript procedures the AI decides to invoke on its own during the flow. They receive the contact's context and return data or actions.
  • Matrices · structured contact memory, accessible from helpers and functions. The agent remembers what it already knows — no reinventing the conversation each turn.

All three work both in the conversational flow and in Cortex vision processes — reading PDFs, validating documents, completing templates.

Built-in editor with JS syntax, test runner and AI generator that writes the code from your description of the problem.

From message to resolution
in four steps

1

The customer reaches out

Via WhatsApp, Instagram, Email, Telegram, SMS or Web Chat. ATLAS identifies the contact and the conversation automatically.

2

The AI replies

The AI agent processes the message — text, audio or image — and generates a reply with the campaign context and the conversation history.

3

It escalates when needed

If the AI detects it can't resolve, it automatically escalates to the best available operator, with all the prior context.

4

End-to-end follow-up

Every conversation moves through the pipeline. It's measured, audited and resolved. Your team sees the outcome in real time.

Connect ATLAS with
the tools you already use

ATLAS exposes four native integration mechanisms — Functions, REST API, Webhooks and n8n — so you can connect any external system. Here are some of the services our customers already integrate:

ATLAS Functions
Native JS code the AI calls directly. No intermediate servers.
🔌
API REST
Endpoint for every ATLAS resource. Sync CRM, BI and data in real time.
📡
Webhooks
Inbound and outbound events. Trigger flows when something changes in ATLAS, or signal ATLAS from outside.
🔗
n8n
Visual workflows with hundreds of ready-made connectors. For complex no-code automations.
CRM
Salesforce
HubSpot
Pipedrive
Zoho CRM
Automation
n8n
Zapier
Webhooks
API REST
Databases and analytics
SQL
Snowflake
BigQuery
Excel / CSV
Quality, voice and surveys
NICE
Verint
SurveyMonkey
Twilio

Empower your human team.
The AI does the heavy lifting.

ATLAS doesn't replace people. It frees them to do what they do best: sell, resolve and build loyalty.

Customer service team

Less time searching, more time resolving

Your operators stop jumping between tools. Everything lands in a single panel: messages, context, history and tags.

73% of enquiries resolved
without human intervention
Professional on the phone

Instant replies, happier customers

The AI replies in seconds, qualifies leads and sets the stage. When your team steps in, they already have full context.

8s average
first-reply time
Team at work

Smart routing, not round-robin

Assignment weighs performance, workload, availability and team affinity. Every conversation lands with the best operator.

more conversations
per operator

Estimates based on internal pilots. Actual results depend on your volume, team and configuration.

Your industry. Your problems. Your solution.

This is what ATLAS does in industries like yours. If yours isn't here, give us a call — chances are it fits too.

Real estate

The customer writes on WhatsApp on Sunday at 10 pm asking for a viewing. Your team sees it on Monday and the lead has cooled off.

ATLAS replies instantly with the catalogue, qualifies the lead and books the viewing in your Google Calendar — without anyone on your team lifting a finger.

📅 Google Calendar · daily slots, holidays and buffer between appointments 📄 Cortex · reads floor plans and PDF deeds 📊 Pipeline · Lead → Viewing → Offer → Closed
Financial recovery and collections

Hundreds of microcredit defaults, each with its own tone and momentum. Your team can't keep up and every day without contact is older debt.

ATLAS runs T0–T3 campaigns with Meta-approved WhatsApp templates, context-aware escalation to the case handler and structured closure of payment promises — all audited contact by contact.

📊 Matrices · payment promises and follow-up 🎯 T0–T3 campaigns · configurable cadence 🔍 AES-256-GCM · contact-by-contact audit log
E-commerce

The customer asks "where's my order?" at 3 am and heads to Amazon before your team opens on Monday.

ATLAS reads the shipping ticket, checks the status in your logistics API and answers 24/7. Recovers abandoned carts with T0–T3 campaigns.

⚡ Functions · the AI calls your logistics API 📥 Unified inbox · WhatsApp + Insta + Email 🎯 T0–T3 campaigns · cart recovery
Academy and training

Two hundred students ask the same things: "when does it start?", "how much is it?", "are there discounts?". Your team replies the same answer a hundred times.

The AI handles 80% of enquiries with your course catalogue loaded into RAG. It qualifies students with Matrices and only escalates what really matters.

📚 RAG · the AI knows your entire catalogue 📊 Matrices · qualify age, level, interests 🤝 Bidirectional handoff to humans
Professional services

Consulting, agency, firm. You reply from a personal WhatsApp and all the context is lost when someone goes on holiday.

