# AI in Management Strategy Work

Author: Marin Niehues, Niehues Consulting · Published: 2026-04-09 · Updated: 2026-11-16 · AI Strategy & Business Case
Source: https://www.niehuesconsulting.com/blog/ai-in-management-strategy-work/

AI improves management strategy work in condensing information and calculating scenarios, while the decision and its execution remain the task of leadership. For executive management this raises an organizational question: which steps of strategy work can be supported systematically, and what must the company put in place so that better analysis leads to better results in the market? This article sets out the fields of application, describes a delivery approach and names the decision criteria.

This article treats AI as a tool of strategy work. How a company sets up its own AI strategy with goals and a roadmap is described in the article [AI Strategy for Companies](/blog/ai-strategy-for-companies/).

## What is AI useful for in strategy work?

In strategy work, AI is useful above all for building the situation picture and evaluating options, because that is where large volumes of documents and figures must be condensed. Strategy work proceeds in stages, and AI acts differently at each stage:

- Situation picture: market data and competitor monitoring are condensed together with internal figures into a shared fact base.
- Options: courses of action are developed and distinguished from one another.
- Evaluation: assumptions are quantified and scenarios calculated.
- Decision: management chooses and sets priorities.
- Execution: the chosen direction is translated into initiatives with responsibilities and metrics.

AI has the strongest effect on the situation picture and evaluation. Language models condense extensive reports into structured overviews, provided every figure is tied to a verified source, and analytical methods run scenarios with changing assumptions faster than a team can by hand. For options, a model formulates counter-arguments to the preferred variant on request. That improves discussion in the leadership team, because assumptions are on the table earlier. The decision itself stays with management, since it demands judgment under uncertainty and a willingness to take responsibility. In execution, AI supports the tracking of metrics and the preparation of steering meetings, so that leadership spends its time on deviations and decisions.

Four applications show what this looks like day to day:

- Competitor monitoring: a model evaluates competitors' annual reports, press releases and job advertisements and reports changes in offering and pricing, each with a reference.
- SWOT analysis: the model produces a first draft from internal figures and the situation picture, which the strategy team checks, weights and supplements with its own insights.
- Scenario planning: the team defines the driving uncertainties, the model drafts consistent scenarios for them and calculates their effect on revenue and earnings.
- Red-teaming the preferred option: the model is instructed to attack the favored option systematically and names the assumptions on which it would fail.

**AI shortens the time between a strategic question and a robust fact base. The value arises when the organization translates the speed gained into decisions and execution.**

## Where are the limits of AI in strategy work?

The limit lies in tasks that demand judgment about new situations and in figures without a verified source. A field experiment with 758 consultants at a large strategy firm measured this precisely. On tasks within the model's capabilities, participants using AI completed 12.2 percent more tasks, worked 25.1 percent faster and produced results of more than 40 percent higher quality. On a task outside those capabilities, they were 19 percentage points less likely to reach the correct solution with AI than without.

For strategy work, this implies a clear division of labor. Condensing, structuring and drafting belong to the model. Judging whether an insight is correct and what it means for the business belongs to the strategy team. Market sizes and growth rates are particularly delicate: language models produce plausible-sounding figures that appear in no source. Every figure in a paper for executive management therefore needs a reference that a person has opened. How this risk can be contained in operation is described in the article [Hallucinations: Limiting Risks in Productive Use](/blog/ai-hallucinations-in-production/).

**Checklist for AI-supported strategy papers**
- Every figure is backed by a reference that a person has opened.
- Assumptions are marked as assumptions and given a range.
- The model has attacked the preferred option deliberately at least once.
- A named person has approved the paper.
- Confidential documents have entered only the approved environment.

## A model for introduction

Anyone bringing AI into strategy work does best to proceed in three steps. The model follows the stages Understand and Build, followed by Introduce. It can be applied to any strategic question.

- Understand: the strategy process is mapped. Who produces which analysis, with what effort and at what quality, and on what date is the decision taken? From the map come a business case and a sequence of applications.
- Build: a prototype works on a concrete strategic question in the real workflow, for example the competitive analysis for one business area. The strategy team works with it, and the benefit is measured against the previous effort.
- Introduce: clear responsibilities and fixed review standards, combined with targeted enablement, turn the prototype into a routine process that the business adopts.

The sequence has a reason. A prototype without prior process analysis answers a question nobody asked. An introduction without a prototype asks management for a commitment before the benefit is visible in its own workflow. The approach creates both: an evidence-based foundation for decisions and an organization that actually uses the result.

## Criteria for selecting applications

First place belongs to the application that improves an upcoming decision and whose results can be verified. Five criteria have proven useful for selection:

- Effect on decision quality: does the result change a decision the company will actually take in the coming months?
- Verifiability: results must be traceable to sources before they enter a paper for executive management.
- Data situation: the required documents and figures exist and are accessible.
- Confidentiality: strategy data are among a company's most sensitive information. Executive management decides in advance which documents a model may see.
- Link to the strategy calendar: the application engages at a fixed date in the planning cycle.

