What is CPMAI?
CPMAI stands for “Cognitive Project Management for AI”. The method was developed at Cognilytica and is now part of PMI. It builds on Agile and CRISP-DM and adds governance, ethics and a clear link to business goals. The certification is called PMI-CPMAI™.
Why do so many AI projects fail?
Usually not because of the technology. Typical reasons:
- There is no clear business problem: “we want AI too” is not a goal.
- The data is missing, incomplete, or nobody owns it.
- A prototype works in testing but never in daily work.
- Nobody measures after launch whether anything has improved.
AI projects are largely data projects. CPMAI is built for exactly that.
What are the phases of CPMAI?
- Business understanding: Which problem are we solving? Do we need AI for it, or is automation enough? How do we measure success?
- Data understanding: What data exists, who owns it, and how good is it?
- Data preparation: Clean, combine and label the data.
- Model development: Where possible, start with ready-made, pre-trained models instead of building everything from scratch.
- Model evaluation: Does the model meet the business goals from phase 1, not only technical metrics?
- Model operationalization: Bring it into daily work, monitor it, adjust it.
The phases are iterative: you start small, learn and go back when needed.
Do I need AI at all?
This is the most important question in phase 1. Many tasks can be solved without AI:
- Rule-based, repetitive workflows → classic automation
- Reports on the past → dashboards
- Frequent customer questions → good FAQs and forms
AI pays off when a task needs interpretation, adaptation or large amounts of unstructured data, such as texts, documents or images.
What does this mean for small and mid-sized businesses?
You don’t need your own data science team to think the CPMAI way:
- Start with one concrete problem and one metric
- Use existing tools and ready-made models
- Run a small pilot in real daily work, not only in a test
- Measure after launch and adjust
Is CPMAI only for AI projects?
No. CPMAI was developed for AI, but its principles help with any digitalisation project, whether CRM, automation, a data project or a new website:
- Business problem first: Which problem are we solving, and how do we measure success?
- Check the data early: A CRM is only as good as the customer data in it.
- Start small and learn: a pilot in real daily work instead of a big bang.
- Measure after launch and adjust.
We use CPMAI for AI projects. We apply its principles to all digitalisation projects.
Who do we recommend CPMAI to?
Anyone who leads or supports AI or digitalisation projects: project managers, management, IT and data teams. The method creates a shared language between business and technology. If you want to go deeper, PMI offers the PMI-CPMAI™ certification.
How do we use CPMAI at Kosmo Team?
For AI projects we follow the method, for all other projects its principles:
- In the IT-Check we work through phase 1: where does AI really help, and where is automation enough?
- Before we start, we agree on metrics and measure them after launch.
- We pay attention to data and governance, including GDPR and the EU AI Act.