The MIDAS framework.
Methodology for Integrated Diagnosis, Automation and Selection. It is the methodology the Nexus team applies to ground the decision to automate a process, avoiding return-on-investment promises that real data cannot support.
Developed by the Nexus team and academically peer-reviewed (Val Picado, 2026, Universidad Intercontinental de la Empresa).
From opaque process to defensible metric
The framework articulates five blocks that run in order, each producing an output that feeds the next. Completing them does not guarantee savings: it guarantees that the decision to automate is made on verifiable criteria.
Starting point: organize the company's processes into a neutral taxonomy with five hierarchical levels, from the macro category down to the specific task that is a candidate for automation. Without this map, any later attempt at evaluation is built on quicksand.
Category → Process group → Process → Activity → Task. The MIDAS framework applies it without renaming anything, to preserve traceability with the literature.
A four-phase funnel that progressively discards unsuitable processes until the optimal candidate for a proof of concept is selected. It is what separates truly automatable processes from those that only look automatable.
Phase A: eligibility (5 criteria). Phase B: feasibility × benefit quadrant. Phase C: digital readiness of the data. Phase D: technology typology.
The candidate process is documented in BPMN notation before touching any technology. It makes explicit the human roles, the systems involved and the critical error points worth mitigating.
This is the phase many consultants skip, and the reason so many deployments fail by fitting poorly with real operations.
Turns the manual time of the process into a defensible economic figure: hours/month × cost/hour × 12. It builds the As-Is state and, with analyst judgement, allows the To-Be scenario to be simulated.
The Nexus self-assessment only calculates the As-Is for a methodological reason: the To-Be demands human judgement and real data. Calculating ROI from generic assumptions is not serious work.
Four technology categories according to the nature of the data and the legacy system: low-code RPA, workflow management systems (WfMS/BPMS), platforms with built-in AI (cognitive) and general-purpose LCNC.
The framework does NOT recommend a closed commercial platform: it presents a typology to choose from case by case, with representative platforms for each and an indicative price range.
Change management
It operates across all five blocks and deals with what is almost always underestimated: internal resistance, Citizen Developer training and Shadow IT governance. Without this layer, even the cleanest technical implementation can fail.
Where the framework was tested
The framework is validated on three simulated SMEs representing deliberately different sizes, sectors and levels of technological maturity. The third case is deliberately negative: it confirms that the framework also advises against projects when the investment does not offset the savings.
LogiTrans Levante S.L.
Process: Orders received by email with a PDF template
NovaTalent Consulting S.L.
Process: Onboarding with disconnected ATS, CRM and invoicing
ArteVivo Comunicación S.L.
Process: Monthly social media reports
What MIDAS does not promise
The self-assessment calculates the current As-Is state and projects a range of manual-time reduction based on empirical evidence gathered from the literature. Nothing more.
It does not calculate ROI, payback or absolute savings. Those calculations demand analyst judgement, real process data and a concrete sizing of licenses and implementation hours. Any firm figure without that prior work is an act of faith dressed up as a number.
Nor does it recommend a closed commercial platform. The final selection depends on the existing IT context, budget constraints and team profile; the framework proposes a technology typology with representative platforms, not a purchase.
That restraint is deliberate, and it is the opposite of the sector's usual sales pitch. It is also what makes the framework a useful decision tool rather than a pamphlet.
Academic references
The framework rests on published, peer-reviewed academic literature. The seminal source of the MIDAS framework itself is the last entry in this list.
- APQC. (2024). Process Classification Framework cross-industry — Version 7.4.
- Arno, K. (2025). Business Process Management lifecycle for SMEs.
- Dolk, D. R., & Dolk, K. (2025). Time-Driven Activity-Based Costing for digital transformation.
- Fang, Y. (2025). Low-Code Platforms and the democratization of enterprise development.
- Markevičius, M. (2025). Change management failures in SME digital projects.
- Mou, Y. (2025). Customer expectations and the immediacy threshold.
- Moreira, A., Silva, R., et al. (2025). M4BPA — Methodology for Business Process Automation.
- Pastel, J. (2025). Empirical error-rate reduction after intelligent automation deployment.
- Pervaiz, S., & Ijaz, M. (2025). Adoption case studies of LCNC platforms in SMEs.
- Raković, M., et al. (2025). Shadow IT risks and governance for citizen developers.
- Raza, A. (2025). Audit trails and compliance in digitalized SMEs.
- Simon, F., et al. (2025). The digital divide between SMEs and large corporations.
- Tekleab, A., & Vähäkuopus, J. (2025). API-first integration vs. manual data transfer.
- Tyagi, M., et al. (2025). Reactive vs. proactive management models in SMEs.
- Val Picado, Á. (2026). Diseño de un marco estratégico de automatización inteligente basada en herramientas LCNC para el diagnóstico y digitalización de procesos en PYMES. Trabajo Fin de Grado, Universidad Intercontinental de la Empresa.
- Viswanadhapalli, V. (2025). The Citizen Developer phenomenon: opportunities and limits.
Apply the framework to your own process
The self-assessment walks the five MIDAS blocks over the process you choose. Six guided steps, with a downloadable PDF at the end containing the As-Is analysis, sector benchmark, representative platforms and a roadmap.