Enterprises are losing millions of dollars due to flawed and inefficient contract analysis processes, resulting in revenue leakage, non-compliance, higher cost of operations etc. This is particularly true for those who have adopted a traditional approach to contract analysis by leveraging a team of lawyers or supplementing them with template or rule-based software to semi-automate the process. Such approaches have gaping flaws such as inflexibility, inaccuracy, inability to learn from reviewer-corrections, being non-scalable and more. To know more about such challenges, I would urge you to read my earlier blog . In that blog, it is recommended to switch to an effective Artificial Intelligence (AI) based contract analysis tool, which can progressively learn, adapt, accurately predict and extract the key elements of the contract, mitigate risks and more. This blog focuses on how AI Enabled XtractEdge Contract Analysis business solution can help enterprises realize value in their contract-analysis journey.
XtractEdge Contract Analysis
XtractEdge Contract Analysis is an AI enabled business solution, built on XtractEdge, which leverages Machine learning (ML), Semantic Modelling and Deep Learning to automate and transform the process of analysing and reviewing contracts. This application identifies pre-defined legal clauses in contracts, determines contentious clauses, and scores each contract clause for risk in the given context. A front-end conversational interface allows end-users to infer, search, query, and reason with the model using natural language. It uses parallel neural pathways with text and vision-based learning to enhance model prediction.
XtractEdge Contract Analysis Key Capabilities
- One-time training: Its intuitive workbench can be used to quickly train the ML model, using some sample contract documents. Sensitive content within those documents can be masked if needed. Contracts between an organization and its partners or vendors are often bespoke and focussed to their mutual needs, terms and conditions. Traditional template and rule-based applications, which adopt a one-size-fits-all approach, are not adaptable for handling such custom contracts.
- Accurate: Its machine learning models can accurately predict and extract clauses (or Intents), and their underlying elements (or Entities) from contracts.
- