Three business cases, data from the Polytechnic University of Milan, and a product vision. This is what emerged from an afternoon of conversations about real-world cases.
On May 25, 2026, the third edition of Jamio Community Day—the Jamio community's annual gathering with customers, partners, and stakeholders—was held at Copernico Isola for S32 in Milan. This year's theme was " of Change."
There were no product presentations on stage. There was a social cooperative that nearly quintupled its turnover in seven years, an icon of Made in Italy struggling with a decades-old management problem, an IT partner that has made saying no to customers a valuable practice. And—in the second part of the day—one of Italy's leading digital observers, with market data that put everything the speakers had discussed into perspective.
The format—a roundtable discussion moderated by Martin Arborea (Openwork) and Francesca Lopa (DWIT/eGlue), followed by an external interview and a closing speech by the CEO—confirmed the choice of an event that prioritizes substance over form: real stories, measured results, and declared challenges.
KEY DATA FROM THE EVENT
- Aldia Cooperativa Sociale: Onboarding lead time reduced from the 85th percentile of 6 days to an average of 1.5 days, with 98% of hires completed within 5 days.
- Bialetti Industrie: Accounts Payable management (1,000-1,200 invoices/month) transformed from fully manual to semi-automatic. Estimated savings: one week of work per month for the accounting team, with volumes doubled compared to the previous year.
- 75% of large Italian companies have at least one AI project, but less than half have a policy for the use of consumer tools (data from the Milan Polytechnic Observatory 2026).
- 85% of European cloud spending is managed by American operators subject to the US Cloud Act (data from the Cloud Ecosystem & Sovereignty Observatory, Politecnico di Milano).
- Global AI investment 2026: $750 billion versus aggregate revenues estimated at around $60 billion — token costs are still artificially low compared to the real cost of the infrastructure.
The process before the technology
The cases presented converge on a principle that management literature has known for decades but that organizations, even today, continue to overlook: digitalization produces value when it follows organizational analysis, not when it precedes it.
Elisa Saragaglia, CIO and board member of Aldia Cooperativa Sociale, described a systematic approach: organizational analysis, identification of critical issues, definition of needs, and selection of tools. This seemingly obvious sequence is, in practice, the exception, not the rule. When she arrived in 2018, Aldia had 700 employees and a turnover of €15 million. Today, it has 3,000 employees and €70 million. "It's not innovation for its own sake, but rather to support the cooperative's evolution within a highly competitive environment."
The onboarding process —critical for a cooperative that works on contracts and must manage hiring peaks concentrated in just a few days—has become a testing ground. The results, as indicated by the key data, are measurable and confirmed. The secondary benefit, difficult to quantify but likely of comparable value, is the reduction of interfunctional conflict: where previously departments passed on responsibilities, today they work on shared data.
Jessica Mariotti, Italy Accounting Manager at Bialetti Industrie, highlighted an increasingly common problem in mature organizations: highly customized ERP systems that, over time, become a constraint rather than a resource. The request to develop a native connector was impractical. The answer was to seek a lateral approach—an application layer that could fit where the core system couldn't reach, without increasing its technical debt. "It's not easy to assert yourself, to make people understand that this isn't a top-down intervention, but something that serves to improve everyone's lives."
The project—which began as an operational initiative and was subsequently approved by management—produced concrete results in a context where volumes were simultaneously growing. What the literature calls Extended ERP found direct and measurable application here.
DATA
- Aldia: Onboarding lead time reduced from 6 days (85th percentile) to 1.5 days on average — 98% hires completed within 5 days.
- Bialetti: Passive cycle (1,000-1,200 invoices/month) from fully manual to semi-automatic. Savings: one week of work per month for the accounting team, with volumes doubled compared to the previous year.
The business case is built first, not after
One of the most significant methodological issues that emerged from the discussion concerns the timing of building the business case. The prevailing practice—building the business case after the project has been approved—is questioned by all three cases.
The model described by DWIT requires that the business case estimate be an integral part of the analysis phase, preceding the commercial proposal. This shifts the conversation with the client from "how much does it cost" to "what value does it generate"—a distinction that changes the type of relationship and the level of commitment to the project. Both client companies confirmed that the benefits estimated in the preliminary phase were achieved or exceeded in the final results.
The topic then intersected with the discussion on artificial intelligence. Totaro presented an anonymous case study: AI agents deployed for after-sales service at an energy company—functional and scalable, but with token costs exceeding the cost of the human operators they would have replaced, assuming the business model remains unchanged. The business case only holds if volumes grow significantly. Without this, the technological investment incurs costs without the corresponding expected value.
DATA
- In both business cases presented, the benefits estimated in the preliminary business case were achieved or exceeded in the final results.
- Case study: AI agents in the energy sector: token costs are higher than the cost of replacing human operators, assuming the scope remains unchanged. The business case holds up only under conditions of strong volume growth.
Artificial Intelligence: Governance Before Adoption
The AI section produced some of the day's most relevant observations, avoiding both uncritical enthusiasm and prejudicial rejection. The speakers' common ground was: adoption governance, not technology availability, is today's priority challenge.
Data from the Milan Polytechnic Observatory indicates that 75% of large Italian companies have at least one AI project. However, less than half have established policies regarding employee use of consumer tools—a gap that exposes organizations to concrete security and privacy risks.
