Restructuring · Artificial Intelligence
AI in Restructuring: Where Artificial Intelligence Actually Helps, and Where It Does Not
Few terms promise as much in restructuring right now as artificial intelligence. The expectation is a system that spots the crisis earlier, analyzes it faster, and delivers the right measures along with it. Part of that is real. Another part is marketing.
This article separates the two. It shows where AI gains measurable time and precision in a turnaround, and where it hits limits no model can remove. The short version up front: AI shortens the path from data to insight. The path from insight to realized, bankable impact it does not walk alone.
01
Why AI matters in a restructuring at all
A restructuring runs against the clock. The Chief Restructuring Officer has to stabilize a frightened organization in 12 to 18 months, secure liquidity, and convince the financing parties at every interim update. Information in that situation is scarce, and it is often scattered: across spreadsheets, across emails, inside people's heads.
This is exactly where AI starts. It can work through large, unstructured volumes of data faster than a team could by hand. In a phase where every week counts, that is not a convenience but a real lever. The decisive question is only this: at which point in the turnaround process, and with what reliability?
02
The crisis stages and the fitting AI support
A crisis develops in stages. Observations from restructuring practice suggest that a large share of turnarounds only begin in advanced stages, when room to maneuver is already tight.4 The earlier an early-warning signal is recognized, the more options remain. AI moves that detection point forward.
The table below maps the typical stages to realistic AI contributions. It is kept deliberately sober: what is listed here is analysis and proposal work, not a decision.
| Crisis stage | Where AI supports | What stays with people and mandate |
|---|---|---|
| Strategy crisis | market and competitive analysis, scenario simulation, detection of strategic drift | judging which scenario is viable |
| Earnings crisis | automated cost and margin analysis, forecasting of measure effects | prioritizing by impact and feasibility |
| Liquidity crisis | liquidity forecasts on a daily and weekly basis, early warning | negotiation with banks and creditors |
| Near insolvency | contract and receivables analysis, simulation of realization scenarios | legal and fiduciary responsibility |
The pattern is the same across all stages. AI delivers the groundwork, faster and broader. The decision, the negotiation, and the accountability stay with people. Whoever confuses the two builds in a risk that becomes especially expensive in a crisis.
03
The three fields where AI shows real impact
Beyond the stage logic, three fields can be named where the benefit is documented and not merely asserted.
Issue detection and early warning
Under time pressure, delays matter, and AI can deliver transparency. Better still, it can prioritize: which delays matter because the activities pay into a large financial effect. But it is not only delays that matter. Missing information, non-actionable items, and deviations in operational KPIs warn early that important effects may not materialize later. AI can mine for all this scattered information.
Action proposals
From the given project context, AI can make proposals for milestones and activities. That does not replace the judgment of an experienced restructurer, but it shortens the path from finding to a structured action plan considerably.
Focused, personalized information
Across a running program, reports contain lots of data, only a small fraction of which is relevant to each participant. The additional information often wastes their time and drains mental energy. AI can deliver personalized reports on impending tasks, issues in one's own area of responsibility, and similar items, so everyone is empowered to act without undue waste and delay.
These three fields share one thing: they make the people involved faster and better, instead of replacing them. That is exactly how AI is built into ChangeMaker®.
04
Where AI hits its limits
AI is no cure-all, and in a restructuring the honest naming of its limits is not an admission of weakness but a precondition for trust. Three limits matter most in practice.
Data is never neutral.
A model is only as good as the data it works with. In a crisis the data is often patchy, outdated, or polished. Distorted inputs produce distorted proposals. Whoever does not check this automates their own blind spot.
A forecast alone is not enough for a bank; it needs evidence.
An AI can create awareness around delays, but cannot force action. Banks want to see how the financial impact of restructuring measures moves through the maturity levels. This requires embedding AI-generated reporting into a short-looped steering logic. The program organization does this, not the AI.
AI does not mobilize people.
A restructuring rarely fails for lack of analysis. It fails because a frightened organization does not take action. AI can recognize issues. But real progress and tangible results are the best remedy for organizational anxiety, and this AI cannot achieve. That is a leadership task, not a model task.
05
From proposal to bankable measure: the maturity level
This is the most important transition in this article. An AI that helps with detailing measures, creating transparency around delays and other issues, and delivering tailored information to each participant is useful. It helps to be fast and focused. But more is required to get to a bankable outcome.
In ChangeMaker®, every measure runs through a maturity-level (Härtegrad)
workflow: defined maturity stages from idea through concept to the realized, earnings-effective
measure. An AI proposal starts as an idea, not as a success. Only when the evidence criteria are
met does a measure count as creditor-ready: substantiated, earnings-effective, auditable.
Whatever does not meet the criteria stays visibly open.
