Faster, higher-quality analyses and sharper longlists

Harald Friedrich is responsible for strategy development at VINCI Energies Industry & Infrastructure in Germany; around 70% of his work is dedicated to M&A, with the goal of closing strategic gaps in the business areas in a targeted way.

The central bottleneck before StrategyBridgeAI was manpower: every search mandate required setting up a dedicated team, since most of the research is done in-house and only partly outsourced. The databases in use were not AI-supported and were heavily oriented around NACE codes, which quickly made searches cryptic and made fine-grained scaling difficult.

The strategy team has been using StrategyBridgeAI since the first quarter of 2026. The platform is used both for market analyses and for targeted company searches – on the buy side to identify target companies in specific segments, but also on the sell side, where previously unnoticed opportunities can be found.

The Hawkeye report (outside-in business analysis) is used for decision papers as well as the team's own market studies, which are also further processed by the analyst team in Paris – alongside other tools used there, such as Orbis, PitchBook, or Capital IQ. It is highlighted positively that StrategyBridgeAI's results are not purely automated, but additionally quality-checked by humans – this extra control step is seen as a clear contribution to result quality.

A clear advantage over the previous, NACE-code-based databases is that the search is not bound to rigid classifications, allowing more specific searches for companies that are actually relevant.

Particularly positive feedback is given for the blueprint field: the user simply enters, in the search field, what kind of companies they are looking for. The AI then researches matching companies and refines a “search blueprint” from this input – i.e. the concrete criteria the search should be based on. This blueprint is visible to the user at any time and can be corrected and further developed, so that the search input improves with every iteration.

Two points are on the team's wish list. Company searches initially return worldwide results, with no direct way to narrow them to a single country such as Germany from the outset – so longlists come back larger than needed and have to be filtered afterwards. And for very generic search terms, competitor delineation could be finer: genuine competitors are then sometimes hard to distinguish from non-competitors within a large result list.