From consequential work to governed AI production
Who We Are
LevelUp360 helps organisations redesign selected business workflows for AI and carry them into governed production.
LevelUp360 is designed for implementation providers and internal transformation teams operating across business, architecture, security and delivery. The focus is regulated and evidence-intensive environments where an AI system must be explainable, reviewable and owned after it goes live.
Clients can enter at the point where they are blocked: deciding whether to proceed, redesigning the selected business process and workflow, defining the controls and evidence, carrying the system into production, or providing architectural and delivery oversight while their team builds.
LevelUp360 combines bounded specialist services with Govern360, the product that makes the repeatable workflow discovery, readiness, evidence and governance work more consistent and reusable.
The difficult part of AI adoption is often not the model. It is making the work clear enough for AI to participate reliably.
A procedure may be documented while important dependencies, exceptions, decisions and handoffs remain in people’s heads. Different teams may understand the same process differently. People can usually find an experienced colleague, ask for clarification and work around what is missing. An AI system cannot safely fill those gaps for itself.
Automating an incomplete workflow scales the weakness already inside it. LevelUp360 starts by understanding how the work actually operates, redesigning what needs to change and deciding where deterministic automation, AI, escalation or human judgement belongs. The controls, evidence, architecture and ownership are then designed around that operating boundary.
Founder And Technical Leadership
LevelUp360 is led by Manuel Tomas Estarlich, a Principal Architect with more than 20 years across infrastructure, cloud, enterprise architecture and AI delivery. That includes seven years designing secure cloud platforms in regulated financial services under direct security and audit scrutiny, together with public-sector transformation and large-scale Azure platform delivery.
The process-design and automation experience behind LevelUp360 predates current AI. Manuel began building Delphi business applications that automated operational processes in 2002. He later designed SharePoint workflow solutions, Dynamics 365 case-management systems and Power Automate processes. He also automated migration and technology-delivery workflows through PowerShell, DevOps pipelines and infrastructure as code.
These systems automated different types of work, but the discipline was the same: establish the required outcome, understand how the work moves, remove steps that should not exist, and then decide what rules, systems and people should do. Manuel now applies that method to agentic AI workflows, including where a model may interpret context or select a path, when the system must escalate, and where a person retains control.
The problem behind LevelUp360 also comes from working across organisational boundaries. In process-heavy environments, delivery crossed business analysis, design, architecture, security, platform and operational teams. Procedures often existed, but people still struggled to identify everything another team required or to navigate the process successfully.
By learning the requirements and documenting what was missing, Manuel became the person colleagues approached when they needed to move work across those boundaries. When the process was not understood, the consequences were delay, rework and frustration. As AI agents begin to participate in the same workflows, making that operating knowledge explicit becomes more important, not less.
How We Work
We do not start with tools.
We start with the required business outcome and how the work actually happens. We establish what is known, what remains uncertain, who owns each decision and how success will be measured before recommending a technical intervention.
Every engagement follows the same principles:
- Understand the actual workflow, not only the documented procedure.
- Redesign the work before deciding where AI belongs.
- Treat security, governance, failure and operating constraints as design inputs.
- Keep evidence connected to decisions, controls, approvals and change.
- Name the people who own the outcome during delivery and after handover.
- Measure the complete workflow, not only the performance of the model.
What We Bring
- Firsthand experience carrying complex, regulated work across organisational boundaries.
- One connected path from business decision and workflow redesign through controls, architecture and production.
- Practical governance that exists in workflows, systems, evidence and change processes, not only in policy documents.
- Direct senior involvement and bounded, outcome-led engagements.
- Decision and implementation artefacts that business, security, audit and delivery teams can use.
What We Do't Do
- We do not assume AI is the answer when workflow redesign, deterministic automation or integration would solve the problem better.
- We do not treat a compliance checklist or policy document as an implemented control.
- We do not force a platform into a workflow before the problem, evidence and operating constraints are clear.
- We do not replace legal, compliance, risk or data-protection functions; we give them a concrete workflow, control and evidence record they can review.