How AI drafting, terminology enforcement, domain expert review and QA audit combine into a single accountable delivery process.
- Why "we use AI" tells you nothing
- The eight steps
- What compounds
Why "we use AI" tells you nothing
Almost every language provider now uses AI somewhere. The claim by itself is not a differentiator — what matters is where it sits in the process, what checks follow it, and who is answerable for the result.
The meaningful question to ask a vendor is not whether they use AI, but what happens to the AI output before it reaches you.
The eight steps
Client submission, AI draft generation, terminology matching, domain expert review, LISA QA 3.1 audit, format restoration and delivery, terminology extraction, continuous optimization. Two of those deserve attention.
Domain expert review is the step that converts a draft into a deliverable. It is performed by someone with background in the subject, not a generalist proofreader — because the judgement required is technical, not linguistic.
Terminology extraction is the step most providers skip. Every completed project feeds the terminology base, which means the asset improves with use rather than being rebuilt each time.
What compounds
The first project with a new client is the hardest and the least consistent, because nothing has been established yet. By the third, the termbase carries most of the decisions and both consistency and turnaround improve.
This is why we treat terminology as a deliverable rather than a by-product. It is the part of the work that keeps paying after the files are delivered.