Newforma buys Lumi to give project questions a shared meaning

Newforma announced its acquisition of Toronto-based Lumi AI on 21 October, adding an analytics engine to its software for architecture, engineering and construction. Financial terms were not disclosed. Lumi’s engine is intended to connect plain-language questions with the meanings of the underlying project records, while showing how it reached an answer.[1]

The integration is still ahead: Newforma says Lumi will extend its Vojo assistant and power analytics in Konekt’s project dashboards, with rollout details to follow. Proposed applications include spotting delayed responses to requests for information and comparing a project’s submittal cycle with a firm’s historical baseline.[1]

What does “late” mean?

Imagine asking which consultant is responding late. A useful answer needs a definition: late against the contractual deadline, the project team’s target or the consultant’s usual turnaround? It may also need to distinguish an original request from a revision. Those are different questions even if all the records contain a date and a company name.[2]

Lumi’s existing setup documentation shows the practical work behind its semantic layer. Administrators choose the tables and fields the system can query, give them understandable names and describe what they mean. Consistent names help the system connect records across tables; explicit units and definitions for status values reduce ambiguity.[2]

The company’s knowledge-management tools also let teams define relationships between tables and apply filters. They show basic information about each field’s contents, including missing values and the range of values present. Those are useful ingredients for checking what a query is counting before trusting its total.[3]

A construction team could therefore agree the definition of response time once and apply it across repeated questions. In our example, counting calendar days from first issue is different from counting working days from the latest revision. An apparently precise average can answer the wrong question if that choice remains implicit. The semantic work makes the choice available for inspection.[2]

An assistant with project context

Newforma introduced Vojo in May, alongside a broader roadmap covering natural-language search, queries across projects and analysis of models and documents. That earlier announcement positioned the assistant within the company’s existing project-information systems. Lumi adds a specialised route from a question to structured analysis.[4]

The acquisition’s practical appeal is that project managers already know the records against which an answer can be checked. A useful analytics system should let them follow a finding back to the requests, dates and decisions behind it. That shortens the distance between a dashboard and the project conversation it is supposed to inform.[3]

Report an error or suggest a correction