AI agents
AI solutions for multi-stage processes noticeably reduce your day-to-day workload.
AI agents can provide more active support to your business processes than conventional AI assistants. Not only can they answer questions, but they can evaluate information, determine the next steps or prepare, initiate or perform tasks as well. The key point is that where required, users retain control of the actions carried out by the agent.
In contrast to conventional automations based on workflows with clearly defined, structured inputs, AI agents are also able to handle unstructured information. Whether these are enquiries expressed in natural language, emails or incomplete statements, agents can interpret them and determine the process steps required. We can help you deploy AI agents in a way that makes sense, implement them reliably and integrate them into your Microsoft and unitop landscape.
Making AI technology usable
AI agents with Microsoft and unitop – our solutions at a glance
Microsoft BC agents
In Business Central Standard, Microsoft now provides AI agents that use data available in BC to perform tasks largely autonomously. They analyse data, understand contexts and connections, carry out tasks independently or initiate processes – typically starting out from an enquiry by email to compiling quotations and orders.
Find out more
unitop AI agents
unitop AI agents support ERP-related processes in the context of Business Central and unitop, but also beyond. You can collate information from unitop processes or other sources, prepare individual steps of a procedure or initiate defined actions. They can either work directly in the system or in an overarching way using additional applications and can help configure processes in a scalable and more rapid and efficient way.
Find out more
From answer to action
How AI agents differ from AI assistants
An AI assistant helps perform individual steps in a procedure – it answers questions, summarises information, compiles suggestions or helps analyse data. An AI agent, on the other hand, is more suited to carrying out multi-step processes and procedures. It can extract information from a context, use that information to determine what steps are necessary and use associated actions to initiate or perform tasks. Depending on the type of agent concerned, it can link multiple steps together: It collects information, evaluates interim results, asks additional questions and then launches defined actions. At the same time, AI agents operating in an ERP environment are always bound by existing permission and control mechanisms. When performing critical steps in particular, feedback is provided to the users (“humans in the loop”), so that decisions remain traceable at all times and can be actively confirmed.
AI agents to ensure structured business processes
The true value of AI agents particularly comes to the fore in contexts where processes are recurring, information-intensive or time-critical. They can reduce the workload in specialist departments and enhance customer service by preparing tasks, giving instructions, combining data or initiating defined steps of a process or procedure.
What sets them apart from simple automations is the fact that AI agents are more context-oriented. They can process natural language, contextualise incomplete information and determine the next steps to the best of their ability.
Typical scenarios for AI agents
The following are examples of typical scenarios where AI agents can provide assistance. Which specific tasks are taken on by an agent will depend on data sources, permissions and technical implementation.
An AI agent can evaluate incoming information, check suitable data sources and prepare a procedure for further processing. This means that your teams can receive a structured basis for the next steps much more rapidly than before.
During verification processes, an agent can highlight any anomalies, bring together relevant information and bring the persons responsible onboard. This will ensure that decisions and releases are managed clearly at all times.
Whenever data are stored in multiple applications, an agent can bring relevant information together and collate it for processing or decision-making. This will typically include information from ERP data, documents, communication or other specialist applications.
An AI agent can prepare or initiate defined steps if specialist rules, permissions and control points have been laid down in advance. In ERP-related scenarios, this may involve checking datasets, marking anomalies or preparing releases.
Innovating. Transforming. Succeeding.
Implementing AI agents dependably and securely
When using AI agents successfully, the starting point is not the technology, but clearly defined tasks, reliable data and a security framework that’s suitable for your processes.
How we work
Implementing AI agents step by step
Contact
Your contact for AI agents
Are you looking to identify areas where using AI agents to support your processes would make sense?
We will help you to find the right areas and work with you to implement AI agents dependably within your Microsoft and unitop landscape.
Frequently asked questions about AI agents
What is an AI agent?
An AI agent is an AI solution that doesn’t only provide information, but can also help perform tasks that are part of the process. It can evaluate information, determine the next steps and initiate or prepare defined actions. The precise actions an agent can perform will depend on data access rights, permissions and the technical implementation.
How do AI agents differ from AI assistants?
AI assistants help with individual tasks by providing answers, summaries, analyses or suggestions, for example. AI agents are more oriented towards performing multi-step processes and procedures. They can combine information, evaluate interim results and prepare or initiate additional steps.
Are AI agents the same as automations?
No. Conventional automations follow fixed rules and predefined processes. AI agents can work in a more context-oriented way, can process natural language and respond to different initial situations. Nevertheless, AI agents also require clear rules, permissions and verification by specialists.
Which solution is more suitable – Microsoft BC agents or unitop AI agents?
That depends on where you plan to use them. Microsoft BC agents are provided by Microsoft as standard solutions for all BC customers. unitop AI agents are tailored by us to meet the needs of individual, sector-specific unitop solutions or, if necessary, are adapted to the customer’s specific needs. This makes them especially suitable for specific processes in individual sectors or for individual customers. Other than that, certain unitop AI agents also make it possible for data sources outside Business Central to be connected. In many scenarios, these two approaches complement each other.