Everything your organization runs on.
Everything your household runs on.

Shape the work your field's way

Run three levels or four, and rename every one of them — an industry starter pack renames the whole hierarchy in English and French in one action, so a construction firm sees phases and a clinic sees programmes. Every item keeps a permanent Lplan#### code that never renumbers, plus nestable sub-tasks, file attachments and priority labels you define yourself, with your own colours.

Know that "done" means done

Work is submitted by the person who did it and approved by someone else — the app will not let one person do both. Approvers can request changes instead of rejecting outright, rejected work can be resubmitted, and every submission, note, edit and sign-off stays on the item's history with who did it and when.

Capacity you can actually act on

Utilization is computed from this week's real load against a capacity that has already had vacations and Canadian statutory holidays subtracted — all thirteen provinces and territories.

"Who should do this?" ranks your people on availability and demonstrated experience with similar work, and it makes no AI call at all. Bulk rebalancing shows you its reasoning before it moves anything.

Forecast the whole business

Seven views over an 8 to 52 week horizon. The delivery bands are P50, P80 and P95 from a bootstrap simulation calibrated on your organization's own estimate-versus-actual history — so the range narrows as your estimating gets better, not because we made it look tidier.

AI-written estimates are excluded from that calibration, so the forecast can never narrow its own bands on its own guesses. Save scenarios side by side, or ask the deadline solver what would have to change.

Run the money properly

Every organization gets its own real double-entry ledger. Entries are immutable — a correction is a reversal, never an edit — and the income statement, balance sheet, cash flow and equity statement articulate with each other by construction.

Present the same books under GAAP·ASPE, IFRS or SYSCOHADA. Budgets roll up from children to parents, and task actuals reach the ledger only when an admin explicitly posts them. Management-grade books, not audited.

Add only what you need

Sales, Product, Operations, Marketing, HR and Finance are all off by default, and an admin picks which features appear in the sidebar at all. Most organizations never turn most of them on, and that is the intended outcome.

The genuinely unusual one is Operations: a site is a function. A depot or a clinic is a node in the hierarchy you already have, so there is no second org chart to keep in sync and no parallel reporting line.

A range, honestly drawn

The fan opens as it goes further out because that is what your own history says. A forecast that stayed narrow twelve weeks out would be telling you something nobody can know.

Two columns that must agree

Every posting has an equal and opposite side, so the statements cannot quietly drift apart. If a number looks wrong, the ledger will tell you which entry made it wrong.

Every draft is a card you confirm

The assistant proposes; you accept. Edits, calendar blocks, meetings, recipes, a week's meal plan — each arrives as a card showing exactly what would change before anything is written.

External email is never sent on its own

Mail leaving your organization always waits for a human click. There is no setting that turns that off — not a default we chose, an absence of the option. Meeting invites work the same way, because an invite becomes a notification on other people's phones.

The item assistant cannot write at all

It has no tools and no privileged database client. It returns a proposal, and your own browser applies it through exactly the same permission check you would hit doing it by hand — so it can never do something you could not have done yourself.

A list AI may never touch

Hours, money, budgets, completion, approvals and priority are closed to it. Those are the fields people are judged and paid on, so a machine does not get to move them.

Every AI write is on the record

Each one is logged with the prompt that caused it, notifies the item's owner, and can be undone. The history also records whether you took the suggestion verbatim or edited it first — so nobody has to guess later how much of this was the machine.

Other people's words are data, not orders

Anything the assistant reads — a forwarded email, a note someone else wrote, a meeting caption — is fenced as content to be read, never as instructions to follow.

The clever parts aren't AI at all

Item signals, the "who should do this?" ranking and bulk rebalancing are ordinary arithmetic over your own data. They behave identically with AI switched off, which is why the no-AI plans are a real product and not a crippled one.

Sold separately, disabled centrally

Messaging, email and the meeting note-taker are separate capabilities, each off unless enabled for your organization. Usage is capped per person and per organization so one enthusiastic week cannot surprise you.