Martin Arborea introduced the topic of shadow AI —the phenomenon whereby organizations that don't make decisions about artificial intelligence find their employees using it anyway, often with sensitive corporate data on free platforms. He cited a direct case: a salesperson uploaded a confidential quote to an AI tool, indirectly allowing a competitor access to the company's strategy through targeted queries on the same platform. The message wasn't anti-AI. It was: govern before acting.
Piva contextualized the cost dynamics within the global investment landscape: $750 billion invested against estimated aggregate revenues of around $60 billion. The cost of tokens is still artificially low compared to the actual cost of the infrastructure. Economic assessments conducted today could change significantly in the coming years, making the business case built on this foundation more fragile than it appears.
DATA
- 75% of large Italian companies have at least one AI project in progress by 2026, but less than half have defined policies for the use of consumer tools (Osservatorio Politecnico di Milano).
- Global AI investments 2026: $750 billion against estimated aggregate revenues of around $60 billion.
Digital sovereignty: a question of processes, not just data
Stefano Mainetti, Scientific Director of the Cloud Ecosystem & Sovereignty Observatory at the Polytechnic University of Milan—one of Italy's leading digital observatories, which this year incorporated the word "sovereignty" into its name—offered the geopolitical and market framework in a conversation with Martin Arborea.
The structural fact is known but not yet internalized in boardroom agendas: 85% of European cloud spending is concentrated in American operators subject to the US Cloud Act, which provides for the possibility of data access upon declaration of national security interest—regardless of the physical location of the servers. Three recent events have made this risk more tangible: Broadcom's acquisition of VMware, with price increases of up to 200%; the case of an operator with data geolocated in Europe but subject to North American jurisdiction; and the general asymmetry in cloud contract termination clauses.
's position Mainetti is not one of digital nationalism. Sovereignty does not imply technological autarchy, but awareness: knowing which assets are strategically important to keep under legal control, which contractual clauses govern data restitution, and where the points of non-negotiable dependency lie.
Arborea raised a dimension that is still underrepresented in the debate: application sovereignty, or control over company processes and workflows. When an organization outsources its operational logic to an external platform it doesn't govern, the risk of dependency extends beyond data to operational capacity itself. Mainetti acknowledged the scale of the problem, citing the National Cybersecurity Agency's ACN qualification as an increasingly relevant regulatory benchmark for the private sector as well.
It closed with an unexpected market finding: the shares of Salesforce and SAP—enterprise software giants with prescriptive workflows built over decades for millions of customers—have been declining for six consecutive months. The proposed interpretation: the financial market is already betting on a transition to more dynamic and intelligent application models, capable of evolving without requiring rigid and costly implementations.
DATA
- 85% of European cloud spending is managed by American operators subject to the US Cloud Act (Cloud Ecosystem & Sovereignty Observatory, Politecnico di Milano).
- Broadcom's VMware acquisition: Price increases of up to 200% for European enterprise customers.
- Salesforce and SAP stocks have been declining for six consecutive months — the financial market anticipates a transition to more dynamic application models.
The Product Vision: Sovereign Cognitive Orchestration
The closing remarks by Salvatore Latronico, CEO of Openwork, linked the day's stories to a specific product vision, articulating Jamio's development directions for 2026-2027.
Latronico opened with a quote from Carl Schmitt— "Sovereign is he who decides on the state of exception" —applied to the current context: organizations today operate in a regime of permanent exception, with unstable markets, distributed knowledge, and rapidly changing technological dependencies. In this scenario, sovereignty isn't about owning the entire supply chain. It's about being able to govern exceptions when they arise.
The proposed framework— Sovereign Cognitive Orchestration —is based on a premise: AI agents behave as nondeterministic actors, similar to humans, and must be governed with similar tools: access control, performance measurement, cost management, and decision traceability. The difference compared to human governance is quantitative, not qualitative.
Announced development directions include: AI agent orchestration and management capabilities—including agents developed on third-party platforms—natural language application generation (called “natural language building” to distinguish it from vibe coding, which produces ungoverned code), native integration with Power BI for analytics and audit trails, and maintaining the National Cybersecurity Agency’s QC2 qualification.
The most significant strategic element of the project concerned the product development model: the roadmap is built through ongoing dialogue with customers and partners, with a systematic requirements gathering process that translates into generalized functionality—not ad hoc customizations. This model, possible only for a manufacturer with direct relationships with end customers, represents a competitive advantage that is difficult for large international vendors to replicate.
“Having a local manufacturer collaborating with you — we believe that, beyond the technological capabilities we bring to the table, is perhaps one of the most decisive factors.”
Concluding remarks
The evidence gathered at Jamio Community Day 2026 converges on some propositions that deserve attention from business decision makers.
The primacy of process over technology remains a necessary—and often overlooked—condition for the success of digitalization projects. Organizations that choose technology first and then adapt processes systematically face adoption and return-on-investment issues.
The preventive business case: it is not a procedural formality but an organizational alignment tool: building it before starting defines the success criteria in a measurable way and creates the conditions for an objective evaluation of the results.
AI governance: this is today's most pressing and least addressed challenge. The gap between the percentage of organizations with AI projects and the percentage with adequate usage policies represents a concrete systemic risk.
Digital sovereignty: it's a matter of controlling processes rather than data. Organizations that outsource their operational logic to external, ungoverned platforms expose themselves to a form of dependency that no contractual data clause can fully address.
