Our approach
Your dashboards show what you measure. They don’t always tell you what is happening. B-ADSc reveals how your organisation works: its objectives, its decisions and the way it regulates itself.
We combine three strengths: listening to the whole front line through conversational AI, organisational engineering grounded in B-ADSc, and the experience of our partners.
The goal: understand what is holding you back, then build an organisation that decides better, delegates more clearly and adapts to its environment.
100%
of the scope invited. More than 85% participation.
2.2–6×
Documented ROI of actions identified in the first half of 2026
In 4 weeks
from scoping to action plan
Senior
Partners do the analysis. No junior staff.
The issue
Your organisation chart shows who is responsible.
It rarely shows who actually decides.
You have written procedures, a capable ERP and detailed dashboards. Yet delays pile up, trade-offs contradict each other and the same decisions keep moving up to the same level.
Why? Because these representations are not enough to explain how the organisation is actually steered. What result are we aiming for? Who decides? With what information and how much autonomy? How does the decision-maker learn the effects of their choices?
Organisation chart
It places people and lines of authority.
Processes
Process mapping describes the planned sequence of operations.
Information system
IT architecture describes digital resources and how they are orchestrated.
B-ADSc
B-ADSc completes these representations with the actigram: the hierarchy of activities that pursue objectives, make decisions and evaluate the results of their conduct.
This model is called the actigram. It links your strategic objectives to everyday decisions, without confusing an activity with a job, a task or a piece of software.
B-ADSc in plain terms
An organisation is a set of activities
pursuing goals.
Janusz Bucki designed the Decision Analysis of Complex Systems, or B-ADSc. It describes every activity through three inseparable elements.
Its purpose
The result to achieve or maintain, and its reason for being.
Its conduct
The decisions and know-how that make it possible.
Its resources
The resources and sub-activities it steers.
An activity is distinct from whoever runs it. A person, a team with a shared line of conduct or a technical system can run it. B-ADSc first models the decisions that are needed; it then examines who can take them, and with what resources.
At every level, four practical questions
1
What objective have we been given?
The result our environment expects from us.
2
How do we understand the situation?
The state of the situation on which we base our decisions.
3
What state do we want our resources to reach?
The objective our decision sets for our resources.
4
What changes are reported to us?
The information that will let us reassess the situation.
A simplified view of an activity. It separates the decision, the actions, the internal objective and the feedback. It does not replace the full B-ADSc notation. Dashed lines: possible feedback.
Two functions do the steering. Decide: choose the course of action according to the objectives pursued and the situation as understood. An interface then translates the decision into actions. Evaluate: interpret the reported changes against the target state, update the understanding of the situation and report back when needed. This understanding, or a change in objectives, may lead to a new decision.
What the method reveals
What B-ADSc makes visible,
and your dashboards rarely show.
Objectives that pull apart.
Two activities can meet their own objectives while working against the overall objective. B-ADSc examines how objectives are delegated and which arbitration rules help or prevent them from converging.
Decisions taken at the wrong level.
Some choices climb to senior management for no reason. Others are left to people who lack the information or authority they need. B-ADSc examines at what level, and over what time horizon, each decision should be made.
Scarce resources with no sharing rules.
Expert time, industrial capacity, budget or access to a platform: when several objectives call on a scarce resource, arbitration rules decide how it is allocated. Those rules must stay suited to the context.
Indicators that do not help steer.
An accurate indicator can still be useless if no activity knows how to use it to evaluate, decide, report or learn. Conversely, critical information may never reach the activity concerned. B-ADSc helps rationalise indicators according to how they serve steering.
Digitising an incoherent organisation can amplify its dysfunctions.
In the age of AI, digitising a broken process can speed up its dysfunction. You first need to decide which decisions to simplify, redistribute or hand over to a technical system.
Example
Why are quotes
approved too late?
A real situation; the case around it is a composite.
