If you’ve ever taken a longer look into Applica’s revolutionary intelligent document automation, you’ve most likely heard of TILT—our proprietary technology that enables unprecedented usability in the automation of all kinds of business documents. It’s what allows Applica to extract information from documents that feature unpredictable content and running text. Just as importantly, it makes this process fast, reliable, lucrative, scalable, and operable by end users who don’t need to be engineers.
TILT is an acronym that stands for “Text-Image-Layout-Transformer”—a functional description that highlights the way our ground-breaking technology has an unprecedented ability to read graphical embeddings. It is thanks to the versatility of TILT that Applica’s deep learning-based solution is able to automate all sorts of documents, not just structured ones (like tables and forms) and unstructured ones (such as contracts full of text), but also ones that combine structured, unstructured, and graphic elements in unpredictable ways. Having been out of reach for the text automation solutions of yesterday, graphical embeddings (such as signatures, nested tables, and radio buttons), have been one of the major obstacles along the path to full-spectrum document automation. Thus, what Applica’s developers have managed to accomplish really is changing the very scope of text scrutiny and document automation.
Another distinctive advantage of TILT is the model’s generativity, which maximizes the utility of each piece of training data. This is what allows the machine to learn on a minimal number of samples. And the less data the solution needs to see, the less time, effort, and cost is required for setting up each business operation. Fittingly, then, the acronym TILT also communicates a shifting of perspective. After all, Applica really has “tilted” the technology to strategically expand the field of possibility for the entire sector.
Best of all, with TILT Applica brings you the most capable and versatile back-end in the industry and the most intuitive front-end to match. Our progressive user interface deploys a natural question-and-answer format and assists you in training our TILT deep learning model to perform tasks exactly as you need them done. And because our easy-to-use system lets you use plain human language to micro-calibrate TILT, it’s easy to apply horizontally across all of an organization’s departments, so that more processes can be automated, for unprecedented speed, accuracy, and scalability. Unlike coding, which requires expert IT knowledge, asking questions comes naturally to everyone. Users simply complete phrases using ordinary language. Sometimes, the questions and answers need some refining. That’s when synonyms or including additional details come in handy. Anyone is bound to get the hang of it because it’s no more difficult than using the world’s most popular Internet search engine.
And the training is a cinch, because instead of hundreds of sample documents only a small amount will suffice. How is this possible? Because our TILT model is pre-trained exclusively on a mind-bendingly large set of various relevant business documents and not on irrelevant random online text, as is typically the case. In terms of the time this saves—both for the individual worker training the AI and for the company deploying a process—is a quantum leap forward from the hundreds of training documents many competitors would require. In fact, Applica is the only commercially available generative language model trained on millions of business documents. TILT really is the only solution of its type available on the market.
Another distinctive feature of TILT is the simultaneous way it comprehends all aspects of the documents it processes. In contrast to the sequential step-based nature of traditional models, our non-sequential approach eliminates compounding errors and in fact safeguards against errors through a process that entails constant fact-checking and recursive verification. From the semantics of a document, which has to do with its type and purpose, to the specific data communicated through context, graphics, layout features, and raw text—Applica does not prioritize any of the above to the detriment of the remaining elements. What you get is what is there, holistically.
.png)
