Emil: the AI employee who prepares the introduction of AI

You rarely read about one of the biggest pain points when introducing AI. In large companies it comes up almost every time anyway. Before anything starts, all stakeholders have to be brought on board properly. Exactly these alignments beforehand cost an enormous amount of time. And this is precisely where Emil comes in.

This article describes a complete process that a single AI employee at scoreprise.AI is responsible for, from the first minute to the finished report. Not as an idea, but as a workflow we actually run.

Why discovery determines the quality of every AI solution

Before an AI employee is created at scoreprise.AI, we have to genuinely understand the process. Capabilities, integration, role within the team. This only fits precisely when we know how the process actually runs today, and not from the perspective of a single person, but from the perspective of everyone who lives the process every day.

An example from business intelligence makes this tangible. A BI process runs through sales, management, marketing and the BI team. In the projects we accompany, that is regularly 20 relevant voices, sometimes 30. Each one contributes information that determines the quality of the later solution. The sales person knows which analysis they are missing in the customer conversation. Management sees the steering logic. Marketing knows the campaign reality. The BI team knows the data sources and the pitfalls in between.

If one of these voices is missing in the preparation, it shows up later. Namely in the implementation, when the AI employee is already being built and suddenly a requirement is on the table that no one mentioned before. That is where it costs the most, because then we have to readjust instead of building it correctly from the start. Discovery is therefore not a preliminary skirmish. It is the part that determines success and rework.

How companies gather these voices today

To this day, companies gather these voices through workshops. Whole days spent in such sessions, just to capture every perspective in a controlled way. Calendars from five or six departments have to be aligned. People sit in a room and talk in front of their colleagues and often also in front of their own boss. That is expensive, it is slow, and it colors the answers. In a large group, hardly anyone openly says where their process really gets stuck.

Our aim is different. We want to relieve the customer before our AI employee even starts the actual onboarding. Instead of demanding a whole workshop day from the customer, an AI employee takes over the discovery at our company. His name is Emil.

Who Emil is

Emil is an AI Research Assistant, one of our AI employees. Not a chatbot and not a form with a nice interface. The difference matters to us. Assistants respond. AI employees work independently. Emil conducts the discovery on his own, in the form of real conversations with the relevant people in the customer company, and he does so by voice, in a natural, German-sounding spoken dialogue. No robot tone, no accent, but a conversation that feels like a well-prepared one-on-one.

Emil is responsible for the complete discovery process. From the first briefing through every single interview to the consolidated report with a clear recommendation on how we should proceed. Our contact person is only present in the meeting themselves at two, at most three points. Emil carries the rest.

How the process begins

At the start there is a single real meeting. A kick-off with the main stakeholder on the customer side, often a VP Sales or a Head of Business Unit, especially when it comes to introducing intelligence solutions. In this meeting we introduce Emil, we clarify the goal of the project, and we brief Emil on the customer. After that he takes over.

In parallel, we set up the customer folder in the portal, for example Kelvion, and name the main stakeholder. The main stakeholder receives a link through which they enter their team, with name, role and email. In doing so, they enter all relevant voices, across the departments the process touches. This way the customer defines for themselves who has to be heard, and no one is forgotten just because they are not part of the obvious core team.

The special role of the main stakeholder

A detail that is easily overlooked but makes a big difference. The main stakeholder is not interviewed about themselves. They are the source, not the subject. Emil interviews them exclusively about the other stakeholders they have invited. So about their role, about the relationship within the team and about who is likely to push and who to hold back. In short, the main stakeholder confirms the map on which Emil then conducts his conversations. In the portal they are therefore also shown visibly separated. At the top themselves, below them their team.

How Emil interviews

Every entered stakeholder receives a personal invitation link to a voice conversation with Emil. They then speak whenever it suits them. No shared calendar, no waiting for the one free day. And now comes the part that distinguishes Emil from every form and every standard survey.

Emil does not interview blindly by a fixed schema. Before every single conversation he consolidates what has been said so far. So he already knows the answers of the colleagues who spoke before this stakeholder. On this basis he recognizes patterns, uncovers contradictions between the stakeholders and tailors his questions precisely to the role of the next conversation partner. He does not go into the conversation with an empty list, but with an already formed picture that he wants to sharpen deliberately.

