Most companies are deploying AI without the governance or testing controls to manage it safely. A recent Grant Thornton survey found nearly 8 in 10 executives believe their company would fail an AI governance audit today. This article explains why that gap exists, why it matters, and what software testers can do about it.
Companies are spending on AI, but not controlling it
There is no doubt that companies are investing in AI. Leaders see it as something they need if they want to stay competitive. According to the survey, 75% of boards have already approved major AI investments. But at the same time, many of those companies have not set clear rules for how AI should be used. Almost half have not defined governance expectations, and almost half do not have strong programs to manage AI risk.
Companies are pushing AI into their systems, but they are not building the structure to control it. A lot of this comes from pressure. Leaders feel like they need to move quickly or they will fall behind. But moving fast without putting guardrails in place can lead to bigger problems later. See why certified software testers are needed more than ever as AI writes more code.
AI is starting to act on its own
New types of AI, often called agentic AI, can take action without someone telling them every step. These systems can complete tasks, make decisions, and interact with other systems on their own. That can be helpful, but it also increases risk. If something goes wrong, it is not just one bad answer – it can turn into a chain of actions. A system might use bad data, make a wrong decision, and then trigger other systems based on that decision.
For example, an AI system could:
- Act on incomplete or incorrect data
- Produce results that break company rules
- Start automated processes that should have been reviewed by a person
- Repeat the same mistake across different parts of the business
Without strong testing, companies might not catch these problems early. Read more about whether AI will replace software testers and why software testers now need broad skill sets to manage these risks.
The companies doing it right are getting better results
The survey also showed something important. Companies that have fully integrated AI and use it in a more structured way are seeing better outcomes. About 58% of fully integrated AI companies reported revenue growth, compared to just 15% of companies still running AI pilot programs. This shows that AI can deliver real value. But that value comes from using it the right way, not just using it quickly. Companies that take the time to build testing and control into their AI systems are the ones that benefit the most.
Governance depends on testing
A lot of people talk about AI governance like it is just about rules and policies. But those rules do not mean much if they are not backed by real testing. Companies need to be able to answer basic questions:
- Does this AI system give reliable results?
- What happens in unusual situations?
- Can we explain how it made a decision?
- What happens if the data changes?
These are not just policy questions. They are testing questions. If a company cannot answer them, then it does not really have control over its AI systems.
AI testing is different from regular testing
Testing AI is not the same as testing normal software. With traditional systems, you expect the same input to give the same output every time. AI does not always work like that. AI systems can give slightly different answers each time, change based on new data, depend heavily on how good the data is, and be hard to fully explain.
Because of this, testers need different skills. They need to know how to test systems that are not fully predictable. They also need to look for problems like bias, unfair results, or unexpected behavior. This is one reason why many companies are struggling. They are using AI, but they do not have enough people who know how to test it properly. See why ISTQB AI Testing certification matters for your career and how to learn AI testing to future-proof your QA career.
How ASTQB AI Assurance Pro helps
The AI Assurance Pro designation from ASTQB was created to help solve this problem. It focuses on two main things: helping people learn how to use AI as part of testing, and teaching how to test AI systems themselves.
To earn the designation, you need to complete a set of ISTQB certifications that cover both basic testing and AI-specific topics like testing with generative AI and AI system behavior. This matters because it shows that someone has real skills, not just general knowledge. They understand how AI works and how to check that it is working correctly. Read about the most popular ISTQB certifications that form the foundation of this designation.
Why employers should care
Right now, most companies do not have enough people who can test AI systems well. That is a big reason why so many executives say they would fail an audit. Hiring or training people with AI testing skills can help fix that. The AI Assurance Pro designation gives employers a way to find people who understand both testing and AI.
These professionals can help:
- Build better testing strategies
- Reduce risk in AI systems
- Improve reliability
- Help the company get closer to being audit-ready
As AI becomes more important, these skills are going to be in higher demand. See how to get software testing jobs and why software testers need ISTQB AI testing certification to stay competitive.
Waiting will make things harder
Some companies may think they can deal with governance later. That is risky. As AI becomes more common, there will be more pressure from regulators, customers, and leadership to prove that systems are working correctly. If a company waits until something goes wrong, it will be much harder to fix. It is usually easier and cheaper to build strong testing and controls early instead of reacting later.
The role of testers is changing
Testers are no longer just checking if software works. They are helping companies make sure AI systems are reliable and safe. That is a bigger responsibility. Testers who learn AI skills will have more opportunities and be involved in more important work. They will help shape how AI is used in their companies.
AI is not slowing down. Companies will keep using it more. The real difference will come down to who can control it. Right now, most companies are not there yet, which is why testing, governance, and structured programs like AI Assurance Pro are becoming so important. For broader guidance, explore our software testing career tips.
Ready to build your AI testing credentials? Start with how to get ISTQB software testing certification, see where to get ISTQB AI certified, or learn about the ASTQB AI Assurance Pro designation directly.