How KPMG scaled learning to 100,000+ learners inside Salesforce

Every other story in this series is anonymized. This one is published with a name on it. KPMG, the global professional services firm, took on a hard mission: build and roll out digital training for public healthcare organizations across Italy's Lombardy region. The programme ran on Daniwoo LMS, inside Salesforce, and the published case study reports four results. More than 100,000 learners onboarded. A 92% course completion rate within the first three months. Over 120,000 hours of training delivered with real-time, SCORM-compliant tracking. Administrative workload reduced by 30%.

This article is the blog version of that case study: the mission, the evaluation, the four-phase build, and what the numbers mean for anyone planning a large-scale deployment. It exists because the rest of this series makes architectural claims, and claims at this scale deserve a named, verifiable anchor. Everything here is traceable to the published case study, which you can download in full.

The mission: 100,000 learners in public healthcare

The engagement gave KPMG a hard target: design and deploy a digital training initiative for public healthcare organizations in Lombardy, reaching more than 100,000 learners. Public healthcare is one of the most demanding environments a learning programme can land in. The case study lists the constraints plainly: strong security, compliance, and reporting requirements, a highly regulated public-sector environment, a distributed learner population, and tight deadlines.

The shape of the problem mattered as much as its size. This was a rollout across several health bodies, not one. Learner profiles varied widely across those bodies. And the deadlines left no room for a long build.

The day-to-day need underneath all of it was consistency. Training had to be digital, tracked precisely, and reported centrally. And it had to achieve that without adding technical complexity or fragmenting data across systems. In a regulated public environment, a completion that cannot be reported centrally, with audit-grade confidence, might as well not exist.

That mix of scale, regulation, spread, and speed is what makes this deployment worth writing up. It is the stress test that most LMS architecture debates only simulate.

The programme sits inside the training effort around FSE 2.0, the Fascicolo Sanitario Elettronico 2.0: Italy’s national electronic health record initiative. Regional FSE 2.0 adoption and training workstreams are funded by the European Union through NextGenerationEU, under Mission 6 (Health) of Italy’s PNRR recovery plan. That funding context explains the bar: a publicly funded, EU-backed programme is audited by design, and its training reporting is not optional.

Why the evaluation ruled out an external LMS

The deadline did the maths. Every risk in KPMG's published assessment costs weeks: integration overhead between the LMS and Salesforce is an integration to build, data synchronization complexity is a sync layer to debug, increased operational risk is a review a regulator can fail, and longer implementation cycles were the one thing the calendar could not absorb. The published conclusion: a traditional external LMS would have fought every constraint at once.

Readers of this series will recognize that list. It is the integration tax that the native versus integrated comparison documents in the abstract, showing up as a concrete call made by a global consulting firm. When learner data must be tracked precisely and reported centrally, every sync boundary is a place where precision can quietly die.

So the requirements KPMG retained pointed one way. A 100% native Salesforce architecture, with no external platform in the data path. Experience Cloud integration, to reach learners spread across many bodies. SCORM support with audit-grade reliability, because compliance reporting depends on trustworthy completion data. Multi-audience management at large scale. Automated workflows for surveys and feedback. And rapid deployment, because the deadlines were not moving.

That list reads like a table of contents for the complete Salesforce-native LMS guide. The difference is that here, each line was a hard requirement on a live public-sector programme, not a talking point.

The build: four phases to production in weeks

The build went from kickoff to production-ready in just a few weeks, through four phases documented in the case study. The compression came from where the work happened: setup in a Salesforce sandbox, on a platform that ships its workflows preconfigured, rather than an integration project between two systems.

Phase 1: Architecture and alignment. Technical validation, the access model, and data governance mapping. In a public healthcare context, deciding who can see which learner data is not a detail to defer. It is the foundation everything else inherits.

Phase 2: Sandbox configuration. SCORM validation, multi-audience structuring, and workflow automation, all configured and tested in a sandbox before any learner touched the system. This is where the SCORM playback and tracking architecture was proven against the programme's own content, not assumed from a datasheet. The multi-audience structure, serving varied profiles across several health bodies through Experience Cloud, follows the model the partner and customer academies guide describes.

Phase 3: Stakeholder testing. Reporting validation, user acceptance testing, and governance. The stakeholders who would own the reporting confirmed the reporting, before launch rather than after.

Phase 4: Production rollout. A controlled launch with monitoring, support, and adoption tracking, rather than a big-bang release to 100,000 people at once.

The case study credits the speed to the teams working side by side: sandbox testing, content checks, rollout prep. Nothing on the delivery list required building an integration, because there was nothing to integrate, and that absence is the whole few-weeks story. Daniwoo set the platform up for multi-audience learning paths, SCORM completion validation and tracking, automated surveys and feedback, and real-time reporting inside Salesforce.

