Skip to content
Advancing Bits to Master Atoms

Reusable methods.
One verified result becomes a method your agents reuse and adapt.

MS Duplex turns a procedure confirmed to work into a method your agents apply on the next case like it. On a related case, the method is adapted and proposed for checking, and the methods your team trusts are shared across people and agents.

Learn and improve · one of the fourteen capabilities of MS Duplex

A satin-silver stamping die on a graphite plinth beside a neat row of identical silver parts, each with the same small cobalt-blue inlay.A satin-silver stamping die on a graphite plinth beside a neat row of identical silver parts, each with the same small cobalt-blue inlay.

Your team gets

  • Proven procedures applied consistently
  • Know-how that stays when people move on
  • New team members start with the team’s best methods

What reusable methods mean for your team.

A reusable method is a procedure your team has confirmed to work, kept so your agents can use it again. It holds the steps, the conditions they were confirmed under and the result that proved them, so the next agent to use it knows what it is applying and why.

MS Duplex uses a method in two ways. On the next case like the one it was confirmed on, the method is applied directly. On a related case, where something differs, it is adapted to the new conditions and proposed for checking before anyone relies on it. Once the adapted version is confirmed, it becomes a method too.

Methods your team trusts are shared across the people and agents your rules allow. Each one is re-verified where it is used, so a method that worked in one setting is checked again in the next before it is trusted there.

Why research needs methods that stay with the team.

A research team’s know-how is built one confirmed procedure at a time. Too often it leaves with the person who worked it out.

Every team has procedures that took long trial and error to get right: the order of steps that gives a clean analysis, the settings that make a simulation behave, the checks that catch a bad batch early. They are rarely written down in a form an AI can apply, and they are applied unevenly from one person to the next.

When the AI cannot reuse them, every agent works the procedure out again, or skips it. Results vary with who ran the work, new team members spend their first months learning what colleagues already know, and when an expert moves on, the method often goes with them.

  • Consistency matters

    A result is easier to trust and to reproduce when the same proven procedure produced it each time.

  • Know-how is fragile

    Procedures that live in one person’s head or one notebook leave when that person moves on.

  • New members start behind

    People joining the team should begin with its best methods in hand, ready to apply.

  • Settings change

    A method proven in one setting needs checking before it is trusted in another.

A set of meshing gears on a graphite plate, one cobalt-blue master gear driving satin-silver gears of different sizes.A set of meshing gears on a graphite plate, one cobalt-blue master gear driving satin-silver gears of different sizes.

How reusable methods work in your team’s day.

Your team confirms what works as it does today. MS Duplex keeps it as a method and puts it in front of the next agent that needs it.

  1. 1

    A procedure is confirmed

    An agent or a researcher follows a procedure, and a test, a measurement or an expert confirms that it worked.

  2. 2

    It becomes a method

    MS Duplex keeps the steps with the conditions they were confirmed under and the evidence behind them.

  3. 3

    It is reused or adapted

    On the next case like it, the method is applied directly. On a related case, it is adapted and proposed for checking.

  4. 4

    It is shared and re-verified

    Methods your team trusts reach the people and agents your rules allow, and each one is re-verified where it is used.

An example, in any field where an agent runs analyses.

A team’s agent prepares and runs a recurring analysis. One researcher’s careful procedure becomes the way the whole team’s agents do it.

  1. 1

    One good result

    A researcher finds an order of preparation steps that removes a persistent error from the analysis. An expert confirms the result, and MS Duplex keeps the procedure as a method, with its conditions.

  2. 2

    The next case like it

    A colleague’s agent meets the same kind of analysis on new data. It applies the method directly, and the result comes back clean.

  3. 3

    A related case

    Another agent meets the analysis with a different input format. MS Duplex adapts the method and proposes the adapted version for checking. The researcher confirms it, and it becomes a method of its own.

  4. 4

    A new team member

    A new hire’s agent starts on the same work with both methods already in place, each with the result that proved it.

Reusable methods connect to the rest of the layer.

A method is only as useful as the learning behind it and the rules around it.

  • Continual learning

    A procedure is learned the moment it is confirmed, so the method is ready from the next request.

  • Solving new problems

    On a new case, MS Duplex assembles an approach from the methods your team already trusts, and once it is verified it becomes a method too.

  • Personal, team and organization layers

    A method can stay personal, be shared with a team or become an organization standard, as your rules decide.

  • Provenance, review and unlearning

    Every method keeps the result it rests on, so anyone can see why it is trusted and correct it on its own.

What your team gets.

Your team’s best procedures, applied consistently by every agent and kept with the evidence behind them.

  • Proven procedures applied consistently

    The next case like the one a method was confirmed on gets the same proven steps, whoever runs it.

  • Adaptation with a check

    On a related case, the method is adapted and proposed for checking before anyone relies on it.

  • Know-how that stays

    A method stays with the team when the person who worked it out moves on.

  • New members start ahead

    New team members and their agents begin with the team’s best methods already in place.

  • Sharing by your rules

    Methods reach the people and agents your rules allow, from one researcher to the whole organization.

  • Methods on record

    Each method keeps its steps, its conditions and the result that proved it, ready to review.

Common questions.

What counts as a method?

Any procedure your team has confirmed to work: a sequence of steps, a choice of settings, a set of checks. MS Duplex keeps it with the conditions it was confirmed under and the evidence behind it.

What happens when a new case is only similar?

MS Duplex adapts the method to the new case and proposes it for checking before anyone relies on it. Once the adapted version is confirmed, it becomes a method for the next case like it.

Who can use a method once it exists?

The people and agents your rules allow. A method can stay with one researcher, be shared with a team or become a standard across the organization, and none of it is shared with another company.

How is this different from a shared notes file?

A notes file holds what someone wrote down, and the model rereads it and judges what applies. A method in MS Duplex is kept with the result that proved it and the conditions it holds under, applied where those conditions match, and adapted and checked where they differ.

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.