Motivation
At SCRAI, we are committed to help companies and organisations with guidance and specific support to build better, more responsible AI systems. Our support involves each step along the way, covering the full AI lifecycle from inception to operations. Therefore, we are very much interested where organisations stand today in their AI adoption and use, and how they engage in the process of developing, evaluating, integrating, and operating AI.
Survey
The following questions formed the basis of the survey. We think they are highly relevant for business and technology leaders as they are for SCRAI:
What’s the state of AI in companies and organisations? How do organisations and companies utilise AI today?
Do they develop their own AI systems, applications, products, and services?
Do they evaluate and procure third party AI offerings?
Do they integrate AI with their own proprietary data, IT systems, and solutions?
Are they deploying and operating AI? And if so, in which domains – inside the organisation and/or customer-facing in the market?
The survey is a timely, ad hoc reality-check rather than a scientific study.
There were between 61 to 66 participants (depending on the specific question).
In three rounds, different groups participated over the course of the first quarter 2026: members of the HSG AI Alumni Club, participants of the ZHAW CAS KI und Organisation, and attendees of the After the Algorithm festival.
The survey participants came from various industries and organizations of different sizes from SME to multinational enterprises located in Switzerland. It should be noted that participants are from a group of business and technology leaders with an affinity, stake, or interest in AI. This selection of persons may be better informed than other professionals regarding AI initiatives, projects, and investments of their organisations.
Results
Key findings of the survey:
65% of organisations say that they develop their own AI systems/applications. Thereof, 56% have their own internal data science/AI teams, 23% rely on both internal and external development (mixed or dedicated teams), while 21% exclusively use external development partners for development.
84% of organisations evaluate and procure AI systems/applications. Thereof, 61% integrate procured AI systems, modules, or application with their own IT systems, platforms, business processes, and connecting it with their proprietary customer and company data.
83% of organisations have AI systems/application in production/operation. Thereof, 63% only use AI for internal purposes, 25% have AI deployed both for internal and external applications, while 12% use AI only for external applications (customers, partners, suppliers, or recruiting).
Reflection
A surprising result may be the relatively high percentage (65%) of organisations that develop their own AI systems and applications. However, the way the question was asked did not account for differences between highly sophisticated, advanced data science, machine learning, and AI development (e.g. developing and training own AI models) and simpler customised applications based on tools like MS Copilot or OpenAI’s ChatGPT. A more specific questionnaire in this area may put the “own AI development” into perspective as we expect the share of sophisticated internal AI development to be lower.
The 84% of organisations that evaluate and procure AI today will certainly grow close to 100% in the near future. As many commercial software vendors (e.g. SAP, Microsoft, Salesforce, etc.) include functionality and features powered by AI, there is little room for organisations to stay outside of the AI adoption. At the same time, domain-specific AI vendors gain traction (e.g. specific banking applications, supply chain management, etc.). Also, organisations increasingly integrate foundational AI models (for example Large Language Models / LLM from the likes of Anthropic Claude, OpenAI ChatGPT, Google Gemini, Microsoft Copilot, DeepSeek-V3, Alibaba Qwen, Mistral’s various models, or Switzerland’s Apertus) in their own proprietary developments when they shift the focus from personal productivity tools (that provide little sustainable competitive advantage) to sophisticated business applications, processes, and products, and services.
The explanation of the seemingly large 83% of organisations that already have AI in production/operation needs a more differentiated look at the internal vs. external deployment. The main contribution to the 63% of the “internal” deployment of AI originates probably from personal productivity tools like ChatGPT or Copilot. The 37% of “external” AI deployment (consisting of 12% “external” and 25% “both internal & external”) includes not only customer-facing applications, but also AI tools for partners, suppliers, or for recruiting (if external AI touchpoints with applicants are provided). Still, we can see that organisations are more hesitant to deploy AI externally than internally, most likely because they consider customer-facing AI riskier for their business and reputation. From a strategic innovation viewpoint, this approach could be challenged because most successful innovation projects are connected to strategy, solve real problems (high intensity, frequency, impact), focus on customers and business value (not just cut costs).
Outlook
The reflection above showed that later surveys could shine a light into the realities of AI usage in organisations regarding strategic focus and investments:
Where do companies see competitive advantage to build their own AI solutions?
Do smaller or larger organisation profit more from the AI revolution? How do AI investments and usage vary between large and small organisations?
What are the different approaches for companies in different businesses, i.e. B2B vs B2C space? How do industries compare in where they focus their investments and innovations?
Additional surveys could also go into more detail to investigate what functions, processes, markets, customer segments, and use cases are prioritised by different organisations. For the development of AI, the level of sophistication and integration of third-party AI models/systems would be interesting.
For SCRAI, the next crucial level is of course the application of Responsible AI principles, approaches, and methods along the AI lifecycle. Which organisations are already applying RAI practices? What are the gaps they want to close? Where do they struggle and why?
On our mission to improve the state-of-the-art of Responsible AI it is important to understand where research meets practice and how methods and tools are put to most effective use for organisations.
Let us know if you are interested in deeper insights and which questions are most relevant for you.
Ivo Ruckstuhl, Managing Director, SCRAI, June 2026
