HORIZON NORD
FR Discuss your operationsLet's talk
AI & automation

AI agents that prepare the work.

A project update to assemble, invoices to check or records to find: we configure useful tasks in monday.com. You define access and approval points. Your team reviews what the agent prepares.

Discuss your operations

Bernard, Diane, Nora and Sam, Horizon Nord's four AI agents

AI that preps the work, in your boards.

This example shows one possible setup: a delay is identified, its impact is prepared for the project manager, and the people affected are notified after approval. The tools and steps are selected around your needs.

Review construction use cases, guardrails and limits

A monday.com workflow where an AI agent detects a delay, gets project manager approval and notifies subcontractors
monday workflows orchestrates the steps. monday Agents analyzes the impact, prepares the decision and sends notices after approval.
Sidekick

monday.com's built-in AI assistant, day to day: ask a question about your boards, it answers with your data. It summarizes, drafts, and helps each person in their daily work.

AI Agents

Complete sequences: read, classify, link to the right project, prepare the follow-up. Agents configured on your rules and your processes, not on generic templates.

monday workflows

Automation with AI blocks: they read your documents, pull the information out of your invoices, contracts and forms, then file it in the right places, with no manual re-entry.

Claude, Anthropic's AI

Claude connects directly into monday.com: it analyzes your boards, writes summaries and prepares files from your real data. You approve before anything goes out.

With Direction by Horizon, OCR by Horizon extracts supplier invoice data for review in monday.com. Your team checks the amounts and allocation before approval.

A crew of AI agents, configured for you.

Four example roles, built in monday.com around your operations. They prepare the work. You keep the decisions.

Field assistantAssembles the morning briefing
Bernard, AI field assistant agent
Bernard
Accounts payableLinks invoices to projects
Diane, AI accounts payable agent
Diane
DocumentationSpots missing pieces
Nora, AI documentation agent
Nora
BidsAssembles the files
Sam, AI bids agent
Sam
Your agents, custom-builtConfigured for your operations
Not four rigid products
Bernard · Field assistant

Gathers reports, blockers and follow-ups; the daily briefing is assembled in monday.com before you walk in. Your team decides: priorities and commitments.

Diane · Accounts payable

Gathers invoices, links each one to the right project, then prepares the approval requests. Your team decides: approve, send back, or release payment.

Nora · Documentation

Spots missing documents and keeps every project file complete and accessible. Your team decides: what goes out to the client or the inspector.

Sam · Bids

Gathers the context and assembles the bid files. Your team decides: the price and the send.

Four examples, not a catalog: your agents are configured on your operations, with the names your team recognizes.

Explore Direction by Horizon for construction.

See the simple AI agent pilot offer

The limits, in black and white.

AI becomes useful when it knows your system, its role, and the moment your team must decide. This is what we write into every mandate, before any configuration.

Agents prepare. You decide.

Commitments, payments and important messages have a responsible owner. We define the approval points with you and test those limits before go-live.

AI sees only what you allow.

Every agent has a precise mandate: the boards and documents it can access are listed and documented, board by board.

Access and processing are documented.

The selected tools determine where information is processed. We specify the data shared, services used and permissions required before go-live.

Defined validation criteria.

We select a task, sample records and the checks required. The errors and exceptions observed help determine the adjustments needed before production.

Simple first.

A fixed rule may be enough to assign a task or flag a deadline. We use AI for steps that require reading, classifying or preparing information.

What you want to know before calling.

Your question isn't here? You'll talk to the person who would do the work at your company.

Discuss your operations

Is AI mandatory in an implementation?

No. A monday.com implementation can be simple, with no AI at all. Agents come in only where they create concrete value, and only if you want them.

Who decides, in the end?

People, always. Agents gather context, link information and prepare follow-ups. Every commitment, approval or important send stays with your team.

How does it work, concretely?

Sidekick assists your teams day to day, AI Agents run complete sequences, the AI blocks inside monday workflows read your documents, and Claude, Anthropic's AI, connects directly into monday.com. Everything is configured on your real files, with your real words, before go-live.

Is our data safe with AI?

Data access and processing depend on the tools selected: monday.com AI features, agents or external services. We specify accessible boards, information shared and human review steps before go-live.

Last stop

Your system should give you
real direction.

Book your time slot: 45 minutes, no commitment. Then:

You talk to the person who would do the work. In English or French.
You leave with the first process where an agent would create value. Follow-ups, invoices or documents: we name the starting point, whether you continue with us or not.

Not ready to book? Write to us: a real person reads your message within one business day.

monday.com certified partner

If the calendar does not load, open the booking page.

Prefer to write? Tell us where information gets lost or where follow-ups slow down: a real person reads your message and replies within one business day, in English or French.

Optional: to get a text about your request.