Contextual AI holds the key to its business value

Today, content management systems or services that claim to use AI typically have the memory of a goldfish. That is, they apply image or pattern recognition very well, but beyond that the value of the AI application is very limited. As each document is scanned and its content ‘read’ and input into the target system (e.g. invoice details into the relevant accounting system), this knowledge is treated as brand new: there is often little, if any, contextualisation. Documents and their contents are not linked intelligently with those residing or extracted from other documents. For example, a purchase order is not automatically being linked with a customer record in the CRM system, to provide new insights about buyers.

There is no doubt that process automation at this basic level is delivering decent value in time and cost savings, as well as enhanced customer services. But the next step in enterprise digital transformation, the one that will provide a real competitive edge, is the intelligent integration of content management more intrinsically within the wider enterprise ecosystem. When intelligent connections are made between documents, data and systems across different functions, this enables more effective decision-making and can help to streamline more complex processes.

Beyond image & pattern recognition

Single- process AI-enabled use cases, as delivered by intelligent order or invoice automation solutions, are ripe for integration with other enterprise systems now. This is about combining intelligent learning software with advanced content management to build synthesised institutional knowledge – so called “contextual AI”. Here, every document, and the data within it, can contribute to an intelligent knowledge bank that provides a holistic view of the customer, supplier, product development process – or any other strategic goal.

Next-generation content management platforms and services apply AI in new and smarter ways. Once the technology has identified the document type, it is able to process that information by inputting data into a database, or triggering the next step in a workflow.  

Take the ability to boost and transform the impact of customer relationship management systems (CRMs), in driving more targeted sales efforts and better customer service. CRM systems have long promised a “360 degree” account view (just as supplier management systems do in supply chain management). Yet these potentially very powerful systems are only as effective as the information fed into them, which up to now has relied on respective teams building up that picture of everything they know about a given account. That is, supplementary knowledge from accounts and so on, is not being fed or made available to these systems automatically, through direct, smart intelligence sharing across the enterprise. And this failure to join the dots leads to inefficiency. For example, a sales person won’t automatically be able to see that a customer has been asking the support team for new capabilities, which could be addressed in a new sale. They are thus missing out on a revenue opportunity by living in their local information silo.

With enterprise-wide, contextualised AI-enabled content automation, the scope for business process transformation and new efficiencies grows considerably – for instance through the circumvention of protracted processes based on reliable new account insights. An invoice from a recognised supplier which is identifyied as tallying with a purchase order on the system, for instance, could now trigger payment without the need for manager approval, saving time and cost while giving loyal suppliers a better experience.


There is another important consideration in all of this, and that is building in the scope to keep adding more value over time, as AI technology develops.

Adopting a single-application AI is less likely to provide that path. A software application with ‘built-in AI’ will be locked into a specific AI framework such as Google TensorFlow, Microsoft Azure Cognitive Services, a Python-based framework, or linked to specific capabilities for pattern matching/image recognition or for natural language processing (BERT, ERNIE, etc.).

Hard-wiring in one tech approach is a risky move given how quickly technology advances and changes. A better approach would be to adopt an open content architecture, which supports any combination of current and future AI options, on a ‘composable’ basis. In this scenario, companies over time will be able to continue to connect intelligent systems in different ways, that do not depend on the specific built-in AI of individual applications.

None of this is to say that the role of and value added by human teams has in any way diminished. But given the extent to which the Great Resignation and hybrid working have heaped the pressure on roles and skills retention, it follows that smarter automation offers significant value in enabling each team member to truly excel, supported by richer insights that have been pulled together, automatically and in a timely fashion, by cross-enterprise, AI-enabled content services.

About the author

Dr. John Bates is SER Group’s big-hitting new CEO, a tech visionary, automation expert and experienced CxO with a PhD in computer engineering from Cambridge University. SER Group is a leader in Intelligent Information Management, headquartered in Bonn, Germany.

Customer Service