Roadmap

Three states, and no dates anywhere.

Shipped means you can go and use it now. Building means somebody is on it this week. Considering means it is wanted and not committed to. There are no quarters anywhere on this page, because we could not honour them.

Shippedin the product today

Create a workspace and use it this afternoon. The feature inventory lists every piece of it, and marks the rough edges as rough rather than as finished.

Buildingin progress right now

Under way this week, and still with no date on it: an estimate we cannot keep is worth less to you than the plain statement that somebody is on it.

Consideringwanted, not committed

Wanted by somebody and committed to by nobody. Some of it will never be built — saying that now is cheaper for both of us than saying it in a year.

Shipped

The charter, its signatures and its numbers are all built.

Not a beta, and not a waitlist: create a workspace this afternoon, click the demo charter, and every count on the cover beside this is computed out of the document you are looking at.

The full inventory is on the features page, and it marks the rough edges as rough rather than calling everything finished.

From the document · cover

5departments8handoffs7stages2lifecycles14metrics4offerings2campaigns6committed goals

Every one of those counts, and the alignment score beside them, is computed from the charter itself — not typed into a cover page.

What “shipped” covers

The document

A seeded operating model — departments, handoffs with SLAs, lifecycle stages, offerings and talk tracks — published with your logo, a contents block, computed semver versions, tracked-changes redlines and a print stylesheet.

The agreement

Who must sign is derived from the structure, every signature is pinned to a hash of the wording it approved, edits mark it stale, and only the owner ratifies — which is the one act that bumps the major version.

The numbers

Monthly actuals, a baseline computed from each metric’s declared basis, goals that derive from campaign targets, forecasts frozen before the month they predict, and a scoreboard that keeps the model’s record either way.

Building

Three things are in progress right now.

Two of them are holes we can name precisely, which is the only reason they are being worked on ahead of anything else. Nothing here carries a date; when one lands it moves to the features page and off this list.

  • A designed public site

    building

    The home page, this roadmap, the feature inventory and the how-it-works page, rebuilt as one public zone with its own type scale and chrome — deliberately not the signed-in app's.

  • Custom metrics

    building

    Today every metric is stage-derived: volume, conversion and time, per lifecycle stage. Anything that is not a stage — a margin, a support load, a renewal rate — cannot be tracked at all. That is the single biggest hole in the product.

  • Signature revocation history

    building

    The audit log records signatures but not their withdrawal, because revoking deletes the row. For a document whose whole value is provable agreement, a withdrawn signature has to leave a trace.

Considering

Wanted. Not committed to.

Everything below is genuinely on the table and none of it is promised. Several are blocked on the same thing — CRM data — and saying so is more useful than listing them as if they were independent.

  • CRM sync, HubSpot first

    considering

    The CRM-binding fields already exist on every lifecycle stage and bind to nothing. This would fill the monthly actuals automatically, with manual override kept, and is the prerequisite for most of what follows.

  • Offerings scenario model

    considering

    Sell X of offering A, at its ACV, on an N-month term with an R% renewal rate, and see the revenue shape that produces. It replaces the parked calculator, and is only worth building because offerings already carry terms, renewal rate and recognition method.

  • Product / line-of-business segmentation

    considering

    Actuals split by product line, in a way where the totals still reconcile. Segmentation that quietly stops adding up is worse than none.

  • Scheduled review reminders

    considering

    Review cadence is prose inside the charter today, not a timestamp, so nothing can fire off it. Making it a real schedule is the work.

  • Sub-monthly granularity

    considering

    Weekly or daily numbers need CRM event data to mean anything. Without it, finer periods would be invented precision — which is why the version of this without CRM sync is on the not-planned list below.

  • Seasonal forecast model

    considering

    It unlocks at twenty-four months of history by its own rule: a seasonal model fitted to twelve months is fitting noise.

  • Blog and changelog

    considering

    Scaffolded, and deliberately unlaunched until there is something worth reading on it.

Waiting on account setup

Not engineering work, so it does not sit in the queue.

These three are waiting on third-party accounts and the decisions that go with them, not on code. They are listed separately because putting them in “Building” would flatter the engineering picture, and putting them in “Considering” would understate how settled they are.

  • Billing

    Subscription per workspace and trial gating, through Stripe. The work is an account and a decision about pricing, not a hard engineering problem.

  • Password reset and team invites

    Both need transactional email — a Resend account and a verified sending domain. Until then, invites and resets are the two gaps in an otherwise complete auth system.

  • Dedicated database with row-level security

    App-level tenant scoping is live and verified today: cross-tenant reads and writes return 404. Database-level RLS is the belt on top of those braces, and it needs its own database project.

Not planned

Two things we have decided against.

A refusal is a commitment too, and it saves everyone a conversation. If either of these is a requirement for you, this is not the product for you — which is worth knowing before you sign anything into it.

  • Sub-monthly reporting without CRM data

    A weekly number derived from a monthly entry is invented precision. If the events are not there, the chart should not pretend they are.

  • A mobile app

    The web app fits a 390px viewport, and the document collapses into stacked records rather than a grid you scroll sideways. A wrapper around that would be a download, not a feature.

Nothing on this page has a date,
and everything on the features page is real.

Create your workspaceSee what is built