Shadow AI is not the problem
It’s a signal that some organisations are running the wrong AI pilots.

Most organisations talk about shadow AI as a risk, which would make sense, right? I mean, the thought of people using unapproved tools, for unknown purposes, sharing potentially sensitive data with LLMs, using vibe coding tools to hack tools and products they can’t adapt, launching chatbots that may or may not advise customers appropriately, asking Claude or Chat GPT to accelerate huge swathes of their day jobs, writing agents, even to send emails. The list goes on, the panic increases.
Shadow AI, that is AI that employees just find and use themselves; undirected, unapproved and ungoverned, might sound like a nightmare, but could be hugely valuable.
It’s valuable as a signal, because it shows you where people genuinely find value. Official AI strategies are often instigated by leadership, but shadow AI starts on the factory floor, and reveals what people are actually trying to do faster, better, or differently.
Shadow AI is not a failure of strategy, it’s simply evidence that the workforce is already experimenting with the future.
While corporate AI aims high:
- enterprise AI platforms
- predictive analytics programmes
- automation initiatives
- AI centres of excellence
Employees use AI more simply:
- summarising long reports
- analysing messy spreadsheets
- drafting comms
- interpreting policy
- generating code
I myself use AI in my day to day to summarise complex documents, to structure content, create media, to vibe code new tools, and to make accounts admin easier. My daughter uses AI to summarise topics and learnings, to revise for tests, and to create images and presentations. Before she’s 12, she’s figured out how to use AI for herself, she hasn’t been taught. The point is, AI adoption, in many cases is not strategy led, but need led.
Much of it is small and incremental, but as a whole it accelerates so we can work much more efficiently. Small improvements lead to immediate productivity gains.
If you listen carefully, shadow AI acts as a kind of bottom-up discovery process. It reveals where AI genuinely helps people do their jobs. You might be surprised; the ways you thought people would use AI could just be the tip of the iceberg.
This dynamic is not new. When cloud tools first emerged, organisations saw the same behaviour with shadow IT. Employees started using Dropbox, Google Docs, Slack et al often in violation of IT policy. It was considered risky and uncontrolled. I even led initiatives to shore up shadow IT with Mobile Device Management, and policy. But many of those tools eventually became standard infrastructure.
The biggest risk is not that shadow AI exists. It’s that organisations fail to understand the valuable insight it offers. If leaders assume AI adoption is limited to official pilots and products, they could miss the reality that employees are already integrating AI into their work.
The smartest organisations won’t try to eliminate shadow AI but will try to learn from it. First identify how employees are already using AI tools. Then understand where those uses create productivity gains or new value. Then build safe, supported versions of those capabilities.
The smart AI leaders will stop trying to crack down on shadow AI, and use it to prioritise the AI pilots that actually matter; to really listen.
Katie Lips is CEO and Founder of Rebel Futures. She helps organisations grasp the opportunities at technology inflection points, from the early web and pre-iPhone mobile to cloud-based SaaS and now AI. She has held senior roles at EY and Aimia, and was Head of Digital Policy at Which?. Katie holds a Harvard Kennedy School qualification in Leading in Artificial Intelligence and is a Founder Institute graduate.
Related: explore our work on AI Adoption.