Building AI analysis and portfolio synthesis: prioritize without overwhelming teams
Health score, recommendations, BACS analysis, and organization synthesis: useful AI ranks actions; it does not replace the manager.
Octowise Team
Useful AI ranks actions; it does not replace the manager. Octowise combines three layers: statistical signatures on sources, per-building analysis (health score, risks, recommendations, BACS), and portfolio synthesis at organization level. Each layer has data prerequisites, and relaunches are explicit.
Three layers, three questions
The signature answers: “is this source leaving its usual regime?”. Building analysis answers: “what should we look at first on this site?”. Organization synthesis answers: “which branches or territories concentrate the risk?”.
Mixing them into a single “magic score” produces unexplainable rankings. Octowise keeps them separate, including BACS analysis, which is not a duplicate of building analysis: it relies on the structured register and a confirmed ADEME DPE.
Building analysis: an actionable score
Health score, risks, opportunities, and recommendations are generated from building data and the linked DPE. Copy follows the user’s language. Internal identifiers and GUIDs must never appear in a synthesis: we talk about building names, not technical keys.
If analysis has never been launched, the empty state is not a load error. The user decides when to run or rerun, for example after updating sources, DPE, or the BACS register.
- Health score, risks, opportunities, operational recommendations.
- Separate BACS analysis, gated on confirmed DPE and the register.
- Explicit relaunch: no silent recompute that overwrites a review in progress.
- Localized syntheses, with no technical identifiers exposed.
Organization synthesis: prioritize the portfolio
The portfolio view aggregates analyses by org-chart branch. That is the executive layer: where to concentrate investment, which action plans to open, which sites are too poorly documented to be analyzed honestly.
A null synthesis (no analysis yet) is distinct from a technical failure. On web and mobile, the absence of analysis is an empty state, not an error toast.
AI does not invent missing data
Without source history, no robust signature. Without confirmed DPE, no BACS analysis. Without an org chart, portfolio synthesis has no reporting grain. AI makes gaps visible; it does not paper over them.
That is deliberately less spectacular than a copilot that “knows everything”. It is far more useful in a steering meeting, where each recommendation must be traceable to portfolio data.