The “AI-washing” problem in the market
Since 2024, practically every technology provider claims to have AI capability. Slide decks, projects rebranded as “AI-powered” and teams who yesterday were doing something else. For a CIO in a German utility this is a real issue: most of that capability is borrowed, subcontracted or bought on the market from profiles that churn every few months.
The question that is rarely answered with facts is a simple one: how does a consultancy actually build real AI capability that holds up over time, without depending on the hiring lottery? At Principal33 our answer rests on three pillars we have been working on for years: structured internal training, a Senior-Junior development model, and active participation in industry events and community.
“In AI, the differentiator is not the model you use — it is the people who know when, where and why to apply it.”
Why building AI capability is different from buying it
Buying AI talent on the open market has three problems for a partner working with German utilities. First, cost: AI profiles with real experience in energy or SAP IS-U are expensive and scarce. Second, volatility: attrition in senior AI profiles in 2024 and 2025 remains well above other technologies. Third, lack of sector alignment: a generic data scientist does not understand MaKo, GPKE or the regulatory weight of the Bundesnetzagentur or BSI.
That is why building capability internally is not a romantic decision: it is an operational one that protects the client from market churn.
The three pillars of real AI capability
Three pillars sustain how we build AI capability internally, instead of renting it from the market.
01 · Structured internal training
A continuous programme combining ML, GenAI and MLOps fundamentals with utility-sector specialisation: forecasting, anomaly detection, billing incidents, customer communication.
02 · Senior-Junior development
High-potential juniors developed on real projects under senior mentorship. Applied learning, accumulated knowledge, stable client relationships and sustainable margins.
03 · Events & community
Regular presence at E-world energy & water, SAP for Utilities and applied-AI conferences. Learning, validating and contributing back through articles and university collaborations.
Pillar 1 — Structured internal training
Our AI training is not a series of disconnected courses. It is a continuous programme that combines technical fundamentals (classical machine learning, deep learning, GenAI, MLOps) with sector specialisation. For utilities that means training the teams on real sector cases: demand and generation forecasting, anomaly detection in meter readings, billing incident classification, customer communication on complex invoices.
The programme runs on three complementary formats: periodic technical sessions led by seniors, hands-on labs on sector datasets, and cross-project peer reviews on real cases. Nobody is certified in AI just by attending a course; internal certification requires having worked on a real case under senior supervision.
Pillar 2 — Senior-Junior talent development
Instead of buying expensive seniors on the market, we bring in high-potential juniors and develop them on real projects under senior mentorship. This model has three direct consequences.
First, learning is applied, not theoretical. Juniors learn by solving real problems for real clients, not in a classroom. Second, knowledge accumulates inside the company: when a junior grows into a senior, they carry the sector memory forward instead of exporting it to another employer. Third, the client relationship is preserved: teams stay stable because career growth happens within the company.
The model also allows us to keep margins sustainable without inflating the client invoice with senior-priced profiles for tasks a well-mentored junior can handle perfectly.
Pillar 3 — Events and community
A consultancy that wants to talk about AI with credibility cannot stay inside its own walls. We take part regularly in reference events for the sector, such as E-world energy & water, SAP for Utilities gatherings and technical conferences on AI applied to regulated industries. These forums serve two purposes: we learn what others are doing and we validate our hypotheses in conversations with real clients, not in internal meeting rooms.
We also contribute back to the community through technical articles, panel participation and university collaborations. This is not marketing: it is how we keep the teams exposed to the state of the art and how we attract talent that wants to work on serious projects, not lab pilots.
Why this matters to a German utility CIO
For a German utility evaluating partners for its AI strategy, the relevant question is not “do you have AI capability?”. Everyone will say yes. The useful question is “how do you build and maintain that capability over time?”.
The answer to that question predicts three things: whether the teams starting the project will still be on it in eighteen months, whether sector knowledge accumulates or leaks, and whether the commercial proposal is sustainable or dependent on subcontracting to third parties.
Where Principal33 fits
At Principal33 this model has been running for years and is the reason we can offer senior, German-speaking, nearshore DACH teams with a low attrition rate. Our AI capability is not a product we bought: it is a capability we have built internally, calibrated specifically for the needs of the German utilities sector and for the regulatory framework it operates in (Bundesnetzagentur, BSI, KRITIS, GDPR).
The shift in perspective
In AI, sustainable competitive advantage does not come from access to the latest model — that is now a commodity. It comes from having people who understand the sector, who know when to apply AI and when not to, and who have been working together for years. That kind of capability cannot be bought: it has to be built. And building it takes time, commitment and a certain obsession with the quality of your own team.
