From Amsterdam to the Agent Era: Our Key Takeaways from Sitecore City Tour

Sitecore City Tour Amsterdam did not land as “just another date in the events calendar.” It felt like a signal that the ecosystem - customers, partners, and platforms - is stepping into a new operating reality for digital experience.

Yes, there were impressive product sessions, live demos, and hands-on code jams. But what truly defined the day was a shared recognition that we are moving into an agent-driven era, where the organizations that win will be those with the operating foundations to use AI safely, consistently, and at scale.

Horizontal had the opportunity to host a panel, “Same Platform, Different Worlds,” featuring leaders from IFS, HCA Healthcare, and Greystar Europe, representing technology, healthcare, and real estate. Different industries, but a strikingly similar set of questions around speed, governance, and value from AI.

Below are the key reflections we are taking away, featuring themes echoed by partners, community highlights and our panel.

1. Community over Features: The Real Headline

The day opened with a partner breakfast that set the tone: relationships first, slides second. It was a reminder that trust in this ecosystem is built in conversation and collaboration, not just in roadmaps.

Throughout the event, this mindset showed up in the way sessions were run: less hype, more practical dialogue about how AI is changing modern marketing and experience delivery. The energy was one of listening, challenging, and co-designing for real world impact, supported by a strong partner network and a packed agenda.

Our takeaway: The organizations that benefit most from AI will not be the loudest promoters, but the ones who combine innovation with trust, alignment, and the ability to execute together.

2. From “AI Assists Me” to “AI Operates with Me”

In previous years, the dominant storyline was copilots, AI as a smart assistant that helps individuals do their jobs faster. In Amsterdam, the narrative had clearly evolved: this was the year of agents.

Agentic workflows, autonomous execution, intelligent routing, and orchestration were at the forefront. SitecoreAI’s direction, illustrated through Sitecore Studio (including Agentic Studio, App Studio, Marketplace, and Connect), underscored a move toward an extensible ecosystem where agents are not a novelty layer but a core part of how work gets done.

Our takeaway: AI is no longer an add-on inside marketing workflows, it is reshaping the delivery model itself, shifting teams from managing isolated tasks to managing systems.

3. “Pages → APIs → Agents”: UX Is Being Re-Architected

One of the most quotable threads came from the Vercel perspective: we are moving from page-centric digital experiences to agent-centric ones. For years, we optimized journeys around navigation, clicks, and page hierarchies.

Agents flip that paradigm. Users express intent; agents leverage tools, APIs, and workflows to return outcomes rather than screens. Vercel’s technical session highlighted how coding agents can interact with SitecoreAI using skills, bringing platform intelligence into the build process itself.

Our takeaway: Conversational, outcome-led experiences are rapidly becoming the default expectation. Headless architectures, robust APIs, and composable design are no longer “nice-to-have patterns” they are prerequisites for delivering agentic experiences.

4. Microsoft’s “Frontier Organization”: AI Adoption as an Operating Model Change

Microsoft’s contribution framed AI adoption as a maturity journey most enterprises are now somewhere along:

Human with Assistant: today’s common scenario, with AI supporting individuals on discrete tasks.

Human-led Agents: agents perform defined tasks within clear boundaries.

Human-led: agent-operated organizations: humans set intent and guardrails; agents orchestrate and execute core processes.

Breakout discussions echoed this model, emphasizing that moving up this curve is less about installing new tools and more about rethinking operating models, decision rights, and governance.

Our takeaway: The future is not about AI replacing people; it is about humans setting strategic direction while AI scales execution. That requires deliberate change in how organizations are structured and run, not just another platform feature.

5. Platform AI versus Organizational AI: The Readiness Gap

A recurring theme across the day, and in our “Same Platform, Different Worlds” panel, was the gap between having AI-capable platforms and being an AI-ready organization. Being live on a modern DXP with embedded AI features does not automatically translate into value.

Our panelists, representing three large enterprises with distinct constraints and audiences, described similar pressures: faster time-to-market, rising expectations from stakeholders, and a market that refuses to stand still. Their different starting points on SitecoreAI illustrated how much organizational context shapes adoption.

