Client. A national public agency in the production area, with procedures that affect companies, importers, and citizens: auto-parts permits, industrial design models, foreign trade, competition oversight.
Approach
A case of two links in the method —the sociology of organizations and software engineering— applied to the place where digital transformation fails most: the State. The rule that governs it is a single one, and it is the antithesis of marketing hype: a broken process is not digitized; it is redesigned first. Computerizing an inefficient procedure only makes it faster in its dysfunction.
That is why the first step was never technical. It was reading the real process —how a case file actually moves, with what delays, which actors, which incentives— and separating it from the process in the manual, which almost never match.
The problem identified
The agency’s procedures were slow, opaque, and manually intensive: paper case files, redundant steps, criteria that lived in people’s heads rather than in explicit rules. The stated problem was “we want to digitize.” The real one was that, underneath, the process itself was undesigned: digitizing it as-is would have frozen the inefficiency into software.
Functional assessment
I modeled the processes with BPMN and UML and a systematic effort in requirements engineering: interviews, observation of the real flow, mapping of actors and decision points. For each procedure —auto-parts, industrial design, foreign trade, competition oversight— the assessment distinguished which steps add value, which are genuine control, and which are inherited friction that could be removed.
Building the solution
- Process redesign before system: simplify the flow, make criteria explicit, remove redundant steps —the part that truly drives efficiency.
- Formal modeling (BPMN/UML) as a bridge between the business and whoever builds the system, so the requirement does not degrade along the way.
- Validation with data (SQL) to ground redesign decisions in evidence, not impressions.
Information and data layer
Digitization turned scattered case files and criteria into structured, queryable information: traceability of every procedure, explicit rules instead of tacit judgment, and a basis for measuring the process itself. In the State, that traceability is not just efficiency: it is transparency and the possibility of audit.
How the work was conducted
The work belongs to a stage well before agentic tooling: the instruments were those of classic process analysis —BPMN, UML, requirements engineering, SQL—, carried by hand. I record it as an honest reconstruction; I do not expose specific procedures or data from the agency. What I assert is the method: redesigning state action from the perspective of those who use it, before installing technology.
What this case proves
- The bridge rule in the State: understand and redesign first; automate afterward.
- Business ↔ technical translator: formal modeling that avoids the “broken telephone” between the area and the system.
- Root of the trajectory: the sociotechnical discipline applied to the public sector, the foundation of everything that came after.