How to weight these criteria against the company's other AI initiatives is described in the article [Prioritizing AI Use Cases](/blog/ai-use-case-prioritisation/). The following overview shows how the three most common organizational forms of use compare.

**Organizational forms of AI use in strategy work**
Criterion | Use by individuals | Central analytics team | With Niehues Consulting
Effect on decisions | Inconsistent, hardly measurable | Selective, varies by topic | Measurable and aligned with the strategy calendar and decision dates
Quality assurance | Individual | Defined, but detached from management | Review standards and source rules embedded in the workflow
Data confidentiality | Unmanaged | Partly regulated | Data release set in advance by executive management
Scaling | None | Limited | From prototype to routine operation with clear responsibilities

## Why execution determines the benefit

A better analysis on its own changes nothing in the market. Strategies reach their goal when they are translated into initiatives with named owners as well as resources and deadlines. The article [Strategy Execution: Why Strategies Fail in the Organization](/blog/strategy-execution-why-strategies-fail/) describes this transition as the real weak point of many programs.

AI sharpens this situation in one respect: it delivers analyses faster than a decision path with monthly committees can process them. Where decision rights are unclear or steering meetings take place only every few weeks, the time gained in analysis goes unused. The question for executive management is therefore whether the decision path can absorb the new speed. From the delivery perspective this means: every AI application in strategy work is given a recipient and a date as well as a metric by which its contribution can be read.

**Four review questions before starting**
- Which strategic decision is due in the coming months, and which analysis prepares it?
- Who is accountable for the quality of AI-supported results?
- Which documents may enter a model, and where is the model operated?
- By which metric will executive management recognize that the effort pays off?

## Niehues Consulting as a partner for AI in strategy work

The task sits at the interface between AI management and organization. [Niehues Consulting](/) supports executive management and division heads from business analysis through the prototype to introduction into routine operation. Marin Niehues works from the delivery perspective: he builds organizations that deliver and brings AI into operation. His experience comes from projects ranging from VC-funded scaleups to DAX 40 corporations and from assignments in 20+ countries on six continents.

The approach follows the Niehues Delivery Model. Three of its review questions matter most for strategy work: which business result should change measurably? Does the prototype prove its benefit in the real workflow with real data? Who is accountable for operation from day one? In concrete terms this yields an analysis of the strategy process, a business case for the selected applications, a prototype on a real strategic question and the transfer into operation with enablement of those involved. How this fits into AI management as a whole is described in the article [AI management: what executive leadership must steer](/blog/ai-management-what-executives-must-steer/). The entry point is a free 30-minute initial consultation, which you can arrange via [the contact form](/#kontakt).

**Recommendation for executive management**
Start with one concrete strategic question and a clear acceptance: understand the process, build a prototype in the real workflow, then introduce routine operation with responsibilities and metrics. That is how the speed of analysis becomes speed of decision.

**Sources**
- Dell'Acqua, McFowland, Mollick et al.: Navigating the Jagged Technological Frontier, Harvard Business School Working Paper, 2023

## Frequently asked questions

### Can AI make strategic decisions?

AI prepares strategic decisions, and management takes them. It condenses information, calculates scenarios and formulates counter-arguments. The decision on direction and priorities demands judgment under uncertainty and personal accountability.

### Which AI tools are suitable for strategy work and competitive analysis?

Three classes of tool suit strategy work: general language-model assistants under an enterprise license, research tools that cite their sources and analytical tools for scenarios and figures. Assistants such as ChatGPT or Microsoft Copilot condense documents and produce drafts. For competitive analysis, what counts most is that every statement points to a reference and that confidential data stay in an approved environment.

### How do you use ChatGPT for a SWOT analysis or scenario planning?

For a SWOT analysis, give the model the verified situation picture and your own figures and have it draft each quadrant, which the team then weights and supplements. For scenario planning, the team defines the two driving uncertainties, and the model drafts four consistent scenarios with early indicators. In both cases the rule holds: figures come from your sources, the model structures and formulates.

### Will AI replace strategy consultants?

AI replaces a large share of the research and condensing work that used to absorb consulting hours. Judgment about new situations, moderating between interests in the leadership team and accountability for execution remain human tasks. The field experiment with 758 consultants also shows that blind trust in the model lowers accuracy on difficult tasks.

### Which parts of strategy work are suited to AI first?

The situation picture and evaluation offer the quickest benefit. There, many documents and figures are available for a model to condense, and the results can be checked against sources. The selection in each case follows the criteria above.

### How long does it take for AI to have an effect in strategy work?

A prototype on a concrete strategic question shows its benefit within six to ten weeks. In routine operation, AI takes effect from the next planning cycle, once recipients and dates are fixed. Without a fixed slot in the strategy calendar, the prototype remains a one-off trial.

### How do strategy data remain confidential?

Executive management decides in advance which documents a model may see and chooses the operating model accordingly. Access rights and logging, together with clear responsibility for approval, are part of the introduction concept.