In every conversation Emil works out the needs, wishes, efficiency-gaining requirements and the real pain points of the stakeholder. How does the process actually run today. Where is time lost. What would an AI employee need to be able to do so that this one person is noticeably relieved. Which systems depend on it. And because Emil brings the knowledge of the previous conversations with him, he can follow up where a form would have stayed on the surface. A prepared one-on-one conversation simply goes deeper here.

The result is not a stack of isolated interviews. It is a growing, self-sharpening process of insight. Every conversation makes Emil smarter, and every additional stakeholder benefits from the knowledge of the previous ones. In the end there is a picture of the organization that is denser and more honest than anything a single workshop day could deliver.

What the portal does in the meantime

In the background, all of this runs through a portal that makes the process visible and controllable. For us it is the command center, for the customer it stays pleasantly simple.

At the top level are the customer folders, one per running scoping process. When we open a customer, we see the complete stakeholder list with their respective status. Who is invited, who has already had their conversation with Emil, whose interview has been analyzed. A progress view shows at a glance how far the discovery is, without us having to ask anyone.

Below that are the details per stakeholder. The conversation that was held, the points that were worked out, the patterns Emil recognized and the places where statements from different stakeholders contradict each other. These contradictions in particular are worth their weight in gold, because they show where in the company different views of the same process exist that never sat side by side before.

The people on the customer side see none of this except their own, warm entry point. The main stakeholder gets a plain page on which they enter their team. Every stakeholder gets their personal link to the conversation. All hosted in Germany and GDPR-compliant, transparent for everyone who takes part. For the participant it does not feel like a tool, but like a short, respectful conversation in which their view really counts.

From the last interview to the finished report

Once Emil has conducted the conversations, he turns to us. He briefs us on the entire customer. And that is not a static report to read through, but a real dialogue. We can talk to Emil, follow up, challenge his hypotheses, dig into details. At this point Emil is our sparring partner about the customer, not just a reporting tool.

After that we go together into a final conversation. We ask all our questions, Emil asks the ones still open. What comes out of it he incorporates into the report one last time and sharpens it.

This is how a consolidated picture of the entire organization emerges. A report across all stakeholders, enriched with Emil's experience from the process and, this is the decisive part, with his concrete recommendation on how we should proceed. We then present this report to the main stakeholder in the closing meeting.

For us, this usually leaves two to three appointments per customer. The kick-off, a possible interim point and the closing presentation. In between, Emil works autonomously. The most time-intensive phase in sales, the discovery, no longer depends on our calendar, but runs through an AI employee who can run any number of scoping processes in parallel.

What the customer gains from it

The value lies on several levels, and none of them is marketing.

• Depth. Instead of the single view of one manager, the customer gets a complete, multi-layered picture of their own organization. 20 to 30 real voices instead of one assumption.
• Honesty. In a one-on-one conversation, people often tell a neutral, well-prepared conversation partner more about their true pain points than in a large group in front of their own boss. Exactly these honest points are the ones a good solution is later built from.
• Relief. No full-day workshop, no calendar tetris across five departments. Everyone speaks whenever it suits them.
• Quality of the solution. Because the scoping is based on the real, consolidated needs of all users and not on the assumption of an individual, the later AI employee meets the actual need. This avoids exactly the expensive readjustment that otherwise shows up in the middle of the implementation.

Why Emil is a logical instance that is needed

Finally, the point that goes beyond this one product. Big decisions are, to this day, set from the top. AI decisions in particular often come top-down out of IT, because a medium is missing that truly brings all levels on board. Not out of malice, but because gathering all perspectives properly has so far been so expensive and so slow that it was simply skipped.

Exactly this medium is what Emil can be. He makes gathering all voices so cheap and so fast that there is no longer a reason to skip it. This turns the logic around. Introducing AI no longer has to be imposed from above, in the hope that the users will go along. It can build on what the users themselves have said.

We tested Emil early on with existing customers. Their reaction was clearer than we had expected. This kind of interview, they said, should really be used for every internal alignment. And that is exactly where the true scale of the idea lies. Emil emerged as an AI employee for the discovery in the introduction of AI. But a system that gathers every relevant voice cleanly, honestly and without great effort and condenses it into a clear picture is needed everywhere organizations make big decisions.

Emil simplifies the introduction of AI by taking over the most expensive and slowest step before it. And at the same time he shows how one of these decisions can even come about in the future. Not set from above, but heard out of the organization.