Our takeaway: Technology is increasingly the easier part. The harder work is organizational readiness: adoption, alignment, operating maturity, and the ability to embed AI into real processes.

6. The “Adoption–Impact Gap” Is Widening

One line that resonated came from a partner recap: most enterprise employees are now using AI weekly, yet only a minority report clear, measurable revenue impact. Even if you debate exact numbers, the pattern is familiar - usage is up, impact is uneven.

The missing element is rarely “one more feature.” It is the scaffolding that turns experimentation into outcomes: clear goals, robust measurement, and alignment between strategy, data, content, and operations.

Our takeaway: Many organizations are rushing to automate the “easy 80%.” The next step is designing for conversion, decisioning, and value - something that depends on foundations, not enthusiasm alone.

7. AI Exposes Your Operating Reality - Good or Bad

A key theme we see in our own work was strongly validated in Amsterdam: AI amplifies whatever system you already have. If your current delivery model is built on duct tape and heroics, AI will not fix it; it will break it faster and at scale.

The tension leaders voiced was less about whether AI can generate content, and more about what happens next: how to validate outputs, stay on brand, maintain quality, and avoid small errors escalating into major incidents. Scenarios such as incorrect localization, hidden CMS debt, and poorly governed personalization rules illustrate how easily AI can introduce operational and commercial risk when guardrails are weak.

Our takeaway: Speed without structure simply accelerates risk. The real risk is not AI itself, but unsupervised AI running on top of fragile foundations.

8. Strategy, Operations, and Guardrails Must Interlock

At Horizontal, we regularly describe the need for scaffolding and guardrails to support AI-enabled experience delivery. The conversations in Amsterdam reinforced this view.

Strategy defines the why and what: target outcomes, customer expectations, and the experiences that matter most.

Operations define the how: teams, workflows, tools, and what “good” execution looks like day to day.

Guardrails define the rules of the road: quality standards, governance, lifecycle, and which activities can be automated or must remain human-reviewed and auditable.

Or, as we often summarize it: strategy is the map, operations is the engine, and guardrails are the seatbelt and speed limits.

Our takeaway: The differentiator is not AI capability in isolation, but an operating model where strategy, operations, and governance function as one system.

9. Discoverability and Trust Are Being Rewritten

One quote from the event captured the broader context: we have moved through two major eras of the web and are now in a third, shaped by large language models, where visibility is earned and trust becomes the core currency.

Search is shifting from a list of links to AI-generated answers, and in many cases, a customer’s first interaction with your brand is now mediated by an AI summary rather than your homepage. Early research suggests that when AI summaries appear in results, users are less likely to click through in the traditional way, which fundamentally changes how content must perform.

Our takeaway: Content must now be consistent, structured, and credible in environments you do not fully control, raising the bar for governance and quality at scale.

A Practical Checklist for “Post–City Tour” Teams

If you left Amsterdam energized, the next step is turning that momentum into action. Some questions we are asking with clients:

Where should AI assist, and where must humans decide? Map workflows and define safe automation boundaries.

Do we have a clear “truth layer” for brand, product, and policy? If your foundations are fuzzy, AI will replicate that fuzziness at scale.

Can we approve and measure at the speed we can now generate? Review, optimization, and analytics need to match AI-driven throughput.

Are we treating adoption as capability-building, not just communication? Teams need practice environments, sandboxes, and structured enablement.

Are we measuring outcomes, not output? Success is not volume of content, but impact on conversion, self-serve success, cost-to-serve, satisfaction, or pipeline velocity.

Modern digital maturity now looks like building solid foundations while the finish line keeps moving. That is both the challenge and the opportunity.

At Horizontal, we are focused on helping enterprise organizations move from “platform live” to “platform delivering,” aligning strategy, operations, and guardrails so that AI becomes a multiplier of quality and value, rather than a multiplier of risk. If you are rethinking what “AI-ready” really means after Amsterdam, we would be pleased to continue the conversation.

By James Gray, Digital Strategy Director – United Kingdom, Horizontal Digital