All at a glance: Chat with your business data

Artificial intelligence and digital assistants have long since arrived in everyday working life. But many companies are still faced with the question: Do chatbots really bring concrete added value – or are they just a technical toy? The answer depends heavily on the use case, the database and the embedding in existing processes. We take a look at field-tested use cases and show where AI creates real benefits – and where classic tools are more suitable.

AI-powered chatbots are not new – but with large language models like GPT-4, they have reached a new level. They understand natural language, generate context-based responses, and can be used in a variety of ways.

In the corporate environment, however, it is crucial that AI is only connected to relevant systems such as ERP, CRM or HR that makes AI really valuable. After all, real added value is only created when structured, company-specific data can be accessed – for example, through automated processes, personalized evaluations or efficient support in day-to-day business.

The Microsoft 365 Copilot is technically based on OpenAI's GPT models, but only accesses data from its own Microsoft tenant – i.e. Outlook, Word, Excel, Teams, SharePoint or Power BI.

The advantages:

• No data sharing with OpenAI or other customers

• No use of the content for model training

• Strict access control and rights concepts

This allows business data to be used securely and efficiently – whether for creating presentations, writing e-mails or summarizing meetings. About the Copilot Studio , companies can even set up their own agents, whichare also connected to SAP or their own data sources, for example – without additional exports.

SAP Joule is part of SAP Business AI and acts as a generative chat assistant that works similarly to Microsoft Copilot to execute questions and actions based on SAP data.

In many companies, everyday inquiries revolve around recurring topics. For example, in HR: “How do I apply for vacation?”, “What do I have to consider when taking parental leave?” or “What benefits does the company offer?” – questions like these make up a large part of HR communication. An HR chatbot can step in here as first-level support: It provides answers quickly, refers to relevant documents and noticeably relieves the HR team. Support can also be automated for IT inquiries (e.g. ticket overview, password guidelines, device changes) – with a connection to existing systems.

The advantages of chatbots are particularly evident in sales. A sales representative is on the phone with a customer, is asked for a technical specification – and retrieves the information directly via chatbot. No switching to other systems, no searching in documents – the bot delivers the data directly in real time.

The use in service goes even further: A technician stands in front of a machine at the customer’s site with an error message. The service bot provides him with possible causes, suggested solutions and even hints as to which colleagues have successfully solved similar cases. Ideally, this works voice-controlled – without a keyboard or mouse.

Another particularly exciting use case from the field of sales is the “lead enricher” – i.e. the enrichment of leads: Instead of manually researching information (e.g. whether a company uses SAP), an agent takes over this task. It searches career pages, LinkedIn profiles, blog articles or partner references and provides structured answers – such as whether a potential customer fits into the target group. In the second step, this information can even be automatically transferred to the CRM.

Especially in the business development phase , this can save up to two hours per lead – time that can be invested in personal conversations or individual offers. The goal is clear: streamline processes, save time, increase quality.

Despite all the enthusiasm, not every use case is suitable for a chatbot. The question "Show me my top 10 opportunities, sorted by project volume" can usually be answered more quickly via a classic Dashboard or Reporting answer. The strength of chatbots lies in the contextual information retrieval, not in complex evaluations.

So the right question is not "How do we use AI?" – but: "How can we design processes in such a way that a bot can provide meaningful support?" And sometimes the best solution is not AI, but a well-thought-out process or a clean system.

The discussion about chatbots should not start with technology, but with processes. Only those who know the specific needs in the company can develop meaningful solutions. Good bots work with existing data, are specifically trained – and create measurable added value. Whether in HR, service, sales or marketing: The transformation from reactive to proactive AI starts with the right use case – and with the willingness to question and rethink processes.