The symptom
Sales reps wait longer for quote approval than the response time promised to the customer.
The usual analysis
The quick fix would be to add a notification, change tools or speed up the approval pace.
The questions B-ADSc asks
What must approval guarantee first: risk, lead time or margin? Who sets these objectives? How does it decide between two competing requests? What does it know about their urgency and risk? Which decision could it delegate? What signal does the sales rep receive when the deadline slips?
Leads to check
Interviews and observable evidence make it possible to examine delegation, priorities, control rules and information flows. A new CRM would only be justified if the facts show it helps.
Explore the questions B-ADSc asks, starting from a symptom.
A symptom you recognise
An activity
Quote approval
Evaluate: does it match?
Decide: what to do?
What it is asked to do
Its objectives, and who sets them
What it knows
The information that reaches it
What it expects
The results sought and the criteria for judging the gap
What it believes
Its view of the situation
Where it breaks
In what it expects: approval applies the same criteria to every amount.
What B-ADSc finds
The €2,800 quote waits behind the €80,000 one.
What the plan decides
Thresholds by amount, a fast track below the threshold.
B-ADSc helps identify conflicting objectives, poorly defined trade-offs and missing information. Spentia tests these findings against the available facts before building an action plan.
What it brings
Six ways B-ADSc helps you
transform for the long term.
01
Act on mechanisms, not just symptoms
B-ADSc connects objectives, decisions and feedback. This analysis can reveal where bottlenecks arise and which causal hypotheses must be checked before acting.
02
Align objectives, even in a matrix organisation
Some activities must arbitrate between several valid objectives. B-ADSc distinguishes steering one objective, selecting between competing objectives and, for some activities, creating their own objectives. It helps make sharing rules and the conditions for aligning goals explicit.
03
Give autonomy without losing control
Delegating means entrusting decisions, not just tasks. B-ADSc sets out their scope, their time horizon and the information needed to steer results.
04
Rethink organisation, skills and information systems together
B-ADSc separates the activity to be run from whoever runs it. Business teams, HR and IT then share the same language. You design the target organisation from the decisions it requires, then choose the skills, tools and resources.
05
Choose the right automation and AI opportunities
Which decisions can be entrusted to a machine, on what conditions and with what controls? B-ADSc helps describe the objectives and rules required, then assess whether human, automated or mixed steering is feasible.
06
Prepare an organisation that learns and evolves
B-ADSc describes learning from experience, adaptation and, depending on the activity's autonomy, the creation of new objectives. The model can also feed impact studies and simulations, provided the right resources and specific validation are in place.
How Spentia puts the method into practice
About four weeks.
Two to three sessions with you.
At scoping, we set the scope, the timetable and the depth of analysis. We test accounts against the available documents and, where possible, against operational records.
Week 1
Scope
We agree with you on the purpose of the engagement, the decisions to inform, the objectives, the scope, the usable sources and how people will take part.
Week 2
Listen
Aria, our conversational interview agent, gathers the concrete experience of every person in scope. Each interview brings out real situations, trade-offs, exceptions, workarounds and the signals people actually use. These accounts feed the investigation; on their own, they do not prove a cause.
“You say two hours a day. On which tasks?”
Week 3
Model and verify
Our partners cross-check the contributions, reconstruct the activities and their delegation, then qualify the findings. They identify contradictions and the checks still needed.
Week 4
Design and decide
With your decision-makers, we compare scenarios: new rules, delegation, coordination, processes, skills or technical resources. We estimate gains from explicit assumptions, then rank actions by impact, feasibility and dependencies.
Then
What next?
We can support implementation and measure its effects. If your teams carry on alone, the action plan sets out the decisions, the priorities and the conditions for success.
Aria gathers. Our partners analyse and conclude.
What you receive
An action plan built for decision-making,
tailored to your priorities.
How your organisation really works
Activities, objectives, decision responsibilities, interfaces and areas of uncertainty.
