Key Takeaways
- AI raises enterprise value: Three out of four companies expect 25 percent more revenue per employee within three years.
- Few dare to redesign workflows: Only 22 percent of companies are willing to rethink their processes fundamentally with AI.
- Administration comes first: Most companies start using AI in administration and knowledge management.
- Meetings become lasting knowledge: AI-based meeting agents help companies secure knowledge and responsibilities from their meetings.
- Productivity starts with processes: AI only has an impact when it meets a solid system and well-designed processes.
- Keep technology simple: AI adoption succeeds when tools are aimed at concrete everyday problems.
Introduction
Does artificial intelligence really increase the value of a company? A new study from the MIT Center for Information Systems Research gives a clear signal: 75 percent of companies expect AI to deliver 25 percent more revenue per employee in the future.
At the same time, only a small share of these companies is changing its existing workflows in a radical way. Most of them use AI first as a helper in administration and knowledge management.
What does that mean for you as a leader in a mid-sized company? It is not enough to plug AI into old routines. My principle applies here as well: switch on your brain first, then the technology. Whoever uses automation deliberately creates real added value and keeps knowledge inside the company.
In this edition I look at the study results, explain why the gap between expectation and redesign matters, and show where you can start in practice.
What This Week's Study Shows
Leveraging Digital Colleagues for Enterprise Value (MIT Center for Information Systems Research)
- 75 percent expect 25 percent more revenue per employee within three years.
- Only 22 percent dare to redesign their workflows radically.
- AI is used first for administration and knowledge management.
The researchers speak of "digital colleagues". The term is well chosen. An AI application is not just another software licence. It takes over tasks, needs a clear role and must be integrated into the team, just like a new employee.
The Gap Between Expectation and Redesign
The numbers reveal a contradiction. Three quarters of companies expect a substantial jump in productivity. Yet fewer than a quarter are prepared to change the way work is actually done.
That gap is where value gets lost. If you add AI to a process that was designed for paper, email and manual handovers, you mainly speed up the old process. Sometimes you even speed up its weaknesses. The real gains appear when you ask a different question: how would we organise this task if a digital colleague were part of the team from the start?
Why do so few companies take that step? In my experience there are three reasons. People hold on to familiar routines. There is uncertainty about the technology and about how employees will react. And nobody has the explicit task of redesigning the process. AI then remains a tool for the familiar, and its potential stays largely untapped.
Why Administration and Knowledge Management Come First
It makes sense that most companies start in administration and knowledge management. Here you find many recurring tasks: filing, summarising, answering standard requests, finding documents. Results become visible quickly, and the risk is manageable.
A good example is the meeting agent. Many meetings end without a clear output. Decisions are made, but a week later nobody remembers who agreed to do what. An AI-based meeting agent records the conversation, summarises the decisions and assigns tasks. Knowledge that used to disappear after the meeting becomes a resource that the whole team can use.
Such quick wins are valuable. They build trust and show employees that AI relieves them instead of replacing them. But they are a starting point, not the destination.
Productivity Starts with Processes, Not Tools
Whether AI brings value depends far less on the tool than on the problem it solves. Technology is rarely the problem. People and processes are.
Take email. AI can sort, summarise and draft replies. But if your inbox has no system, if everything stays in one folder and nobody decides what happens to a message, AI will not fix that. It will only produce a faster version of the same chaos. First you need a clean system, then automation makes sense.
The same applies to software development with AI. Three prompts are rarely enough for secure software. Without clear requirements, structures and security rules, AI-generated applications carry real risks. Here, too, the order matters: think first, then automate.
And it applies to prompting itself. AI is only as good as its briefing. Unclear instructions and unspoken assumptions lead to poor results, just as they would with a new employee.
How to Get Started
If you want AI to increase the value of your company, I recommend four steps:
- Pick one real problem: Choose a task that costs your team time every week, not the most impressive use case.
- Clean up the process first: Describe how the task should run, who is responsible and what the result looks like.
- Give the digital colleague a role: Define what the AI does, what it does not do and who checks the output.
- Measure the effect: Compare time, quality and errors before and after. Only then scale the solution.
Mentioned Tools & Resources
- MIT CISR study "Leveraging Digital Colleagues for Enterprise Value": Research on how companies use AI applications as digital colleagues and what value they expect.
- Meeting agent: AI assistant that records meetings, summarises decisions and tracks tasks.
- TJ's Technology Tuesday: A 15-minute digital impulse for the C-level every Tuesday from 12:30 to 12:45 CET, with recordings available on LinkedIn.
- digital4productivity podcast: The same topics as an audio format for listening on the go.
Frequently Asked Questions
How do digital colleagues increase enterprise value?
Digital colleagues, meaning AI applications that take over defined tasks, raise productivity by automating processes. According to MIT CISR, 75 percent of companies expect 25 percent more revenue per employee within three years. The value comes from adapted processes, not from the technology alone.
Why do only 22 percent of companies redesign their workflows with AI?
Many companies hold on to familiar routines and are uncertain about technical and human factors. Often nobody is explicitly responsible for redesigning the process. As a result, AI is used for old tasks and its potential remains largely unused.
Where do companies usually start with AI?
Most begin in administration and knowledge management. These areas contain many recurring tasks, results become visible quickly, and the risk is manageable. That makes them a good entry point for building trust.
What does a meeting agent do?
A meeting agent records conversations, summarises decisions and assigns tasks automatically. This ensures that discussions lead to concrete actions and that knowledge stays in the company instead of being forgotten after the meeting.
Why does AI need a good system first?
AI amplifies the process it works on. If the underlying process is chaotic, for example an unstructured inbox, automation only makes the chaos faster. A clear, well-structured workflow is the basis for meaningful AI use.
What makes a good AI prompt?
A good prompt is precise and contains all relevant information: role, context, goal and the expected format. Unclear instructions and hidden assumptions lead to misunderstandings and weak results, just as with a new employee.
Conclusion
The expectations are high: three out of four companies count on AI to raise revenue per employee significantly. But only those who are willing to rethink their workflows will see that effect. Start with one real problem, clean up the process, give your digital colleague a clear role and measure the result. Technology is rarely the bottleneck. People and processes are. If you would like keynotes, training or C-level sparring on AI, Microsoft 365 or the iPad, feel free to get in touch.
