Skip to content
MatterSpace

MS Duplex — The Agent Operating System powered by Runtime Intelligence. Rent the model. Own the intelligence.

Your rules come first.

A rule your team sets always wins over anything MS Duplex has learned.

Rules and behavioural control. It runs On every step, under rules your administrators set. In the Agent Operating System, it is the Runtime Intelligence that powers it all: learning from the work.

  • 0%*
    Stale answers after a policy change, against 72% for cache and retrieval. See the result →
    Stale answers after a policy change, against 72% for cache and retrieval
    MS Duplex0%
    Cache and retrieval72%

* Figures from MatterSpace’s internal benchmarks, not independently verified. Read the disclaimer

A habit says yes. Your rule says ask a manager. The rule wins.

MS Duplex has learned a habit from work your team confirmed: travel expenses from sales are approved at once.

Step 1 of 4: A habit

Finance stops catching the same exception.

A rule your team has set holds on every step, so nobody has to watch for the habit that slips past it.

Step 1 of 4: The rule

The rule is written down: any expense over £500 needs a manager.

A policy changes, and no stale answer follows.

When a policy changed in our test, MS Duplex gave 0% stale answers, on every one of ten runs. Cache and retrieval gave 72%.*

Step 1 of 4: The old policy

MS Duplex has learned the old policy from work your team confirmed.*

A rule holds, whatever MS Duplex has learned.

Your administrators set the rules, and MS Duplex applies them on every step, before anything it has learned. When a step breaks one, it stops, and the reason is recorded.

Step 1 of 5: Who sets it

The rule is set by your administrators, not learned from work.

See it on your own work.

From your own assistant to every agent your company runs, a pilot puts MS Duplex on your own work, with the measures agreed before it begins.