The second effect is time. In well-structured programs the consolidation and reporting effort drops by up to 85%, and per measure the program team saves on average around 8 days of administrative work.2 AI accelerates the analysis; the maturity-level workflow and automatic aggregation accelerate the reporting. Together they hand time back to the mandate.
06
The ChangeBot: AI native, not bolted on
Many tools have, over the past years, laid an AI layer on top of an existing product. It shows when the assistant sits beside the data rather than inside it.
The ChangeBot in ChangeMaker® works within the
PerformanceMap®, that is, inside the structure of objectives, measures,
responsibilities, and financial impact. From that follow the documented functions: drafting work
items like milestones or tasks, generating up-to-date insights on the program, and identifying
risks and other issues.2 Because the assistant knows the
program data, it answers project-specific status and potential questions in context, not from a
detached knowledge base.
The difference is no marketing detail. An AI that understands the hierarchy and the maturity stages of a program can sensibly flag a stalling measure. An AI that only generates text cannot.
07
The other half of the truth: Concerns, Competencies, Coordination
Even the best AI-supported analysis does not bring a restructuring to the finish line if the organization does not pull along. The behavioral science behind ChangeMaker® describes three success factors of any change, the 3C method: Concerns, Competencies, Coordination.
| Dimension (3C) | Question in the crisis | What AI contributes |
|---|---|---|
| Concerns | Will the people involved pull along instead of digging in? | little. Buy-in comes from leadership, not from a model. |
| Competencies | Can they execute the measures under pressure? | some. AI can supply knowledge, but cannot empower and build practice. |
| Coordination | Do the measures interlock across silos? | much. This is the real strength of AI and tracking. |
The pattern repeats: AI is strong in Coordination and weak in Concerns. In a crisis, though, Concerns is precisely the scarcest factor. Whoever fails to place the program at the forefront of people's minds administers the standstill, no matter how good the analysis is.
The evidence supports this link. A McKinsey analysis of 60 publicly listed companies shows that programs with the highest active workforce involvement (21 to 30%) achieved a +67% excess return over 24 months against the industry benchmark, while programs with minimal involvement landed about 18% below it.1 Involvement addresses Concerns, not Coordination. It is the one dimension no algorithm delivers.
08
A case in practice
09
How to place AI in a restructuring correctly
From all of this follows a sober rule of thumb for using AI in a turnaround.
-
Use AI where speed over data decides. Early warning, analysis, scenario simulation, action ideas. Here the time gain is real.
-
Do not rely on AI where responsibility or negotiation counts. Bank talks, legal assessments, the mobilization of the workforce. That stays leadership work.
-
Make sure all AI-generated content passes through a maturity level. It is an idea, not evidence. Only the maturity level makes it creditor-ready.
-
Check the data foundation. Distorted data produces distorted recommendations. In a crisis, clean data provenance is no luxury.
ChangeMaker® combines both in one platform: AI functions for analysis, proposals, and early warning; the maturity-level workflow for creditor-ready measures; and the 3C method for mobilization under pressure, with live data and automatic bank reporting. Data processing and storage for EU customers take place in Germany (AWS Frankfurt), and the ISMS is certified to ISO 27001.3
10
Used this way, AI in a restructuring is neither hype nor threat, but a tool with a clearly bounded benefit. It accelerates the path from data to insight. The path from insight to substantiated impact is walked by the steering system, led by people with a mandate.
Make change. Not plans.
Frequently asked questions about AI in restructuring
What can AI do in a restructuring?
Does AI replace the Chief Restructuring Officer?
Where are the limits of AI in restructuring?
How does AI-generated content become a bankable measure?
What distinguishes a native AI from a bolted-on one?
Does AI make a restructuring more successful?
Sources
- McKinsey & Company, “Seven percent solution? How many employees should be involved in your transformation?” (2021). n = 60 publicly listed companies, Excess Total Shareholder Return over 24 months against a representative industry and region index; highest excess return at 21 to 30% actively involved workforce. To be read as a correlation, not as guaranteed causation.
- First-party data from ChangeMaker® / Principia Mentis (knowledge base, product and training materials): documented AI functions (drafting work items, action proposals, up-to-date insights, identification of risks and bottlenecks, virtual assistant for project-specific questions); consolidation and reporting effort reduced by up to 85%, on average around 8 days saved per measure. Figures from documented customer programs, not an independent study.
- Principia Mentis, information security management system (ISMS) certified to ISO 27001; data processing and storage for EU customers in Germany (AWS Frankfurt).
- For the framing of the crisis stages: observations from restructuring practice, according to which a substantial share of turnarounds only begin in advanced crisis stages. The 3C method (Concerns, Competencies, Coordination) and the maturity-level workflow are the methodological foundation of ChangeMaker®.
See AI and the maturity level in one platform
In a short demo, we show how ChangeMaker® combines AI functions, the maturity-level workflow, and automatic bank reporting on your specific mandate.