For teams that want the right answer at the right moment

Your best expert, available everywhere — and it always cites the page.

Manent reads your manuals, your procedures and your scans, then answers your technicians the way your most experienced technician would — backed by the file, the section, the page. Every answer draws on your documents; when a case falls outside what's documented, it connects you with your expert. The experience of your best people, ready to answer day and night.

Request the audit of my documents Try the interactive demo → The audit report belongs to you, whether you continue with us or not.
100 %
of answers backed by a cited source from your documents
95 %
of excerpts tied to their exact section of the manual*
$0
spent without an approved quote — AI costs capped by contract
1
isolated workspace per client — your data never crosses another's

* measured on a real corpus of 107 documents (457 MB), before/after audit published. Every choice the system makes rests on a documented principle — the full map, module by module.

Verba volant, scripta manent.

"Spoken words fly away, written words remain."

Know-how takes years to build. The day the person who carries it leaves — technician, shop foreman, engineer, owner — that knowledge walks out the door. Manent keeps it: your manuals, your procedures and even the answers of your experts become an assistant that answers the whole team, always citing the exact source — the file, the section, the page.

Sound familiar?

The breakdowns that cost you hours — finally an answer, source in hand

Your technicians know these cases by heart. Manent answers them in seconds, pointing to the exact page of YOUR documents — the official procedure, not a guess.

"The error code comes back after every restart."
Reset the sensor involved, then check that the connector is tight before restarting.
📄 Error code guide › p. 17
"The belt loosens after a few hours."
Check pulley alignment and set the tension cold per the table in the manual.
📄 Maintenance manual › §4.2
"Intermittent hydraulic leak on cold start."
Replace the O-ring and check the recommended oil viscosity for winter.
📄 Winterizing procedure › p. 6
What's already true

Proof first — testimonials follow

We are deploying with our first industrial clients. While we wait for their words, here's what you can verify today.

🔒 One isolated workspace per client, verified by the software 📊 Citation accuracy measured on YOUR corpus 🧾 Readiness audit delivered — the report belongs to you 🇨🇦 Built in Québec, designed for Law 25
"We don't ask you to take an AI at its word. Every answer cites the file and the page — or it stays silent. It's the only promise that holds up in the field."
David Leclair — Founder, Manent
How it works

Three steps, no surprises

You don't need to "prepare" your documents: the mess is our starting point, not a blocker.

1

Drop off your files

As is: PDF, Word, PowerPoint, scans, duplicates included. Our intake inventories every file — nothing is silently ignored — and quotes in advance what reading the scanned pages would cost.

2

Decide in 6 clicks

A guided journey asks you one question at a time: which version is authoritative, whether to include documents with contact details, whether to authorize digitization (exact quote shown). Everything is reversible and logged.

3

Your teams ask questions

Web, mobile, and soon augmented-reality glasses. Unanswered questions become cards to be validated by your experts — your documentation improves continuously, under human control.

Start with the proof

The AI-readiness audit

Before any commitment, we analyze your corpus and hand you a self-contained report, printable, presentable to your management.

  • An overall score out of 100 and the state of each document
  • Duplicates, unreadable scans, contradictory versions, documentation blind spots
  • A prioritized action plan — and the score is re-measured after each fix
  • One-time plan $4,500 to $12,000 depending on volume — a consulting firm bills $20,000 and up
88 / 100
AI-READY
Raw drop-off: 79 → after two preparation passes: 88
Why not 100 %?  →
Published pricing, no surprises

You always know what you're paying

Predictable subscription, AI usage measured and capped by contract — the cap is a technical mechanism, not a sales promise. No per-user fees.

Atelier

$350
per month · CAD

For a small team (2 to 4 people): 500 pages, 150 questions per month, email support.

Choose Atelier

Essential

$750
per month · CAD

2,000 pages of documentation, 500 questions per month, email support.

Choose Essential
Recommended

Professional

$1,950
per month · CAD

10,000 pages, 2,500 questions, field improvement loop, enterprise SSO, quarterly re-audit included.

