In a 2026 McKinsey study, 50% of companies identified AI as one of the key technology investment areas for the next two years. However, nearly a third of organizations are already facing another challenge - integrating AI into existing systems and processes. The biggest challenge becomes the effective use of technology in business processes.
Self-service covers an increasingly larger part of the customer journey
Self-service is increasingly designed not for individual functions, but for a larger portion of the customer service or purchasing process.
According to 2026 McKinsey data, 71% of B2B companies already offer e-commerce capabilities, and in companies that have it, approximately one-third of revenue comes through digital channels. 73% of buyers say they are ready to make purchases online exceeding $50,000.
This also changes the logic of self-service design. What's worth evaluating is not a list of individual functions, but how much of the entire customer journey still depends on employee involvement.
Personalization moves from content to the system itself
We most often associate personalization with recommended products or content, but in business systems its possibilities are much broader.
The system can change what information, priorities, and actions a specific user sees. For a B2B client who regularly purchases the same product group, one scenario is relevant. For an infrequent buyer - another. For a sales manager, the system can primarily show clients whose sales have started to decline, rather than a general summary of all clients.
According to McKinsey data, more than 90% of B2B companies already apply some form of personalization, but the best-performing companies use individual, client-specific personalization approximately four times more often.
This is one of the most interesting directions in software design. Instead of one system for everyone, there emerges a system that increasingly adapts its content and logic according to the user.
Real-time data becomes part of system quality
A modern user interface changes little if it displays yesterday's information, which is why software projects increasingly rely on direct connections with ERP, CRM, warehouse, logistics, accounting, and other systems. The information presented to the client or employee becomes not a periodically synchronized copy, but part of a real-time process.
In McKinsey wholesale trade studies, real-time inventory visibility is already identified as one of the most important digital capabilities. This is especially important in B2B trade, where prices, balances, credit limits, or delivery times can be individual and constantly changing. In such systems, value is increasingly determined not by the number of functions, but by data quality and its availability at the right moment.
Search and navigation logic changes in complex systems
The more functions a system has, the harder it is to fit them all conveniently in navigation. AI allows bypassing part of this structure. The user can simply formulate what they're looking for: show delayed orders, filter products by technical parameters, find a specific client's documents, or compare data from multiple periods.
The shift in the software itself is interesting here. Until now, a person had to know where to look for an answer in the system, but now it's increasingly sufficient to know what answer is needed.
This has particularly great potential in large product catalogs, document systems, customer self-service portals, and internal business systems, where the amount of information has long been a major problem.
Software is increasingly worth designing for other systems too
Until now, most digital products were designed for humans. In the future, an increasingly large portion of information on behalf of the user can be collected and processed by the artificial intelligence systems they use.
According to McKinsey estimates, by 2030, AI systems could mediate $3-5 trillion worth of global consumer trade.
This is relevant not only for e-commerce. In a B2B environment, such systems can collect product data, evaluate technical parameters, check availability, or prepare purchasing decisions.
As a result, APIs, structured product data, and clearly accessible system functions become important not only for integrations between internal company systems.
Modernization increasingly happens in parts
A large portion of companies have systems that don't make sense to replace. An ERP can perfectly manage accounting or warehousing, but its user interface, customer service capabilities, or integrations may be significantly outdated. In such cases, replacing the entire system is often not rational.
Increasingly, a separate layer is created on top of existing infrastructure - a new self-service portal, B2B portal, employee system, product search, or AI function, while the core system remains the source of data and business logic.
This approach allows modernization to be broken down into specific business problems and invest where the impact is greatest. This is much more practical than starting with an abstract goal to "modernize systems."
AI is changing the economics of software development
According to Gartner data, 90% of software engineering leaders are already recording productivity growth using AI, with the average reported increase reaching 19.3%.
The biggest benefit for business is the ability to test the effectiveness of various ideas cheaper and faster.
A prototype, a separate part of the process, or a new user scenario can be created earlier, shown to real users, and only then invest in full implementation. This is especially relevant for larger custom software projects, where a wrong decision at the project's start can cost significantly more than the programming itself.
In this case, AI allows testing more decisions before they become major investments.
What's worth reviewing in your business?
Looking at 2026 software trends, it's worth starting with existing processes and systems.
Where does the customer still depend on an employee, even though the process could be moved to self-service? Where is everyone shown the same information, even though their needs differ? Where does the system have the necessary data, but it's not available in real time? Where is it becoming difficult for the user to navigate due to the growing number of functions and information? Where can an old system be supplemented instead of replaced?
Such questions today often yield more than simply trying to "implement AI."
The software market is moving toward systems that better understand the user's situation, have more context, and cover a larger part of the process. For business, what's most important is to identify those areas where this shift can deliver the best results.