A commercial printing company in the UK, providing bulk paper and electronic documents for large corporations, approached Fountech Solutions to provide an Artificial Intelligence driven product that could facilitate interactivity with readers.
This client were sent raw textual content by their own customers, often this comprised of complicated legal or legislative instruments involving hundreds of pages. Their role was to print these documents onto paper and also provide electronic format copies.
Then they would distribute the resulting products by courier and file transfer to secure locations for the intended readership. They knew that if their electronic versions of documents could be easily ‘interrogated’ by readers to beyond simple keyword searching, their business would become a market leader in its field.
It’s already commonplace to produce e-format documents that are searchable by keyword, because it only requires non-AI driven software to identify any instances of a given specific word, then highlight it to the reader’s attention.
A much more complex concept is a document being interrogable for related content; either internal or external to the document itself. For this to be possible, the software would need to be able to respond to readers’ requests: asking about the definition of words, summarization, paraphrasing, simplifying, translating, classifying and perhaps expanding upon particular sections.
For this to work, AI driven algorithms have to ‘learn’ about given content and process it according to the nature of the request; it simply cannot be done with any practical level of speed or accuracy by non-AI computing. This makes AI especially advantageous when there is a need to change industries, domains, document content etc.
Let’s examine the requirement for a ‘smart document’ to be concerned, perhaps, with a complex legal case, running into perhaps hundreds of pages. The reader might pick up on a certain concept regarding case law over insolvency and disputed invoices. If the reader wanted to find out where else such content was referenced within that document (or perhaps within a given resource library of other legal documents), depending on device capability, they might ask by voice or by typing: “Tell me more about disputed invoices in insolvency caselaw”.
In order to achieve such a complex set of processes, our team had to produce a ‘knowledge graph’, (KG) representing semantic relationships between entities of document content. A KG is an intuitive storage system where data is stored in ‘nodes’, connected to each other with ‘edges’, which describe the relationship between the two nodes. A node is typically a reference to an entity and can have multiple connections.
Each node in turn can be linked to an array of information beyond the name of the entity it is describing, (referred to as attributes). Many nodes can be linked to an attribute; perhaps links to online sources and book titles relevant to the entity in question. Due to the nature of this graph, notwithstanding the constraints of computational resources, the possibilities are, almost literally, endless.
Classification and summarization modules were already available in Fountech’s toolkit as stand alone packages. By integrating these into the KG, we could enhance their efficiency. In order to convert plain text to a KG, we developed the Optical Character Recognition module that would convert scanned text to machine-encoded text. Then we built a Relation Extraction module to extract entities and their inter-relationships from that text.
Next, we built an Information Expansion Module to retrieve related information from the web and other publicly accessible databases. Modules were then designed to render abstractive summarization, paraphrasing and text expansion.
Finally, we designed an application programming interface (API) to interact with speech-to-text and text-to-speech modules, which would enable users to give speech commands or listen to the document with a choice of narrators. A simple Graphical User Interface (GUI) was created to demonstrate the complete solution to our client as a front-end example.
Our client was impressed by the functionality and speed of the solution, especially with the platform’s ability to work with both basic input and natural language speech commands. The fact that the software could listen to voice, speak and translate really consolidated its success.
We have received reports of complimentary comments from end users, and no significant complaints about functionality.
The vCAIO will conduct a situational assessment to identify the needs of, and opportunities for, the organisation. They will provide a report outlining a gap analysis and the baseline from where to start defining the AI strategy going forward
Create a vision of where the organisation should be going
Set standards for strategy implementation through establishing the correct management structures and considering ethical issues
Make things happen by operating at the right level, knowing what to outsource and what to achieve in-house for optimum delivery
Build capacity in the organisation by recruiting the right talent and identifying appropriate vendors or consultants to solve key problems
Become the public face of the new technology, both internally and externally
Understanding your needs and identifying opportunity to exploit
Evaluate your data
Knowledge Gaps | Costs v Benefits | Your Competitors
Latest Industry Tech and how we can enhance it just for you
Our key recommendations and suggested next steps