The results: what the programme delivered

Per the published case study, the programme delivered four headline results. These are results from one project, published by us with KPMG's quote attached. They are not industry statistics, and this article will not dress them up as any.

100,000+ learners onboarded, across multiple public healthcare organizations. The mission's target was met: learners spread across many separate bodies, onboarded onto one platform.

92% course completion within the first three months. Completion rates are shaped by many factors, and the case study does not break this one down. What it establishes is that at this scale, in this setting, the programme reached 92% inside a quarter.

120,000+ hours of training delivered, with real-time SCORM-compliant tracking. Every one of those hours was tracked as it happened, inside Salesforce, against the audit-grade requirement the programme set. Readers who want to understand how tracking architecture holds up under this kind of write load should read the tracking at scale guide. The short version: it holds when the design respects the platform, and this deployment is the published proof.

Administrative workload reduced by 30%, through multi-audience management and native reporting. This is the number that survives longest after a launch. Onboarding is an event; administration is forever. A 30% cut, sustained across every body on the programme, is the difference between a programme that scales and one that consumes its team.

The case study's own summary line is the architectural conclusion of this whole series, stated by the deployment that tested it: "The LMS became an integrated component of the Salesforce environment, not an external dependency."

And the partner's verdict, verbatim:

The full numbers, the build detail, and the programme context are in the complete case study, available for download.

What this proves for Salesforce partners

For systems integrators and consulting partners, this project shows a split that works: the partner keeps its domain, and the LMS handles learning natively. KPMG kept its domain: the mission, the client, the healthcare expertise. Daniwoo handled the learning layer inside the Salesforce environment KPMG was already delivering. Neither party had to become the other.

The case study draws four lessons, and each one is money for a partner. No sync layer means nothing to underwrite for the life of the engagement. Native scale means the platform never caps what the programme can attempt. Sandbox setup and preconfigured workflows replace integration cycles, so delivery is weeks, not quarters. And the LMS ships as a managed package inside the org the partner is already building, so the engagement stays whole.

That last point is the one worth sitting with. For an SI, an external LMS in the solution design is a second vendor, a second data model, and a second support boundary inside your delivery. A native LMS is a package in the org you already govern. The engagement stays yours.

Gianluca Carletti's quote above describes the working relationship that made the rollout possible: responsiveness and Salesforce expertise, operating as an extension of the delivery team. That is the model we run with partners, and it is documented on the partner page for firms considering the same structure. KPMG is featured among our customers.

FAQ

Is Daniwoo really 100% Salesforce-native?

Yes. Daniwoo is installed in your Salesforce org as a managed package, and learning data lives as records in your own database, with no external platform in the data path. A 100% native architecture was an explicit KPMG requirement, and the case study's conclusion states the result: an integrated component of the Salesforce environment, not an external dependency.

How long did the KPMG deployment take?

The platform was production-ready in just a few weeks, per the published case study. The rollout ran in four phases: architecture and alignment, sandbox configuration, stakeholder testing, and a controlled production rollout. The case study credits the speed to close collaboration on sandbox testing, content validation, and rollout preparation, plus preconfigured workflows that cut out integration work entirely.

How does SCORM tracking hold up at 100,000 learners?

The published case study reports 120,000+ hours of training delivered with real-time, SCORM-compliant tracking, against audit-grade reliability requirements set by a regulated public healthcare environment. Tracking at that scale is an architecture question before it is a feature question. The tracking at scale guide explains the design rules that make it hold.

Can one org serve multiple audiences, like employees and external learners?

Yes. The KPMG programme served varied learner profiles across many healthcare bodies from one deployment, using Experience Cloud and large-scale multi-audience management. The same model serves employees, partners, and customers side by side in a single org. The partner and customer academies guide covers the architecture and the audience structuring.

Conclusion

One deployment answers the question every architecture debate in this series circles: does native hold at enterprise scale? Per the published case study, it held: 100,000+ learners, 92% completion within three months, 120,000+ tracked training hours, and a 30% lighter admin load. It did so in one of the most regulated settings there is. And it went live in weeks, inside a partner-led programme, as a component of the Salesforce environment rather than a dependency outside it.

The blog version stops here; the full story doesn't. Download the complete KPMG case study for the programme context, the implementation phases, and the results in detail. And if you are planning a deployment at scale, partner-led or direct, book a demo and bring your numbers.

About the author. Brice Mbouani is the founder of Daniwoo and a Salesforce engineer. He designed and built the Salesforce-native learning platform this series is drawn from, which runs in production orgs at customers including ENG Group. The deployment described in this article is documented in the published case study it draws from.