The firm exists because decisions come back
A regulator reopens a tariff. A court asks for the file, or a new board wants to know why the old one voted the way it did. At that point a deck is no use, because it carries neither the alternatives nor the limits of what was known at the time. We help boards decide while the evidence is still incomplete, and leave a record that survives the audit. Four people do that work. Two of them carry your file, and their names stay on it.
The names on the file did the analysis
Plenary is four people: two partners and two principals. The person who takes the first call is the person who reads the evidence and writes the document the board reads. Between 2019 and 2026 the firm has closed 112 files, across nine sector regulators. We have never handed a file down to a junior after signing it.
Who works on your file
Mireille Osei
Partner, energy and utilities
Mireille Osei spent four years on an energy regulator’s appeal panel. The files reached her finished, years after the vote, and usually the people who wrote them had moved on. Most of those files could not answer the two questions the panel cared about. What did the board know at the time, and what did it decide not to do? She left the panel and started Plenary with Tomas Reinholt, to write files from the other end. On an energy file she starts with the paragraph a regulator will read first, then works backwards into the evidence. She leads energy and utilities.
Tomas Reinholt
Partner, health systems
Tomas Reinholt built clinical trial protocols before Plenary. A protocol is written before the data exists. It states in advance what will count as evidence, what will count as failure, and what happens if a result arrives too late to use. He kept the habit. On a health systems file he settles what would change the board’s mind before anyone goes looking. The firm does not get to choose its standard after the answer is in. He leads health systems. He is the one most likely to say that the question on the agenda is not the question the room is arguing about.
“The vote happens, the evidence lands weeks later, and the conversation that follows is about who was right.”
Nadia Fournier
Principal, financial services
She leads financial services, where a risk committee and a strategy committee can read the same data and reach opposite conclusions.
She rewrites parts of the method after difficult files, which is the only time it changes.
Nadia Fournier writes most of the firm’s method
what would make this wrong
The section every recommendation carries, the one headed “what would make this wrong”, is hers.
a limit of confidence in a plain sentence
So is the rule that we state a limit of confidence in a plain sentence, never as a hedge.
recording two committees that disagree
The method has a section on recording two committees that disagree. It exists because that happened on a live file, and nobody had a clean way to write it down.
Anders Vik
Principal, public infrastructure
Anders Vik keeps the decision archive. Each file carries the question in the words the board agreed on. It carries the evidence and the point where it ran out, the alternatives, and the date each of those was true. Counsel cite it when a decision is reopened. We keep it in a form other people can read without us in the room. He leads public infrastructure, where investment programmes outlast the boards that approve them. At the end of every file he asks the same thing: could a stranger reconstruct this decision from what we are leaving behind.
Before Plenary he ran the records office of a port authority.
A dispute over a dredging contract was settled there by one dated memo, found because somebody had filed it the day it was written.
That job decided what he does now.
He works with transport authorities, water utilities and municipal boards.
The archive itself has been cited in cases.
Four sectors, nine regulators
We work where a decision can be reopened by someone with the power to reverse it. That is a narrow test and it rules out most industries. In practice it leaves four sectors, and the nine sector regulators those decisions have to answer to.
Decision framing, 2 weeks, then evidence review, 4 to 8 weeks, then board handover, 2 weeks.
Energy and utilities
Tariff decisions, network investment.
Led by Mireille Osei
Four years on an energy regulator’s appeal panel, reading the files from the other side.
Reopened by the regulator.
The regulator wanted the reasoning, not the result
A network operator set a tariff on evidence that was still moving, and the regulator reopened the file nineteen months later.
Health systems
Equipment procurement, service reconfiguration.
Led by Tomas Reinholt
Wrote clinical trial protocols before Plenary, and still writes the failure condition first.
Closed, then checked against the evaluation.
The evidence arrived after the vote
A hospital network had to commit to imaging equipment months before the clinical evaluation of one system reported.
Financial services
Leaving a segment, capital allocation.
Led by Nadia Fournier
Writes most of the firm’s method, including the rule about what would make a recommendation wrong.
Reopened at the exit review.
Two committees, one decision
A bank’s risk and strategy committees read the same data and reached opposite conclusions the board could not choose between.
Public infrastructure
Long investment programmes, renewal and concession decisions.
Led by Anders Vik
Keeps the decision archive, the one cited when a decision is reopened.
Rebuilt from the archive.
A decision nobody could reconstruct
A water utility inherited an eleven year investment programme, four hundred slides, and no record of the alternatives.
Evidence review, 4 to 8 weeks, then board handover, 2 weeks.
Two of the four on every file, no juniors on the analysis
Two of the four are on every file, and one of them is always a partner. In energy and in health systems the partner carries the sector. In financial services and public infrastructure the principal carries it. The second name on the file is the one who reads it as the person who will disagree. That reading has to be argued down in writing before anything reaches a board. Nobody here reports evidence they have not read themselves.
The people on the file
That sets a limit on how much the firm takes on at once, and a full cycle runs eight to twelve weeks. The person who wrote the section on the limits of confidence answers for it when the board disagrees.
How a file runsMireille Osei
Partner
Tomas Reinholt
Partner
Nadia Fournier
Principal
Anders Vik
Principal
The first file is dated 2019
There is no founding story worth dressing up. Mireille Osei had spent four years on a regulator’s appeal panel, reading decisions long after the vote and finding no reasoning in them. Tomas Reinholt had spent years writing trial protocols, which fix what counts as evidence before any of it arrives. In 2019 they put the two habits on the same file. The first clients were people who already knew them. The early files took longer than they should have. The method came after the work, not before it. Nadia Fournier wrote most of it over files that had gone wrong, and she still edits it.
We would rather you judged the firm on the record than on the story.
What we turn down
Most of this we learned by taking the work first and regretting it afterwards.
A document for a decision already taken
If the conclusion is fixed, the record will show that it was fixed, and then you own a record that reads badly.
Files where we never meet the board
We write for the people who vote, and we need to hear them ask the question in their own words.
Anything that has to be finished in under two weeks
Framing alone takes two weeks. Faster than that, you are paying us to write down what you already believe.
Implementation
We stop when the decision is made and the record is with you.
Files where the evidence cannot be shown to us in full
We cannot write the limits of confidence around evidence we were not allowed to read.
The deliverable is the reasoning, not the slide
You keep it in editable form, and our copy stays in the archive under the date it was written. Six things are in it, and they are in it every time.
The question, in the words the board finally agreed on
What the evidence supported, and the point where it ran out
Every alternative, and the reason we set it aside
What would make the recommendation wrong
The date each fact was true, because the record has to show what was known at the time
Written for a board session that runs 40 minutes
Five files, with the alternatives left in
We write up five files in full. Five pieces on method sit beside them, including “What a board actually reads” and “How to say what you do not know”.
A regulator reopened a tariff decision 19 months after the vote
In a hospital network the evidence landed after the vote
At a bank the risk committee and the strategy committee read the same data in opposite directions
A water utility inherited an 11 year programme with 400 slides behind it
An insurer’s board asked for eight pages, and the asking changed the analysis
Send the question as your board words it today
It does not have to be the right question. Narrowing it is the first two weeks of work. Tell us who votes, when they vote, and what you already know you will not be able to find out in time. All four of us see what arrives. The one who replies is the partner or principal who covers your sector, and that is the person who would carry the file. If the file is not ours, we say so in the first conversation and tell you who we would call.