CX just got a seat at the table — TheyDo Agent live reveal recap

The deck lands on Friday at 4pm. The budget meeting is Monday at 9am. You know how this goes.

On May 27, we showed what happens when it doesn't have to go that way. Jochem van der Veer, Rebecca Dickson, and Louise Kampmann Kahlen — Journey Management Specialist at YouSee, Denmark's largest telco — spent an hour walking through TheyDo Agent live: what it does, how it works inside a real CX operation, and what it means for the teams that have spent years building customer context that never quite made it into the room where decisions happen.

The recording is above. Here are the highlights.

What you don't want to miss

Louise's problem, which is probably also yours

Louise runs journey work at YouSee across the internet, TV, mobile, and enterprise, with a team that cannot absorb the volume of questions coming at it from the rest of the business. A live poll across the attendees confirmed this isn't unusual: 65% said it takes their team one to two weeks to answer a meaningful CX question, and 26% said it takes longer. "It would be weeks for us. Sometimes months", said Louise.

A six-month-old PowerPoint, and what Louise did instead

A colleague requested access to a deck Louise had built half a year earlier — a familiar enterprise pattern where insights live in files and files live in people's inboxes. Instead of scheduling a meeting to reconstruct the analysis, Louise opened Ask TheyDo, typed the exact question her colleague had sent in chat, and got a document back. The colleague's first response was to ask where the insight came from, which meant she was already acting on it — not reviewing it.

The data sources problem

Jochem asked the audience how many data sources their teams pull from to answer a single CX question. Over two-thirds said four or more, with many landing at seven to ten, or beyond ten, with manual reconciliation sitting on top of all of it. That's not a workflow problem. That's what happens when there's no shared layer that different data sources can map to — and it compounds with every new team, market, or product line you add.

Why AI quality isn't an AI problem

"AI quality is not necessarily an AI problem. It's actually a data management problem." Jochem's framing of the AI landscape is worth watching in full, but the thesis is this: AI outputs in customer experience fall short not because the models are weak, but because the underlying data has no shared structure for the model to reason from. Journey Management provides that structure — a common layer that business performance data, VoC signals, and individual customer interactions can all flow into. Once it exists, AI has something real to work with.

The Friday 4pm message — TheyDo Agent in action

Rebecca sets the scenario — an executive message arriving Friday at 4pm asking for a complete CX update, churn analysis, and funding recommendations ahead of a Monday morning budget meeting — and then shows what happens when the CX team has been building structured context in TheyDo. This is where the product reveal begins.

Three minutes

That's how long it takes TheyDo Agent to produce a sourced, structured executive document from the question. It reads across journey phases, pulls VoC data from Medallia and Qualtrics, draws on metrics from Snowflake and Databricks, and checks what's actively in progress in Jira — then produces something a CX leader can walk into a board meeting with. Not a draft to refine over the weekend. "This is not just the narrative. This becomes the funding conversation."

Same context, new audience

After the executive document, Rebecca shows the same underlying evidence reframed for the product team — reshaped into a sprint brief without rebuilding any of the source material. The context exists once. The framing adjusts to who's in the room. "Instead of CX being a reporting function, you've now been elevated to the most strategic function in the enterprise."

The governance demo

This one is for practitioners. It shows how TheyDo Agent maintains context, not just surfaces it. A journey practitioner uses an Agent skill to govern an entire workspace with a single instruction — updating tagging, linking opportunities to solutions, flagging duplication, and integrating a post-merger dataset that arrived as unstructured files. The alternative is weeks of manual work.

How do you know you got it right?

Louise's answer comes from firsthand use: prompt carefully, review continuously, and use inline editing to refine rather than redo. Rebecca adds that the models score between 80–90% accuracy, every output traces back to the source material, and the activity log shows exactly what the Agent did. Not a black box — a transparent record.

What's next

These capabilities are now live. If AI is enabled in your workspace, your admin can activate it. If it isn't, your CSM can sort that.

TheyDo Agent runs on the customer context your team has already built — not generic AI applied to CX data. Every output traces to evidence and is structured enough to send to a CXO.

We're also building TheyDo MCP, so teams across the business can access structured customer context through the LLMs they already use, without needing to navigate the journey themselves.

Talk to an expert to see TheyDo Agent in practice.