Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Shadow IT is any technology used for work without the knowledge or approval of the people accountable for it: a free SaaS trial, a personal cloud account, an unvetted AI assistant.

What Is Shadow IT?
Shadow IT is technology used for work that the people accountable for security, data protection, and continuity do not know about. It is a free project-tracking trial signed up with a work email, a personal cloud drive holding client documents, a script running on somebody’s laptop that the finance close depends on, or an AI assistant handling customer correspondence. The defining feature is not that it is forbidden — most of it has never been considered at all.
Treating shadow IT as a discipline problem guarantees you will keep finding it. Staff adopt unsanctioned tools because the sanctioned path is slower than the deadline. A marketing manager needs a design tool this afternoon; procurement takes six weeks. An engineer needs a database for a prototype; the internal request form asks questions the prototype cannot yet answer. Self-service SaaS, free tiers, and a card payment have removed every barrier that once made this difficult.
That reframing matters, because the fix is partly operational. If approval for a low-risk tool takes two days instead of two months, most shadow IT stops appearing.
The current wave is generative AI. Staff paste contracts, code, support tickets, and customer records into public assistants to summarise or rewrite them, which is a disclosure to a third party however the provider’s retention policy reads. Samsung restricted internal use in May 2023 after engineers submitted source code to a public service, and the same pattern has repeated across professional services and healthcare since.
The response that works is the same as for SaaS generally: provide an approved tool with an enterprise agreement and data handling terms, publish a short policy stating what may and may not be submitted, and monitor for use of the alternatives. Banning AI outright simply moves the activity to personal devices, where you have no visibility at all.
No single source gives a full picture, so combine several.
Sort what you find into three piles. Sanction the tools that are genuinely useful and acceptably safe: move them to single sign-on, put a contract in place, and add them to your SaaS posture monitoring. Replace the ones that duplicate something you already licence. Remove the small number that are unacceptable, and explain why rather than simply blocking the domain.
Then close the loop. Publish a short list of approved tools by category, give people a fast route to request something new, and review discovery output monthly rather than annually. Cyber Essentials requires an accurate inventory of the hardware, software, and cloud services in scope, so for certified organisations this is a scheme obligation as much as good practice, and the NCSC treats asset management as a foundation for everything else.
If you would like help building that picture, our cyber risk analysis services include discovery of unsanctioned services alongside the rest of your estate.
No. Much of it is an ordinary tool used sensibly by someone who needed it quickly, and some of it is better than what IT would have chosen. The risk depends on what data it holds, whether authentication is covered by your controls, and whether the business now depends on it. Assess each case rather than treating discovery as a disciplinary exercise.
Shadow AI is a subset with a distinctive risk profile. Unsanctioned SaaS mainly creates storage of your data somewhere you do not control. Unsanctioned AI additionally involves submitting information for processing, sometimes for model training, with limited ability to retrieve or delete it afterwards. It also spreads faster, because using a public assistant requires no account creation or procurement at all.
The organisation. Under UK GDPR the controller is accountable for personal data processed on its behalf, whether or not the arrangement was approved internally, and the ICO will not accept that an employee acted without authorisation as a defence. That is precisely why discovery matters: you cannot apply safeguards, contracts, or breach procedures to systems you do not know exist.
Block the genuinely unacceptable and sanction the rest. Broad blocking pushes work onto personal devices and home networks, where you have no visibility and no recourse, which is worse than the original problem. A short approved-tools list, a request process measured in days, and a clear explanation whenever something is refused produce far better compliance than a deny list.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.