Every year, Gartner listens intently to CIOs and other technology leaders about where technology is headed. Gartner’s IT Symposium/Xpo is divided into two parts: a discussion of current trends and predictions for the future.
Gartner Fellow Daryl Plummer introduced the topic for this year’s discussion. He added that we live in a “world where the old normal is being disrupted,” adding that “nothing can be expected to be the same as last year or the year before.”
These are Mr. Plummer’s predictions and my own observations.
1. AI Proficiency Testing in Hiring by 2027
Gartner predicts that by 2027, 75% of hiring processes will include certifications and tests for workplace AI skills.
Employers will want to hire people who are good at using AI. So, individuals should “be playing around with AI all the time, including prompting.” While AI can improve our productivity and creativity, it only works if we ask the right questions in the right way.
However, it’s also important to understand your own business processes, because that’s how you can make them more effective when using AI.
- “You don’t have to worry about being unemployed. You only have to worry about losing your job to someone who can use AI better than you.”
That last statement rings true for most people, in my opinion. However, I think the percentage of people who will need certification is still too high.
2. AI-free aptitude tests will be needed by 2026
We’ve all noticed that our children, as we’ve become more dependent on technology, don’t remember things as well as we did when they were their age.
But we’re all experiencing the same thing. “When was the last time you read a map?” Or when did you write calligraphy? Or when did you drive a manual car?
We’ll have to decide what to let go of and what to keep. People with specialized skills will become increasingly scarce and valuable.
I agree with this. Also, I think that “AI-free” tests make sense for most organizations for some roles, but not for most.
3. Countries Could Be Locked in by Local AI Systems by 2027
By 2027, 35% of countries will be locked into local AI platforms that use their own proprietary contextual data.
This is called AI sovereignty. The goal is to force organizations to use local solutions through regulations and other means. China, the US, and the EU are the three main centers of this power, while other countries are trying to maintain their independence.
For example, if China has manufacturing data that no one else owns, it will ensure that its AI systems have the best data for manufacturing.
Contextual data will be a powerful incentive for people to be locked into or excluded from specific AI systems. Governments and “digital state platforms” are working to solve this problem.
In addition to using Model Distillation as a strategy to reduce the impact of local control, we are also examining open-source systems that are available to everyone. This is also likely to happen in practice.
4. Multi-Agent AI Will Lead the Market by 2028
By 2028, organizations will use Multi-Agent AI systems in 80% of customer-facing business processes.
There was a lot of talk about Multi-Agent Systems at the conference. These systems, which can understand the context, will help to attract customers. Customers do not like boring service tasks. Improving them will benefit everyone.
Even if the customer does not know it, most of the work will be transferred to AI agents. Although these systems will reduce labor, they will increase customer satisfaction and retention.
However, when I spoke to the conference attendees, I found that most of them are not ready for Multi-Agent Systems. But there are many organizations that are starting to implement or are testing AI agents for customer service.
5. 90% of Business-to-Business (B2B) Buying Will Be Done by AI Agents by 2028
By 2028, 90% of business-to-business (B2B) buying will be done by AI agents, and over $15 trillion in B2B spending will flow through AI agents.
It’s not hard to imagine how AI agents will handle the buying of business products, as they can communicate more effectively with other agents. This will have many implications, such as AI Agent Optimization replacing traditional search engine optimization (SEO), and the emergence of new AI-based Machine-to-Machine (M2M) products and services. So we should design business processes for AI agents, not people.
I think this prediction depends on how we count purchases. I expect AI agents—even if they were previously called robotic process automation (RPA)—to be involved in many purchasing decisions.
6. AI could lead to over 2,000 lawsuits by the end of 2026
I wish AI could act like a psychiatrist, but it’s already happening. We all need to be careful because there are not enough safeguards in place.
“Black box agentic AI” has the potential to go astray. So we need to strike a balance between data and security, including appropriate rules to prevent bad behavior. One way to do this is to deploy “guardian agents” that monitor other AI agents.
In addition to good data quality, explainable AI, and ethical use, it’s also important to have good quality assurance programs for AI agents.
7. 22% of payments by 2030 will be programmed for AI
This means that AI capabilities are embedded directly into money. So if you give someone money and they don’t follow the rules, the money will automatically come back to you.
In this case, the value of the exchange will change based on how the business operates, and money will become more like video game items than cash.
I’m a skeptic on this one. I’ve been hearing about smart contracts for years—just look at Gartner’s 2020 predictions. But they’re not yet mainstream. They’ll grow, but I don’t think they’ll be that popular by 2030.
8. The cost-value gap in service contracts will shrink by 50% by 2027
The idea is that business-to-business services are based on labor arbitrage—the difference between the cost of hiring people to work on a project and the value the organization receives.
AI agents will change all of this, allowing service companies to lower prices and keep a larger share of the profits.
AI agents will also be able to uncover tacit knowledge within organizations, which will create new value. Gartner expects to see pricing based on continuous innovation, not just labor costs.
Organizations that hire service companies should be careful to protect their internal knowledge with measures similar to non-disclosure agreements (NDAs).
While this is possible, I don’t think it will happen as quickly as Gartner predicts. Organizations will need more time to adapt.
9. AI regulations will be widespread by 2027, driving $5 billion in investment in compliance
As part of the push toward AI sovereignty, jurisdictions are considering AI regulations. Over 1,000 AI laws were proposed last year, and enforcing them is proving to be incredibly difficult.
In particular, he says, no two governing bodies have consistently defined AI. AI regulation can be both a help and a hindrance.
Governments will start to tax the use of AI. Because they need to limit how much AI is used and what people can do with it.
Ultimately, businesses will find ways to make money through the government mechanisms that are meant to protect us.
To understand this and stay safe, AI literacy is essential, and we need to create a “mind map” of potential laws and regulations.
Regulation is almost certain to increase. Companies will have to invest in compliance. But I predict that it will take longer to tax AI specifically.
Also Read: How to Use AI Chatbots Effectively?
10. GenAI to shake up $58 billion in productivity tools market by 2027
GenAI and AI agent adoption will be the biggest business tool challenge in 30 years. Microsoft is the company that will make this change, but they will have to “do more to deliver the GenAI experience in the performance space.”
“We need to be the ones who set the new dominant design that everyone wants to follow”—that will be with new UIs, plug-ins, new document types and formats. If new vendors are coming to the market, it’s because the right ones haven’t done it yet.
Should an AI agent be able to create PowerPoint slides in a smarter way? Do I want Microsoft to make Copilot in the Office (M365) suite better? Yes, I do.
I agree that others will try to reinvent basic business apps to better use AI. If Microsoft doesn’t, there will be market volatility. Copilot still has a long way to go, but I hope Microsoft continues to improve.
