Where enterprise AI is heading, according to the people building and deploying it.
From 8 to 9 July, the Remobi team spent three days at RAISE Summit 2026 in Paris, one of Europe's largest AI conferences. We spoke with founders, engineers and technology leaders across healthcare, legal, procurement, AI infrastructure and developer tooling, and sat down with six of them for longer interviews on where AI goes next.
We heard the same themes at Big Data London 2025. Enterprise AI is becoming an engineering challenge as much as a technology one.
Here are the trends shaping that next phase.
Key takeaways:
- AI adoption depends on governance as much as technology.
- Infrastructure has become a strategic part of the AI stack.
- Specialist AI tools are solving operational problems faster than general-purpose platforms.
- Human expertise remains the competitive advantage.
- Organisations investing in engineering capability will scale AI more successfully.
Why enterprise AI adoption now depends on governance
Governance has become the first challenge of enterprise AI delivery.
Conversations across procurement, healthcare, legal, product and education all pointed to the same requirement: organisations need AI they can trust in production.
One conversation at the summit turned to public sector procurement, where governance is the biggest barrier to adoption. Public sector organisations cannot deploy AI without confidence that sensitive data remains secure and under their control. Enterprise adoption depends on strong governance, clear ownership of data and trusted supplier relationships.
Healthcare demands the same level of confidence.
Speaking to someone at the summit about healthcare, the expectation was clear: AI will transform the sector over the next three years, but widespread adoption depends on reliability. As they put it, AI is "still sloppy right now" and "not ready for production". In safety-critical environments, systems must deliver consistent, dependable results before they can support everyday clinical decisions.
Related read: AI Enterprise Forum
AI infrastructure is becoming the next competitive advantage
As enterprise AI moves into production, organisations are investing in the compute, storage and orchestration needed to run AI reliably and at scale.
Verda is tackling the efficiency of AI infrastructure, addressing the energy and operational demands created by growing compute requirements. Scaling AI depends on solving the physical constraints behind it.
Backblaze focuses on another critical part of the stack: storage. Training datasets, embeddings, checkpoints and inference data all need to be stored, managed and accessed at scale. Storage rarely dominates AI conversations, but it underpins every production deployment.
Impala AI is optimising inference with infrastructure designed for asynchronous, large-scale workloads, giving engineering teams greater control over performance without manually managing capacity.
Fastino and Featherless are solving a different challenge. Both focus on making AI more efficient in production, whether through specialised small language models or simpler access to open-source models. Their work reflects a broader industry priority: improving performance, reducing costs and making AI easier to deploy.
Why specialist AI tools are creating more enterprise value
Many of the strongest products at RAISE were designed to solve one operational problem exceptionally well.
ENDOless focuses on chronic gynaecological conditions such as endometriosis, combining AI with structured health data to improve long-term understanding while protecting patient privacy.
TACTICA AI operates in defence, national security and critical infrastructure, bringing together multiple sources of operational data to help decision-makers assess complex situations more quickly. One principle summed up its approach: AI recommends. Humans decide.
Developer tooling is following the same path.
Sentry's Seer extends observability into AI-assisted debugging, using production telemetry to help developers identify root causes more quickly. JetBrains gives engineering leaders greater visibility over AI usage, governance and cost across development teams. Datadog is extending observability into AI workloads, helping teams monitor the reliability and performance of increasingly complex systems.
Different problems, same design philosophy. Enterprise AI creates the most value when it is built around a specific workflow, user or operational challenge.
How AI is changing professional expertise
AI is changing how people learn, make decisions and work together.
One of our conversations focused on law, where AI is expected to remove much of the repetitive work behind due diligence, giving lawyers more time for negotiation, strategy and client relationships. It should also help junior lawyers build expertise faster, as AI accelerates routine tasks and exposes them to higher-value work earlier in their careers.
Another conversation looked at product teams, where organisations are expected to manage AI agents much like new team members. As one person we spoke to put it, "you really have to treat them as a team member": define the role, give them the right context and manage the quality of their output. Product teams will play a central role in making that work.
Someone else we spoke to focused on the gap between capability and understanding. AI is advancing faster than people's grasp of how to use it effectively, so organisations are building the tools before building the skills. The hope for the next decade is greater harmony between people and AI through education, understanding and responsible adoption.
“In ten years, I hope and I want that we will be in harmony with AI, because now it's disturbing for the humanity, and we are a very small group of people that trust in this idea.”
Another person we spoke to looked beyond software altogether, pointing to AI increasingly moving into robotics, physical systems and human-machine interfaces.
AI is reshaping work across healthcare, legal, procurement, product and education. Organisations that combine AI with human judgement and domain expertise will create the greatest value.
Related read about the Human Side of AI here.
Five priorities for enterprise AI leaders
- Build governance before you scale. Clear policies on data, security and responsible AI create the confidence to deploy AI across the organisation.
- Treat infrastructure as a strategic investment. Compute, storage, observability and inference determine how reliably AI performs in production.
- Solve specific problems. The strongest AI products are designed around defined workflows, users and operational challenges, not broad capabilities.
- Invest in people alongside technology. AI delivers the greatest value when it's paired with domain expertise, engineering judgement and strong product thinking.
- Develop AI capability across the business. Organisations that teach teams how to work effectively with AI will move faster than those investing in technology alone.
The next phase of enterprise AI
RAISE Summit 2026 reflected an industry focused on delivery.
Across every conversation, the priorities were the same: governance, infrastructure, observability and the specialist expertise needed to turn AI into production systems. This echoed what came out of Remobi's AI Enterprise Forum, reinforcing that enterprise AI has entered a new stage of maturity.
AI will continue to evolve. Organisations that succeed will build the engineering capability, product leadership and operational discipline to evolve with it.
We'll keep following the trends from RAISE and the conversations shaping enterprise AI.
Met us in Paris? Let's keep the conversation going. Didn't? Follow Remobi for more from the people building the next wave of enterprise AI.