I wouldn't use the overengineering of a single script influence your view of the technology you named (S3 data laking, lambda pipelines, build in jenkins, terraform to deploy). Those are well understood concepts and technologies to do reporting. The problem here is that these tools were not in use at the company. Once you have 50 scripts run by different people on different schedules in different excel version, having everything go through S3 (immutable, cheap), ingested by lambdas (scalable, well versioned, monitored), built and deployed in jenkins using terraform (vs say, aws command line scripts) significantly reduces the complexity of the problem.