Qlik's AI Council warns companies to prepare for the most difficult part of AI.

AQlik® This brings a warning to the action of its AI Council regarding the changes that companies must prepare for as AI advances and deepens in supporting decision-making, workflow execution, and day-to-day operations.

The Council's message is clear: the next phase of AI will be shaped by forces that many organizations are still underestimating. Evaluation and accountability will carry more weight. Regulatory environments will continue to fragment. The quality of reasoning will face greater scrutiny. The turnover of models and interfaces will continue. Architectural choices will determine how quickly companies can adapt without having to reinvent themselves repeatedly.

“AI is entering a more difficult phase with more consequences,” says Mike Capone, CEO of Qlik. “The easy conversations are over. Access to powerful models is widespread. The more difficult question is whether AI can operate within the real-world conditions of a business, with reliable data, responsible reasoning, evolving regulatory requirements, and the flexibility needed to continue adapting as the market changes.”

Five perspectives on what companies should prepare.

“Many organizations still treat governance as a set of documents,” says Dr. Rumman Chowdhury, a responsible AI leader, engineer, auditor, and investor. “This approach will fail under real pressure. As AI gets closer to decisions and actions, trust will depend on evidence. Evaluation needs to occur continuously, under real-world conditions, with clear signals of when systems are reliable and when they are not.”

“The next AI divide will be shaped by power, access, and dependence,” says Nina Schick, Author, Advisor, and Founder of an AI Consultancy. “Intelligence is being industrialized, concentrated, and contested all at once. Leaders need to think beyond decisions about tools and focus on whether their organizations are structured to adapt as the configuration of the AI ​​economy transforms.”

“Regulatory fragmentation is becoming an operational reality for global companies,” says Kelly Forbes, Co-Founder and Executive Director of the AI ​​Asia Pacific Institute. “Different markets are advancing at different speeds, with different expectations regarding transparency, work impact, oversight, and acceptable use. Companies that can scale effectively will treat coordination and adaptability as core capabilities from the outset.”

“A fluent result can still reflect superficial reasoning,” says Michael Bronstein, Professor of Artificial Intelligence at DeepMind, University of Oxford. “The systems that matter in business will be those capable of working with structure, relationships, and constraints. It is the context that makes intelligence useful within a real organization.”

“The model layer will continue to change faster than most enterprise planning cycles,” says Mark Relph, Director of Go-To-Market (GTM) for Data and AI at AWS. “Companies should operate on the premise that new models, new assistants, and new orchestration patterns will continue to emerge. The most enduring choice is to remain open, governed, and ready to adopt what works without overhauling the entire system every time.”

Taken together, the council's perspectives point to a more demanding standard of AI readiness. Companies will need systems that withstand scrutiny, operate with reliable context, incorporate better models as they emerge, and remain useful as business, regulatory, and technical conditions continue to change.

This perspective will guide a broader conversation at Qlik Connect® 2026, where Qlik will announce a coordinated set of launches focused on agentic analytics, open and reusable databases, operational trust, and sovereignty-ready implementation. Together, these announcements reflect a practical vision of what enterprise AI demands now: useful under pressure, explainable when questioned, and adaptable as conditions change.