Seppo Kuula
CEO at Cloud1, Doctor of Technology and long-time digital influencer.
The adoption of AI is not a mere technical exercise but a comprehensive business transformation. While the majority of Cloud1's efforts focus on tackling technological challenges, the impact of technical work multiplies when issues are addressed at the business level as well.
In this article, we introduce five of our advisory services designed to support our clients in navigating the AI revolution. The need for these services is evident: when Western European business leaders were asked what factors prevented their organizations from adopting AI, the answers were clear. (Data source: IDC Research Report 2024, Business Opportunity of AI).
The primary concerns and fears businesses have about adopting AI are closely tied to data and AI governance: data management practices (40%), compliance (38%), data security (38%), and privacy (37%) rank as the top four concerns, according to the same IDC report.
To move from pilots to truly impactful production solutions, the fundamentals must be in place. Our advisory team can assist in many ways, and here are a few concrete examples.
1. Data Management
Our data governance work ensures that our clients' data is structured in a way that stands the test of time. Data products, clear ownership, documented responsibilities, and defined usage practices are just some examples of the outcomes we deliver. The majority of the work involves clarifying business concepts, documenting processes, and effectively communicating these aspects.
The scope of this work varies. At its simplest, it involves clarifying and documenting responsibilities and processes for data management. At its most complex, it might include a multi-year process of defining ownership of data warehouse tables, rows, and columns, as organizations progress toward becoming data-driven and leveraging AI.
2. Laying the AI Foundation
Executing an individual pilot project can be straightforward, but large-scale AI production requires certain fundamentals. From a technical perspective, establishing an AI platform is typically essential—this involves creating a shared set of technical components and rules for all projects. From an operational perspective, we recommend organizing an AI center of excellence, initially with part-time resources.
A technically reliable AI platform and the expertise provided by a center of excellence allow project teams to focus on specific business challenges without having to address every technical question. Shared services provided by the platform include monitoring solutions, standardized model updates, testing practices, and compliance management.
While setting up an AI platform requires technical components from Azure, much of the work is consultative—training personnel, developing processes, and supporting organizational changes. After setup, clients can maintain their AI platform independently if they have the right resources in place—or they can continue to rely on our support.
3. Prioritizing Applicaiton Ideas
The generative AI wave generates ideas. In many client organizations, dozens or even hundreds of potential AI use cases have been identified. However, investing in all of them is unrealistic. A preliminary evaluation requires both technical and business thresholds: Which solutions are feasible, and where are the risks too high? Which cases offer enough benefit to justify a short payback period?
We assist clients in filtering and prioritizing AI ideas. Based on our experience, roughly half of the ideas might be worth pursuing as pilot projects, but only 10–20% can be realistically scaled into production. Prioritization ensures that resources are directed toward projects where AI offers maximum value, while deferring projects that lack mature technology or adequate data.
In long-term collaborations, this prioritization process can even be fully outsourced. Generally, we complement the client’s AI team, working together to continually refine an understanding of opportunities and productivity.
4. Service Design for AI Solutions
Most AI solutions—at least in their early stages—interact with humans. At a minimum, systems are built following the "human-in-the-loop" design principle, where AI actions are subject to human approval.
A productive AI solution requires a high-quality user experience. Processes must make human-AI interaction intuitive, or at least structure the solution to ensure that AI gathers the necessary data in a clear and effective manner.
We provide service and user interface design to ensure AI solutions deliver their intended value and operate as safely and responsibly as conditions require. This work is typically carried out as part of individual AI projects.
5. AI Solution Monitoring
AI solutions in production require monitoring—arguably more so than traditional IT systems—both due to the rapid pace of technological advancements and the inherent unpredictability of AI. Unlike traditional systems, AI solutions can perform differently under varying conditions, necessitating statistical analysis for automated monitoring.
Monitoring is crucial both operationally and strategically. How do we ensure that the solution operates correctly most of the time? How do we detect declines in quality or accuracy? What about identifying misuse attempts?
Monitoring also plays a role in lifecycle management. What risks arise when upgrading a language model to a newer version? How do we ensure that our customer service chatbot adapts to evolving terminology and formats? How do we measure whether a solution delivers the business benefits it was expected to?
We support our clients’ AI centers and organizations throughout the lifecycle of their AI solutions—even after implementation, ensuring they remain effective and autonomous.
Success is Built on the Combination of Technology and Business
According to stark statistics, around 80% of AI pilots fail. Unrealistic expectations, insufficient preparation, poor data quality, and poorly designed user experiences are the most common reasons—rarely is the issue with AI technology itself.
We cannot promise that every AI idea can be transformed into a functional production solution. However, we can promise to approach our clients’ AI journeys as a whole, bringing expertise to every phase of the process—from technology and workflows to user interfaces and profitability analysis.