A map of dysfunctions and opportunities
Symptoms, causal hypotheses, scarce resources, trade-offs, failing signals and how effects may spread.
Qualified findings
For each one: its sources, how well it is corroborated, any contradictions and what remains to be checked.
Transformation scenarios
Organisational, skills and technical choices, with the conditions for making them work.
A prioritised action plan
The order of implementation, owners, tracking indicators and documented economic estimates when the data allow.
Where relevant, an AI and automation roadmap
What to simplify, assist, hand over to a system or keep under human control.
For a company, a division or several sites: understand the bottlenecks and build a costed plan.
From first contact to cash collection: identify what consumes sales time and prioritise the actions.
The numbers
Figures that hold up
in front of your leadership team.
Low estimate
Every amount uses the low end of its range.
Prudence × 0.7
Every reported time is reduced by 30% before it is costed.
Counted once
A lost hour counts only once, even if ten departments mention it.
Traced to source
Every finding links back to the interviews it rests on.
Finding 2 of 12 · Discovery
Quoting after inspection blocks the whole chain.
Cited in
74 of 230 interviews
Priority
2 of 3
Direct cost
≈ €480k/year
“We sensed tensions without being able to name them precisely. By interviewing the whole team, the assessment made them legible: where they lie, what feeds them, in what order to address them. The method clearly separates facts from hypotheses, which gives confidence in reliable conclusions and a solid starting point for action.”
Jean-Ghislain de Sayve, Board Advisor & Managing Partner Switzerland, Hunteed
Confidentiality and data
Your teams speak freely.
Your data is hosted in France.
AI speeds up listening. The method structures the analysis. Our partners remain accountable for the conclusions.
Aria gathers the experience of the people in scope. B-ADSc structures the analysis. Our partners review the findings, request the necessary checks and build the recommendations.
For each engagement, we agree with the relevant people in your organisation on confidentiality, informing participants, data access, aggregated reporting and retention periods.
We always separate what is stated, what is corroborated, what is inferred and what remains unknown. That way, neither an AI answer nor an isolated quote becomes, by shortcut, a truth about your organisation.
> 85%
of those invited take the interview.
No one in your organisation reads an individual response.
Every quote is rewritten in a neutral style.
A person alone in their role is never identifiable.
The findings address processes, never people.
Anonymity guaranteed from 30 people.
Below that, we agree with you on how to protect it.
Hosted in France
On sovereign infrastructure.
A sovereign AI model, on request
We run it for you: Mistral, or an open model isolated on our servers.
GDPR and AI Act
The setup can be presented to your employee representatives, if you wish.
The foundation
An engineering method,
born of the automation of complex systems.
Janusz Bucki, doctor of mathematical sciences, developed B-ADSc (Bucki, Analyse décisionnelle des systèmes complexes) to analyse and design complex automation systems. The method also applies to organisational engineering. Techniques de l’Ingénieur publishes his reference paper, “Analyse décisionnelle des systèmes complexes” (AG 1570).
The method applies to activities run by people as well as by technical systems. Spentia combines it with field investigation and the experience of its partners.
Formal
The same rules everywhere
From the front line to the leadership team, every finding is discussed on facts, not impressions.
Published
Techniques de l’Ingénieur
Set out by Janusz Bucki in article AG1570 of Techniques de l’Ingénieur, on 10 October 2002.
People and machines
One model
An activity is distinct from whoever runs it, human or technical. This distinction makes it possible to study which decisions can be delegated to a machine.
Our partners apply B-ADSc on every engagement.
Reza Ghaem-Maghami
Strategy and business development
Philippe Gautier
Processes, systems and transformation · B-ADSc lead
Violette Bouveret
Growth and implementation
François David
Profitability and performance management
Julie Jeantet
Leadership, skills and organisation
What if you started with one question:
how do we really decide?
Thirty minutes with a partner is enough to look at your problem, the scope to examine and the decisions the action plan should inform.