TinMan Professional Services

Implementing AI and expert systems can be a fairly significant undertaking, and is usually at the heart of the system value proposition itself. By utilizing members of our team that are working with a broad set of algorithms and approaches on a daily basis, the time and costs associated with your project can be greatly reduced.   We can even just review and provide written analysis of your project status, in addition to optionally building and integrating. Systems can be stand-alone PC - based or fully web-based. Hosted, local or embedded.


Click here for a sample of a few basic graphical visualizations of live sensor data and neural network for the web.

ai software examples

We can quickly determine which type of AI technology is most appropriate for your application. There are many different types of AI technologies - Classifier Systems, Genetic Algorithms, Bayesian Networks, Fuzzy Logic, Natural Language Processing and Understanding, path-finding, neural networks, and even statistical and analytics-based rules systems. Our approach is to design the system that meets your needs best. Often times, this will involve more than one best of breed approach.

Typical Process Steps:

  1. We'll sign an NDA to protect your IP and project.
  2. Review your application/project requirements, target host environment, status and desired timeline.
  3. Review the data elements of the system, along with determining the frequency of their presentment and availability to the decision making structure. This will include either the review of actual raw data, or the production and discussion of sample data that is deemed suitable for use in development to follow.
  4. Review the types of decisions that must be made along with the method of the delivery and availability of those decisions.
  5. Review all families of AI algorithms (as stated above) to determine the most appropriate approach (which may include multiple approaches). Note: Neural Networks may not be the best candidate, and therefore one or more of our other 50+ algorithms families may be better suited for your application.
  6. Develop a straw-man design of the decision making structure, including layout, approach and design. 
  7. If your project requires natural language processing, we would begin building the 'lexicon' of syntax relative to the use domain.
  8. Provide a stand-in module (semi-functional prototype) for integration and process integrity check while actual custom AI development gets underway.
  9. Custom AI modular component development
  10. Delivery and integration of AI module(s)
  11. Testing and fine-tuning of structure and decision selection
  12. Ongoing support and maintenance