Artificial intelligence

AI in every module.
Your choice of model.

HOP integrates AI directly into your Sales, Helpdesk, Messaging and Marketing modules. For each task, you choose the provider and model : OpenAI, Mistral, Anthropic, or a self-hosted model.

How it works

AI built in, not bolted on

No separate AI module, no dedicated tab. AI acts where you already work.

Transversal

AI acts in your tickets, opportunities, campaigns, messages, directly in each module's context.

Configurable per task

For each AI task, you choose the provider and model. You can mix : one model for translation, another for analysis, another for drafting.

Explicit

Configuration is visible and documented. You know exactly which model is active on which task.

Features

AI in every module

Concrete functions, in the workflow of every team.

Helpdesk
  • Automatically qualify each incoming ticket : urgency, customer sentiment, product concerned, assigned technician
  • Generate a first response and suggest relevant knowledge base articles
  • Detect duplicate tickets and identify recurring root causes or emerging bugs
  • Translate exchanges and generate FAQs from ticket history
Sales
  • Summarise a client account : history, exchanges, context across all modules
  • Prepare a meeting brief
  • Score leads and detect hot opportunities or at-risk deals
Messaging
  • Summarise a thread : decisions made, actions to take, important requests, relevant team
  • Draft or rephrase a response
  • Translate automatically
Marketing
  • Score contacts by engagement
  • Generate an email or sequence draft with subject suggestion, from a brief
  • Summarise campaign performance in plain language
Configuration

For each AI task, its provider, its model type.

In HOP, AI configuration is granular. For each task, you choose the provider and model among the built-in providers : OpenAI, Anthropic, Mistral, or a self-hosted model for organisations that want to keep data within their own infrastructure. Other providers can be added through connectors.

You can mix : a lightweight model for translation, a more powerful one for ticket analysis or content generation. Configuration changes without touching the code.

You choose, above all, where your data goes. Depending on your constraints, HOP can use a US, European, or self-hosted AI provider running on your own infrastructure. Each AI task can have its own rule: an external model for generic rewriting, a model hosted in Europe — or on your premises — for sensitive data.

  • SalesClient summaryOpenAIUSAdvanced model
  • HelpdeskTranslationMistral AIEUTranslation model
  • SalesLead scoringAnthropicUSAdvanced model
  • HelpdeskTicket qualificationSelf-hostedSelf-hostedSpecialized model

Built-in providers as standard. The list extends through connectors, configured by your integrator. Model type and location are configured independently of the provider.

Principle

An assistant, not a black box

AI in HOP suggests, summarises, detects, drafts. It doesn't decide. Every result is subject to human validation before being applied. What AI produces is a starting point, not a conclusion.

AI suggests. A response, a score, a summary. The result is visible before any action is taken.

The human validates. No change is applied without confirmation. You stay in control.

Configuration is explicit. You know which model is active on which task. Nothing runs in the background without your knowledge.

What AI doesn't do

It never makes a decision automatically.

It never sends a message without confirmation.

It never runs an action in the background without your knowledge.

Governance

AI that you govern

As the AI Act's obligations progressively come into force, AI usage will need to be more transparent, traceable and controlled. HOP is designed with this logic from the ground up : tasks that can be switched on individually, an identifiable provider, systematic human validation and complete logging.

Human validation

No AI-produced result is applied without confirmation from a user.

Explicit configuration

Provider, model and active tasks are documented and reviewable at any time.

Per-task activation

Each AI task can be enabled or disabled independently, module by module.

Logging

AI calls are traced : who, when, which task, which model.

For data protection and hosting, see Advanced privacy and Infrastructure & Sovereignty.

AI that adapts to your organisation, not the other way around

A certified HOP expert configures AI to your business context and your choice of providers.

Deployed by a certified partner · European infrastructure · Your choice of AI provider, including European or self-hosted