What’s next in AI? 10 predictions for 2026

At QA, we have the privilege of seeing technology and skills trends as they emerge. Coupled with our internal expertise, it’s a bit like having a crystal ball. So, as we enter 2026, I’m peering into the future and sharing a few pivotal shifts to expect surrounding AI before this year is out – but first, some context to how we got here:

How we got here - 2025 AI review

Over the course of 2025, businesses and individuals alike made the step from preparing for AI change to actively living alongside it. 79% of organizations were using generative AI as of November, up from 65% the year prior. But we have a long way left to go, since only 38% have moved beyond pilots into real adoption.
Last year, we moved from awareness to action. We shifted from talking about AI’s potential to rolling out all-staff training in prompting proficiency, tailored to roles and industries, because organizations are ready to start really applying new, powerful tools to their use cases.
I’ve said before that trust is the issue of our age. And it remains so. One much-debated question in the public mind that I tackled last year was whether AI could replace therapists. My take on this was a resounding ‘no’. While there may be some useful applications in triaging, for instance, human expertise matters. Especially in high-risk situations. Sadly, that message hasn’t always landed, and there have been sobering stories of people turning to AI for help instead of professionals. Regulation in this field is urgently needed to catch up with the tech and how it’s already being used.
There’s another side to that coin. On one face, premature application without safeguards; on the other, AI reluctance and hesitation. There’s a tension between the fear of moving too fast and the risk of falling behind. Fueling and complicating all of this is the growing skills divide.
It’s a new year, but all these questions follow us into 2026, and must be tackled. Attitudes are shifting again. AI is becoming the interface for everything, citizen developers are on the rise and democratized tech skills are a reality.

So, what’s next? Here are my 10 predictions for AI in 2026:

1. Everyone becomes an end user

Generative AI tools won’t just be for business staff. Technical teams - yes, even those building AI applications - will need prompting proficiency of their own. AI literacy will become as fundamental as Excel skills once were.

2. AI inside your favorite tools

Expect deeper integration of AI features within existing software. Microsoft Copilot will lead the charge, with Copilot 365 gaining smarter connections and richer context. We’ll end up with AI that feels less like an add-on and more like part and parcel of your familiar operating systems.

3. Apps inside AI

We already know from OpenAI’s Dev Day that ChatGPT will start hosting apps from early this year,   and I expect Copilot won’t be far behind. Microsoft’s new AI companion, Mico, is a perfect example of this shift, prioritizing a human-centered approach. In short, AI is soon going to be our go-to tool for everything, with all the other tools embedded inside it.

4. Getting over pilot-itis

The 38% of organizations scaling AI will jump to 70–80% by year-end. Those stuck in an endless testing phase - what the team here at QA call a case of ‘pilot-itis’ - will break through their barriers. That means more AI literacy training, more prompting proficiency, and more technical staff learning to manage AI systems at scale.

5. AI development - Where the action starts

IT and Service Management, Marketing and Sales, Product Development, and Software Engineering will lead the way with further AI experimentation this year. To support it, technical teams should learn these use cases. Or, flip the script and train those teams in AI development themselves – democratizing AI in your business while easing skills shortages! Tools like Ela that offer personalized learning assistance for AI upskilling make this more achievable than you might think.

6. Rise of agentic AI

Expect a surge in agentic AI pilots tailored to organizational needs. Technical teams will be starting to experiment now, and pilot soon, so rollouts will accelerate from late summer. Multi-agent systems could be commonplace in many organizations by 2027.

7. Beware of AI slop

AI slop is not just the images flooding your social media feeds; it can be anything. Citizen developers in 2026 will create more code with AI help. As I’ve said before, the best AI outputs come from human expertise. If these citizen developers don’t have the skills to check results rigorously, errors will creep in. Expertise in code review, data science, and model development will be more important, not less.

8. Physical AI experiments

There’s been a lot of buzz around AI hardware and robotics at the tail end of 2025. Generative agents in robotics will grow, but how it will scale is simply too tricky to predict. Costs, safety regulations, and the lack of a standard OS could all slow progress. This is a big shift comparable from the swap from brick phones to smartphones, or robot vacuum cleaners to humanoid robotics. Right now, what’s important to think about it how we address the demand for technical staff at the experimentation stage. We will need more people who are able to build or extend ‘World models’ - to create a digital version of the physical world that the AI can be trained on, and who understand how to build multi-agentic systems and apply rigorous safety and governance processes to AI. 
Is all this really possible now, in 2026, you’re asking? Well, probably. The real question is whether people, businesses and governments will agree it’s ready to use - and I think not.

9. Jobs will change - not vanish

A defining fear factor in 2025 was job replacement. It’s understandable, a fear that I’m empahetic to. So I will be honest that yes, some roles will phase out, but others will be created. On balance, there will be more new roles that ‘lost’ roles. But for the vast majority of people, it’s not as extreme as a total career change. 2026 is about changing how you work, embedding approved AI tools as part of your daily toolkit and embracing them to augment your performance. This actually strengthens you in your career; with AI skills you’re more likely to be promoted, not replaced! 

Futureproof your skills with our top AI skills for 2026. 

10. Governance gets serious about AI

AI governance has been a hot topic last year, and while attitudes do vary across the globe, all businesses can likely agree that we need reliability – which looks like chatbots that give correct information to customers – and AI that works for your business and not against it – which doesn’t look like recommending your competitors, for example.
For this reason, governance will move from theory to practice. Organizations will build structures to prevent those potentially costly missteps. We all remember the cautionary tale from Air Canada a couple of years back; their chatbot gave outdated, incorrect information to a customer regarding bereavement discount, resulting in the ruling that Air Canada must pay damages. 
To ensure compliance with new regulations, both leadership and security teams will need more detailed AI training, but all staff will need mandatory literacy programs to ensure they’re working within policy.

What are the overarching themes of these predictions?

Well, AI is going to be a more operational part of our everyday work, and success depends on skills, governance, and a healthy dose of human judgment. AI is, in a way, inevitable – the part I am interested in is people, and how we steer it forward.

Talk to our experts about your AI strategy for 2026