Choose Professional

Enterprise

from $4,500
per month · CAD

50,000+ pages, 10,000+ questions, service commitment, AR glasses pilot.

Talk to an advisor

Beyond the quotas: $10 per additional 1,000 questions · digitization billed by the page, quote before any spending.

Not sure which plan? Answer 4 questions.

An indicative recommendation in 30 seconds — the diagnosis (audit) will refine it with figures measured on your documents.

1. How many people will ask the assistant questions?
2. Roughly how much documentation?
3. Your content is mostly…
4. About how many questions per month?
Recommendation:
    Request a quote for this plan
    Indicative only — nothing is measured here. The audit establishes the real figures (actual pages, scans, duplicates) and may change the recommendation.
    Security and governance

    Your data stays your data

    • Strictly isolated workspace per client — the isolation is verified by the engine, not just promised by contract
    • Enterprise authentication (SSO) and hierarchical roles
    • Every access, decision and AI expense is logged — a complete audit trail
    • Mandatory human validation: no card published without the approval of your experts
    • The safety instructions in your manuals are cited word for word, exactly as written — or escalated to a human

    The safeguard we're proudest of:

    when a case falls outside what's documented, the assistant tells you clearly, and the question becomes a card to be validated by your experts.

    That's how your documentation grows richer every week, guided by your experts and always tied back to its source.

    Frequently asked questions

    What people ask us first

    Why not bring me to 100 %?
    Because 100 % would be lying to you — and that's the lie we refuse. A vendor who promises 100 % makes the AI answer even where the information does not exist: it makes things up. We distinguish two kinds of "missing." The mess (duplicates, scans, old versions): that, we raise to the maximum. Knowledge that was never written down (a procedure no one documented): no tool can invent it — we capture it with your experts, and each capture raises the score. The percentage that remains is where the AI abstains and refers you to an expert instead of hallucinating: it's a protection, not a flaw. Our tool itself caps its projected index — out of honesty, it never allows itself to display perfection.
    Concretely, which documents do I receive?
    Three self-contained documents (printable to PDF), plus the assistant. 1) The audit report — score out of 100, problems found, action plan: it proves the real state of your documentation, it belongs to you (whether you continue with us or not), and it arrives a few days after we receive your files. 2) The ROI dossier — three quantified scenarios (conservative, typical, favorable), assumptions visible and adjustable: what this documentation can gain you. 3) The client dossier — both combined, plus a measurement bench on YOUR questions. Finally the assistant that answers while citing its sources is the product that follows. Every figure is about YOUR data; anything not measured is shown as "not measured," never a misleading zero.
    We already have a generalist AI assistant — why Manent?
    Manent knows YOUR machines, cites the file and the page of every answer, and keeps your data in your isolated workspace. Where a generalist assistant answers from public knowledge, Manent answers from YOUR documents — the right answer, for YOUR equipment, source in hand.
    Our documents are a mess. Is that a problem?
    It's our raw material. The audit measures exactly the state of your corpus (duplicates, versions, scans) and the onboarding journey lets you resolve the ambiguities in a few clicks. Nothing is deleted, everything is reversible.
    And our old paper scans?
    Our visual reading transcribes them — billed by the page, with an exact quote BEFORE any spending, never re-billed for the same pages. Without your written GO, not a single page is processed: it's a technical lock, not a courtesy.
    Who validates what the AI says?
    Two levels: every answer is verified against its sources before display and comes with its exact citation, and every new knowledge card goes through validation by YOUR experts before publication. The AI proposes, the human decides.
    How long before it's usable?
    The audit report arrives a few days after we receive your files. The assistant is queryable from the first ingestion; it becomes excellent after your onboarding-journey decisions and the digitization of the scans.
    First step, no commitment

    Request your quote in 30 seconds

    Even if your documents are a mess — especially if they are. We get back to you within one business day with the next steps.

    Prefer to explore on your own? Create your space — you can upload your documents and request the audit from inside.

    Or write to us directly: info@manent.ca