The Culture Signals That Keep AI from Scaling

Many organisations have started experimenting with AI. They have introduced new tools, trained employees, and launched pilot projects. Yet relatively few have succeeded in embedding AI across everyday workflows and turning it into measurable business value.

According to McKinsey’s 2025 State of AI report, more than 80% of organisations using generative AI have yet to see a material contribution to earnings.

AI adoption also depends on the signals employees receive from the workplace culture. 

Leaders may say, “We encourage everyone to experiment with AI.” However, employees pay closer attention to what happens when someone tries, makes a mistake, raises a concern, or finds a faster way to work.

When these signals are present, limited adoption should not immediately be labelled as resistance to change. Employees may be responding rationally to unclear expectations, personal risk, or a lack of trust.

To change these signals, leaders should focus on three priorities: modelling responsible AI use, creating clear and safe conditions for experimentation, and aligning AI-enabled workflows with employee involvement. This means demonstrating how AI outputs are reviewed, treating responsible mistakes as opportunities for learning, and communicating honestly about how roles may evolve. Employees should be involved in redesigning the workflows they understand best, while clear guidance should define which tools are approved, what data may be used, when human review is required, and who remains accountable for the final outcome.

AI will not scale through access and training alone. It scales when leaders create the clarity, trust, and working conditions employees need to use it responsibly.

At ALVIGOR, we help leaders uncover the workplace signals limiting AI adoption and translate them into practical action, from creating safer conditions for experimentation to involving employees in AI-enabled workflow redesign.

If your organisation is ready to move beyond isolated pilots and embed AI into the way work gets done, we’d be happy to start the conversation.