The New Digital Literacy: Learning to Prompt
When Cisco Germany began exploring AI, its workforce displayed wide-ranging levels of understanding and engagement. Detlev Kühne, managing director of partner sales, explained that employees’ familiarity with AI varied significantly across the company.
Most rated themselves on the lower end of an internal scale assessing proficiency with key AI tools. Kühne noted that many realized their knowledge gaps only after discussing what AI truly meant in practice.
To bridge these gaps, Cisco introduced foundational training for all employees. The sessions went beyond ChatGPT, highlighting multiple platforms while teaching the art of prompting effectively. Kühne compared prompting to modern-day Googling, stating that the precision of prompts determines the quality of responses.
The training also outlined the legal and ethical considerations of AI, emphasizing GDPR compliance and the EU’s AI Act. Employees were reminded to handle input and output data responsibly to prevent misuse or privacy violations.
Interestingly, older team members, often referred to as “silverbacks,” proved highly engaged during sessions. Their active participation debunked the myth that only younger workers grasp AI concepts. Many younger colleagues were already using AI in daily routines such as vacation planning, summarizing client meetings, or drafting emails.
Circuit: Cisco’s Private AI Powerhouse
Cisco’s internal AI adoption surged in 2024 with the debut of its proprietary tool, Circuit. The platform evolved from earlier projects like Bridge IT and Enterprise Chat AI, expanding the company’s digital capabilities.
Circuit’s launch dramatically boosted AI activity across the organization. Out of more than 80,000 employees, around 50,000 began using it regularly.
Kühne described Circuit as more than just software—it operates as Cisco’s private large language model integrated directly into Webex. Employees can also access it through a web browser for flexibility.
By hosting the platform internally, Cisco ensures that all data remains protected. Kühne said this security layer gives staff confidence to use sensitive information responsibly. Because the AI operates within Cisco’s ecosystem, nothing leaves the company’s servers.
The Hidden Threat of Shadow AI
Despite strong progress, Cisco still battles “shadow AI,” where employees turn to public tools like ChatGPT for convenience. This creates risks of data leaks and regulatory breaches.
To counter this, the company promotes continuous learning through workshops and certifications on safe AI usage. In parallel, its security platform, AI Defense, monitors activity to detect potential misuse.
Kühne highlighted that this system is particularly vital in development teams using coding tools such as GitHub Copilot. He urged businesses to adopt a three-pronged model built on trust, awareness, and strict data protection.
Quality Control and Hallucination Risks
Even with internal solutions, ensuring output accuracy remains challenging. Cisco discovered that verifying AI responses is crucial, as even internal systems can make mistakes.
According to Kühne, some external tools occasionally referenced outdated products instead of current ones. He stressed that no AI system should be treated as infallible. Continuous cross-checking ensures reliability and maintains client trust.
Circuit’s advantage lies in its use of proprietary company data, which significantly reduces—but does not eliminate—the chance of false results.
Small Steps Toward Big Impact
Cisco’s AI framework follows a “Start Small, Then Go Big” philosophy. The goal is to integrate AI strategically where it can offer immediate, measurable value. Kühne said that the company’s primary focus areas include sales and customer experience (CX).
In sales, AI assists employees in understanding client industries, composing messages, and compiling updates before meetings. It even converts customer information into podcasts so employees can prepare during commutes.
Technical teams rely on AI for verifying specifications before delivering client responses. In customer service, AI already handles about 25% of cases, easing workloads and shortening response times. However, human agents remain available for complex concerns.
Key Lessons from Cisco’s AI Journey
Reflecting on Cisco’s AI experience, Kühne outlined five essential lessons for other leaders.
First, companies must identify early adopters who can champion AI within teams. These “fours and fives” act as catalysts for broader adoption. Second, leadership commitment is non-negotiable; managers must model consistent use and share both successes and failures openly.
Third, regular communication is vital. Training should not be a one-time event but an ongoing conversation through workshops and informal gatherings. Fourth, businesses need to assess their employees’ current AI engagement before formal programs begin. Many leaders are unaware of how frequently AI tools are already in use.
Finally, AI should be treated as a central business priority, supported equally by IT and management. Kühne admitted that he began the initiative too late and advised others to start early. Fear of AI, he said, comes from misunderstanding—and that can only be addressed through education and example.
He concluded that AI must become part of company culture. Unlike past tech waves that faded, AI will remain integral to how organizations function and compete.
