MS DuplexPersonal, team and organization layersShow allShow less
Personal, team and organization layers.
Each person, team and organization learns at its own level.
MS Duplex gives each researcher, each team and the whole organization its own layer of experience. Your rules decide what moves from one layer to the next, and none of it is shared with another company.
Remember and govern · one of the fourteen capabilities of MS Duplex


Your team gets
- Learning that respects who should see what
- Team know-how shared by your rules
- One standard across the organization where you want it
What personal, team and organization layers mean for your team.
Everything MS Duplex learns belongs to a layer. A scientist’s or engineer’s own working habits and preferences sit in a personal layer. What a team confirms sits in that team’s layer. What the organization adopts as standard sits in the organization layer and applies everywhere.
Each layer is used where it belongs. A researcher’s agent draws on their personal layer, their team’s layer and the organization layer together, so it works the way the researcher works, with what the team has confirmed and within the standards the organization has set.
Moving experience between layers is a decision your team makes. A method that proves itself in one team can be shared with another or adopted as an organization standard, under your rules and with its source kept. People and agents reach only the experience their scope allows.
Why research needs learning at more than one level.
Research knowledge has owners. Some of it is one person’s way of working, some is a team’s hard-won result, and some is the standard everyone should follow.
When AI learns in one shared pool, a single researcher’s habit can end up shaping everyone’s agent, and a team’s early finding can travel further than it should. When it learns nothing across people, every team rediscovers what the team next door already confirmed.
Layers let a research organization decide what is shared. Personal preferences stay personal, team results reach the people who work on them, and the methods the organization trusts apply wherever it wants them, each with a clear owner and a clear source.
Preferences are personal
How each scientist or engineer likes to work should shape their own agent, and only theirs.
Teams own their results
What a team confirms belongs first with the team that did the work and relies on it.
Standards should hold everywhere
A method the organization adopts should apply in every team and on every agent, with the same source behind it.
Access follows scope
People and agents should reach the experience their role allows, and no more.


How layers work in your team’s day.
Your researchers work as they do today. MS Duplex keeps each lesson in its layer and brings the right layers together for each agent.
- 1
A lesson is learned
A researcher’s correction, a team’s confirmed result or an approved conclusion becomes experience, with its source and the conditions it holds under.
- 2
It lands in its layer
A personal preference stays with the researcher. A confirmed team result goes to the team. Your rules decide which is which.
- 3
Agents draw on their layers
Each agent works from its researcher’s personal layer, its team’s layer and the organization layer, within the scope it is allowed.
- 4
Experience moves by your rules
When a team’s method proves itself, your team can share it with another team or adopt it as an organization standard, with its source kept.
An example, in any field where several teams use agents.
Two teams use agents to rank candidates for testing. Over a month, each piece of experience reaches the people it should.
- 1
A personal preference
A researcher in the first team asks her agent to show its reasoning before the ranking. Her agent keeps that preference, and nobody else’s agent changes.
- 2
A team result
The first team confirms that a scoring method fails outside a known range. The lesson goes to the team layer, and every agent on that team checks the range from the next request.
- 3
A shared method
The second team meets the same kind of case. The first team’s lead shares the method with them, and their agents start from it, with its source in view.
- 4
An organization standard
Leadership adopts the method as the standard for ranking candidates. It moves to the organization layer and every team’s agent applies it, while each researcher’s personal preferences stay as they were.
Layers make the other capabilities fit your organization.
What MS Duplex learns is only as useful as where it is applied. Layers decide that, for every other capability.
Rules and approvals
Your rules decide what moves between layers, and a change to shared experience can wait for a person’s approval, on record.
Reusable methods
A method one team verifies can be shared with other teams or adopted across the organization, and is re-verified where it is used.
Provenance, review and unlearning
Every lesson keeps its source as it moves between layers, so anyone can see where a shared standard came from.
Research memory across sessions
Each session starts from the current state of the work, drawn from the layers that researcher and team are allowed to use.
What your team gets.
Learning that respects who should see what, with sharing decided by your rules.
Personal layers
Each researcher’s working habits and preferences shape their own agent and stay personal.
Team layers
What a team confirms is shared with that team, so its agents build on the team’s own results.
An organization layer
One standard across the organization where you want it, applied by every agent.
Sharing by your rules
Your rules decide what moves between layers, and every move keeps its source.
Access by scope
People and agents reach only the experience their scope allows, with permissions kept under your controls.
Yours alone
The experience in every layer belongs to your organization, and none of it is shared with another company.
Common questions.
Make the next run benefit from the last.
Bring us the research your AI should get better at. We will plan a pilot on your own workflows, with your experts as the judges.

