Cyprus businesses adopt AI faster than they prepare for it

Cypriot businesses should organise their data, establish clear rules and train employees before investing in artificial intelligence (AI), according to Kyriaki Parmakki, co-founder and chief executive of Limassol-based solutions studio essere.ai.

Speaking after hosting the first ClickUp event in Cyprus on October 8, Parmakki said companies needed to establish strong operational foundations before introducing AI tools if they wanted to achieve measurable results.

The event brought together business owners, directors and technology specialists to explore practical applications of AI, including live demonstrations and discussions about how companies could integrate the technology into their daily operations.

ClickUp, a US software company known for its project management and AI tools, held the event in partnership with essere.ai, its official and certified partner in Cyprus.

AI use among individuals outpaces businesses

Parmakki highlighted a gap between the willingness of individuals to use AI and the level of adoption among companies.

According to figures she cited, 44% of people in Cyprus used generative AI in 2025, compared with an EU average of 33%. However, only 9% of Cypriot companies used AI, against an EU average of 20%.

“People in Cyprus are ahead of their companies,” she said.

Parmakki said the strong attendance at the event and the practical questions raised by participants showed that businesses were increasingly interested in how to implement AI rather than whether to adopt it.

The programme included a presentation by ClickUp guest speaker Mary Bornaghi, who discussed the integration of work management and AI. Parmakki outlined the foundations businesses need before adopting the technology, while essere.ai senior AI engineer Neophytos Polydorou demonstrated five use cases using real systems and demonstration data.

Representatives from two companies also discussed how moving their work into ClickUp had changed their operations.

Why businesses need to organise their data first

Parmakki argued that the effectiveness of AI depends on the quality, accessibility and organisation of the information available to it.

“Everything a company does is data. A call, an invoice, a message, an approval. AI can only work with what it can read, so the first question is never which tool, it is where the data lives,” she said.

She outlined four stages of business development based on how information is managed: data held in employees’ heads, information stored in separate files, data organised within systems, and operations supported by AI.

Businesses that skip the groundwork risk receiving generic answers or incorrect information from AI systems. Employees may also resort to personal AI accounts to handle company information when suitable workplace tools are unavailable.

The challenge, she said, is turning individual productivity improvements into measurable results for the wider business.

Parmakki cited McKinsey research involving 1,700 executives, saying 80% reported that AI made them personally more productive, while only 37% of companies saw those benefits reflected in their results.

“The gap between those two numbers is the foundations,” she said.

AI tools can automate repetitive tasks

The event highlighted opportunities to use AI in processes involving large volumes of unstructured information, such as telephone calls, emails, invoices and documents.

Potential applications include answering calls in the caller’s language, recording information and making bookings, as well as processing supplier invoices without requiring employees to open each email individually.

AI systems could also prepare morning briefings that flag tasks at risk of missing deadlines and suggest how work might be reassigned.

Parmakki said these applications could remove repetitive tasks that occur every week, allowing employees to check the results quickly.

Live demonstrations included an AI agent called Ciao handling a support call and making a booking, another system identifying a task at risk and proposing its reassignment, and an agent preparing month-end billing for human approval.

Other demonstrations showed an AI agent filing a supplier invoice and carrying out a weekly security review to check whether its actions complied with established rules.

Clear rules and human oversight are essential

Parmakki warned that businesses must establish clear boundaries before allowing AI agents to operate independently.

“Before an agent goes live we write three things down, what it may read, what it may do on its own, and where a person decides,” she said.

She outlined a cautious approach in which AI could draft messages but not send them without permission, or suggest payments without authorising them.

Decisions involving people or payments should remain subject to human approval, she said. Companies should also log every action and review the records weekly to ensure systems remain within their permitted limits.

The event also examined the risks of employees using AI tools without company oversight. Figures presented during the event indicated that the proportion of employees regularly using AI on their work computers had risen from 15% to 45% in one year, while two-thirds of those users signed in through personal rather than company accounts.

“The question is not whether your people use AI. It is whether the company knows,” the presentation stated.

Parmakki also discussed the choice between developing bespoke software and configuring existing platforms. While building software has become more accessible, she said, businesses must still manage security, permissions, backups, updates and ongoing support. Losing the person who originally built a system can also create problems.

Her recommendation was to develop bespoke technology where it creates a genuine competitive advantage and use existing tools for other needs.

EU AI rules add to businesses’ responsibilities

Regulatory compliance was another focus, particularly the European Union’s AI Act and its implications for companies introducing the technology.

Parmakki said businesses needed to train employees in AI and inform customers when they were interacting with a machine.

The presentation also stressed the importance of documenting training and identifying systems that require greater oversight, particularly those involved in decisions affecting people or money.

Parmakki said a lack of expertise, uncertainty over legal requirements and concerns about data protection were among the main barriers to adoption.

She referred to Eurostat findings indicating that a lack of relevant expertise was the most frequently cited reason among businesses in Cyprus and across the EU that had considered AI but decided not to adopt it. Legal uncertainty and data protection concerns also featured, while cost ranked considerably lower.

Three priorities for companies planning to adopt AI

Parmakki outlined three priorities for businesses seeking to move beyond experimentation.

First, companies should establish a reliable operational system that they own, creating a single source of information for clients, tasks, documents and figures instead of relying on employees’ memories, inboxes and spreadsheets.

Second, they should introduce written rules defining what AI can access, which actions it can perform independently and when human approval is required. Every action should be logged, with a designated employee responsible for oversight.

Third, businesses should train their employees, encourage the use of company accounts instead of personal ones, and select one recurring process in which AI can be introduced and tested.

Parmakki recommended running the chosen process for four weeks and assessing whether it delivered the intended result before expanding to another task.

“That is the difference between experimenting and integrating,” she said.

She added that companies did not necessarily need a lengthy strategy document or bespoke software to begin. Instead, they needed sound preparation and a disciplined approach to implementation.

The event also cited Gartner’s expectation that 60% of AI projects lacking AI-ready data would be abandoned, underlining the importance of preparing business information before introducing new tools.

The central message was that AI adoption is an operational change, not simply a technology purchase. Its benefits depend on how well businesses organise their information, protect company data and prepare employees to use the technology effectively.

Also read: Tech sector contributes 15% of GDP, president says
For more videos and updates, check out our YouTube channel