ATLAS centralises every conversation, every contact keeps its 360° history, and the AI replies in your tone when your team isn't available.

🗂️ 360° CRM · every thread, every decision 🤖 AI agent configurable to your tone 👥 4 RBAC roles · granular permissions

Your industry not here? Chances are we cover it too.

Tell us your case

Start with everything. Scale as you grow.

BASE PLAN — Operational from day one

Everything you need to run omnichannel communication with AI, with no mandatory add-ons.

💬 Text channels

WhatsApp, Instagram, Email, Telegram, SMS and Web Chat. Multiple accounts per channel.

📥 Inbox + Real-time chat

Conversation states, delivery ticks, transfer, escalate, pause, reopen.

🤖 1 configurable AI agent

Audio transcription included, automatic handoff to humans.

📊 CRM + Campaigns + Pipeline

Multichannel contacts, tags, unlimited campaigns, native Kanban.

👥 Teams and roles

4 roles, granular permissions, real-time presence, supervisor panel.

🔐 Security + Audit

AES-256-GCM encryption, full audit log, per-channel health monitor with circuit breaker, 100+ configurable settings.

Scale when you need to

🧠

Advanced AI

Multiple providers, multiple agents per campaign, local models (Ollama), image analysis.

Pipelines and Automation

Multiple pipelines, automatic triggers, native n8n integration, saved views.

📈

Analytics and Compliance

Advanced analytics, SLA thresholds, data export, scheduled reports.

📞

Voice

Inbound and outbound calls, smart IVR, real-time transcription, recording.

Coming soon

Book now and lock in exclusive launch pricing.

Reserve launch price →

ATLAS vs. the rest

The features others charge as add-ons or reserve for enterprise plans, ATLAS includes from the base.

Feature ATLAS Zendesk Intercom Chatwoot
Multi-provider AI ✓ Yes ✗ Proprietary ✗ Fin only ~ OpenAI
Local models (Ollama) ✓ Yes ✗ No ✗ No ~ Partial
Bring your own AI key ✓ Yes ✗ No ✗ No ✓ Yes
Native Pipeline / Kanban ✓ Native ✗ Third-party ✗ No ✗ No
Outbound bulk campaigns ✓ Native ~ Paid add-on ✓ Yes ~ Partial
Smart routing ✓ Multi-factor ~ Enterprise ~ Rules ~ Capacity
Full audit log ✓ Included ~ Enterprise ~ Partial ~ Enterprise
PDF reading and structured extraction ✓ Native Cortex ✗ No ✗ No ✗ No
Bidirectional AI ↔ human handoff ✓ Bidirectional ~ One-way ~ One-way ✗ No
Per-tenant instance and data isolation ✓ Dedicated Docker + DB ✗ Multi-tenant ✗ Multi-tenant ~ Self-host only
WhatsApp by QR (no BSP) ✓ Native WAHA ✗ BSP only ✗ BSP only ✗ BSP only

Clear answers

General

ATLAS is an intelligent omnichannel platform that unifies WhatsApp, Instagram, Email, Telegram, SMS and Web Chat in a single panel. It combines automated AI handling with human operators for sales, support and any organized communication operation.

Yes. ATLAS is delivered as managed SaaS with per-tenant isolation: each tenant runs in its own Docker container with its own database, on European infrastructure (DigitalOcean EU) operated by IMHOIT. Encryption at rest, full audit log, granular access control and the option to run local AI models make GDPR and Spanish LOPDGDD compliance straightforward.

No. ATLAS is designed so any customer service, sales or support team can operate it from day one. The interface is intuitive, configuration happens in the admin panel without touching code, and connecting channels like WhatsApp is done by scanning a QR.

No. ATLAS is a web platform. You access it from any modern browser (Chrome, Firefox, Edge, Safari) without installing anything. You only need an internet connection and your login credentials.

ATLAS is a responsive web app that runs on any browser-capable device. It doesn't require download from app stores. On mobile and tablet it adapts to screen size so operators can manage conversations from anywhere.

Usage and features

Yes. ATLAS includes its own CRM with multichannel contacts, but it also integrates with external CRMs like Salesforce, HubSpot, Pipedrive and Zoho via its REST API and native n8n integration. You can sync contacts, pipeline stages and conversation data between both systems.

Yes. ATLAS lets you configure AI agents that reply automatically with campaign context and the contact's history. The base version includes 1 configurable AI agent. With the Advanced AI add-on you can create multiple agents per campaign, each with its own instructions, tone and behaviour.

