Luboslava Uram is COO and CTO at Solvd Group, a subsidiary of Allianz Group. Transforming the claim management experience.
getty
For years, integration has been one of the most expensive and time-consuming parts of any transformation. Whether it was driven by mergers, platform rollouts or system replacements, integration has often dominated budgets and timelines.
Having led global technology and operational transformation programs across multicountry environments, I have seen integration repeatedly become the single largest source of delay, cost escalation and organizational friction. In many programs I oversaw, the challenge was not the business vision or even the technology itself. It was the complexity of aligning legacy systems, local market processes, fragmented data structures and independently evolving platforms.
Many leadership teams still plan as if this level of integration effort is unavoidable. Integration is assumed to require years of synchronization before organizations can realize value. That assumption is increasingly outdated.
AI is not removing the need for integration. It is changing where cost, effort and time are concentrated—and how organizations can approach integration more pragmatically.
In working with my company's AI platform, I've learned that in global claims environments, insurers often operate across multiple legacy systems, local workflows, repair networks and fragmented data structures, making traditional integration slow and expensive.
Instead of relying only on rigid point-to-point integrations, AI can help interpret incoming documents, normalize data, connect operational events and adapt to local process variations dynamically. In leveraging platforms like these, our goal should not be to eliminate integration complexity entirely, but to reduce the amount of hard-coded customization required and start delivering value much faster.
Traditional integration relies on hard connections: fixed interfaces, strict data mappings and tightly synchronized processes. Each system needs to match another precisely.
This model works in stable environments, but it struggles when organizations operate across multiple legacy systems, different data standards and disparate local processes.
I've found that as complexity grows, integration effort increases faster than expected. Small changes trigger ripple effects across systems, making integration one of the biggest risk factors in large programs.
AI introduces a different integration dynamic. Instead of forcing systems to align perfectly up front, AI can interpret differences in data structures, translate between formats and meanings and adapt to variations over time. This shifts integration from a build-once exercise to a learn-and-adjust capability.
With AI, I've found that value no longer depends on perfect alignment before go-live. Systems can connect earlier, while AI layers absorb differences and improve continuously.
I've seen a few things firsthand when it comes to AI's role in integration. In postmerger environments, AI can help reconcile data and workflows across systems without immediate consolidation. Reporting, routing and decision flows can stabilize earlier, accelerating synergies.
In global platform rollouts, AI can reduce the need for deep local customization by handling variations in data and process interpretation. Integration effort moves away from rebuilding systems and toward managing exceptions.
In partner and ecosystem integrations, AI can shorten onboarding by interpreting incoming data and interactions, reducing ongoing maintenance. In cases like these, integration still exists, but its cost profile changes. AI does not eliminate the need for sound foundations.
Core systems still matter. Architecture still matters. Clear data ownership and accountability remain essential. AI changes how tightly systems must be coupled before value appears, not the need for disciplined design.
This balance is important. Overstating AI’s role creates unrealistic expectations and new risks.
One of the most important lessons I have learned is that AI-powered integration still requires strong operational and architectural discipline. Teams sometimes assume AI can compensate for unclear ownership, inconsistent processes or poor data quality, but in practice, those problems become even more visible in AI-driven environments.
In my work, our company focuses heavily on clear governance, well-defined operational events, shared integration standards and gradual rollout approaches instead of trying to redesign everything at once.
Another key principle we employ is keeping AI close to the product and operational teams using it daily. The most effective integration improvements usually come from continuous learning and adjustment based on real operational behavior, rather than large, centralized integration programs designed up front.
As integration economics shift, there are three assumptions that I believe leaders need to revisit:
• Where Integration Spend Goes: Investment moves from large, one-off build efforts to capabilities that support interpretation, orchestration and adaptation.
• What “Done” Means: Integration no longer needs to imply full harmonization before benefits are realized.
• How Quickly Value Can Be Delivered: When integration can evolve, benefits can appear earlier, reducing time-to-value and exposure.
These are leadership decisions as much as technical ones, and they are part of my day-to-day as a leader in this space. The benefits that come with AI, like increased capacity and better predictability, can help motivate teams to bring projects to higher levels of success.
Integration is no longer just a painful phase to complete and move past. In an AI-enabled environment, it becomes a strategic capability, one that allows organizations to operate across complexity without constant rebuilding.
AI is not making integration disappear. It is making integration more adaptive, more incremental and more economically viable. Organizations that recognize this shift must start planning differently. If they don't, I believe they may continue solving integration challenges using assumptions that no longer fully apply.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。