How Autodesk built the shared context that makes enterprise AI trustworthy

On the stage screen at Forrester CX Forum in San Francisco, the audience was looking at a stark picture of the state of customer experience. According to Forrester's Predictions 2026, just 33% of executives trust the customer experience data flowing through their organization, while 60% of CX leaders can't connect customer experience to business metrics. 

The numbers highlighted a fundamental problem: organizations still haven't built the trusted, connected view of the customer that AI depends on to provide answers.

That's where Florian Vollmer, Director of Experience Insights at Autodesk, picked up the conversation.

"What AI can't synthesize is agreement,” he told the audience of CX leaders and designers. Then, over the next thirty minutes, he explained why AI still struggles to produce answers people trust. The issue isn't the models themselves. It's that most enterprises have never built a shared understanding of the customer for AI to reason over.

Autodesk ran into exactly that problem. What the company built next offers a practical example of what happens when journeys become the shared context from which both people and AI can work.

Where AI hits a wall

Autodesk isn't short on AI. The $7.2 billion design and manufacturing software company, whose tools help build everything from factories to feature films, uses Atlassian, Copilot, and multiple large language models across the business. It's also a deeply federated organization. 

About 8 to 10 groups are responsible for designing customer experiences. Each works a little differently and sees a different part of the customer. That's exactly where AI begins to struggle.

"The LLM gets you surprisingly far," he explained. "Until you ask the same question twice." The answers change because the context changes. Different teams have different information, and the models faithfully reflect it. The issue isn't the quality of the AI. It's that the organization hasn't yet built a shared understanding for the AI to reason over.

Shared context becomes infrastructure

When Florian started Autodesk's journey management practice two and a half years ago, he was, as he put it, "a body of one." Today, that practice has grown into a Journey Management Center of Excellence, where about a dozen people support teams across customer experience, marketing, digital experiences, platform design, and Autodesk's industry groups.

Most of the tools people rely on, such as Atlassian, large language models, and business applications built for individual teams, were already in place. The challenge was getting them to work from the same understanding of the customer.

Journeys provide that common thread. Research, operational metrics, personas, jobs to be done, opportunities, and business data are all connected back to the journey, so teams looking at different parts of the business are still working from the same context.

No one has to abandon the systems they already use. They just stop each working from their own version of the customer, and start working from the same one.

AI drafts. Humans decide.

At Autodesk, AI does the first pass, pulling evidence together in seconds. Customer feedback, research, support interactions, operational metrics, and dashboard data all flow into the system. AI groups related signals, drafts opportunities, and suggests where teams might focus next.

People then review the output, decide whether the signals belong together, interpret their meaning, and determine which opportunities are worth pursuing. Prioritization happens through RICE, using Reach, Impact, Confidence, and Effort as a common framework across product, digital, and marketing teams.

On stage at Forrester, Florian demonstrated what that looks like in action. A narrative built around a single "How might we?" statement, complete with linked insights, supporting evidence, opportunities, and journey context, appeared in about ten seconds. "And it's spot on," he marveled.

Before AI, assembling the same story took roughly two hours in PowerPoint.

Connecting customer experience to business outcomes

At Autodesk, journey management is powered by TheyDo’s experience context platform, which surfaced roughly 50 pain points across the customer experience. Working with the CX action management team, those opportunities were prioritized and acted on through the platform.

The result was a 10-point improvement in Customer Effort Score, along with measurable gains in customer retention.

Florian said the work is increasingly moving beyond operational improvements. Alongside fixing today's friction points, the team invests in what he calls Horizon 2 research, creating vision prototypes for experiences that may be realistic in three or four quarters. The goal isn't to predict the future years in advance. It's to give teams something concrete they can design toward now.

The real job of AI

Autodesk's experience suggests that enterprise AI succeeds or fails long before anyone writes a prompt. Models can summarize information, connect signals, and surface patterns. What they can't do is create agreement where none exists.

"Journeys provide that shared context, empowering decision-making," Florian said. "Not only listening to the loudest voice in the room, but creating an equal playing field through the information that is available in the journeys that we are managing centrally."

That's why Florian sees journey management as more than a way to document customer experiences. It gives people, and increasingly AI, a shared understanding to reason from.

For organizations struggling to trust their own AI outputs, that may be the most important lesson from Autodesk's experience.

Florian Vollmer presented “From Journeys to Shared Decision Context” at Forrester CX Forum San Francisco on June 29, 2026.