Yes. AI agents automatically handle frequent replies, lead qualification and repetitive enquiries. Additionally, with the Pipelines and Automation add-on, you can configure automatic triggers by pipeline stage: notify the team, assign operators, send emails or run n8n flows when a contact advances or changes state.

The AI evaluates each reply against the instructions defined in the agent. When it detects it can't resolve the enquiry — due to complexity, customer request or configured rules — it changes the conversation state to "waiting for human" and notifies the team with all the prior context.

Yes. ATLAS is AI-provider agnostic. The base version includes one configurable provider. With the Advanced AI add-on you can use OpenAI, Anthropic, Gemini, xAI (Grok), Microsoft Copilot, Perplexity, Ollama (local models) and any OpenAI-compatible API, without sending data outside your infrastructure.

A traditional chatbot follows predefined flows with fixed replies. The ATLAS AI agent is different: it understands conversation context, accesses the contact's history, adapts its tone to campaign instructions and generates dynamic replies. It can qualify leads, resolve complex enquiries and escalate to a human when it detects it can't resolve on its own.

ATLAS includes a supervisor panel with real-time presence: who's online, how many active conversations each operator has, reply times and resolution rate. The supervisor can step into any conversation, transfer it or escalate it. The audit log automatically records every action with date, time and IP.

ATLAS allows multiple accounts per channel with no platform-imposed limit. You can have one WhatsApp line for sales, another for support and another for campaigns, each with its own assigned team and campaign. The same applies to Instagram, Email, Telegram, SMS and Web Chat. The number of lines depends on your infrastructure plan, not on ATLAS restrictions.

Yes. ATLAS includes smart bulk sending with priority queue, rate control, scheduled delivery and automatic retries. You can segment contacts by tags, campaign or pipeline stage, and schedule sends to maximize delivery rate without saturating the line.

Pricing and costs

ATLAS supports WhatsApp in two modes: WAHA (QR-based connection, no Meta fees) and Meta's official API (WABA — WhatsApp Business Platform). If you use the official API, Meta charges per conversation according to its category-based pricing (marketing, utility, service, authentication) — ATLAS adds no markup on those costs. With WAHA there are no additional Meta fees, though the connection is subject to WhatsApp's terms of use.

No. All conversation history, actions and contact data belong to the platform, not the operator. When an operator leaves, the administrator can suspend their account and reassign their active conversations. The full history, including the audit of their actions, remains intact and accessible.

Security and data ownership

Strictly speaking, neither. ATLAS is managed SaaS with per-tenant isolation: each tenant runs in its own Docker container, with its own database, on European infrastructure (DigitalOcean EU) operated by IMHOIT. You don't share runtime or schema with other customers — your data isn't mixed with anyone else's in a multi-tenant database. For organizations with own-cloud or strict on-premise requirements, we evaluate custom installations on request.

You. Always. Every ATLAS customer runs on a dedicated instance — its own Docker container and database, physically and logically isolated from the rest — on European infrastructure operated by IMHOIT. The data is yours: exportable and deletable on demand, and never resold or shared with third parties. For organizations with special deployment requirements (own cloud or strict on-premise), we evaluate custom installations on request.

No. Your phone number remains yours. ATLAS connects your WhatsApp line via QR pairing (self-hosted) or through Meta's official API, but in neither case does it transfer ownership of the number. If you decide to stop using ATLAS, your number and your contacts remain entirely yours.

Support and Onboarding

ATLAS includes technical support by email and chat. During onboarding, the IMHOIT team walks you through initial setup: channel connection, campaign creation, AI agent configuration and team training. The goal is to have you live in hours, not weeks.

Yes. The Web Chat widget includes a native consent gate: it shows a mandatory disclosure — configurable text and link to your privacy policy — before capturing the visitor's first message. Acceptance is audited with timestamp and IP. The gate can be enabled or disabled per widget from the admin panel.

Yes. ATLAS keeps a 360° profile per contact with multiple identities: phone, email, Instagram, etc. When an anonymous visitor identifies themselves on any channel, ATLAS automatically detects and merges the prior history. The operator always sees the full conversation, no matter which channel each message came in on.

Ready to see ATLAS in action?

25 minutes with our team. No commitment. We'll show you how ATLAS fits your real operation.

Let's talk

Prefer to write to us?

Tell us your use case, your current channels or the pain you want to solve. We reply in under 24 business hours.

[email protected]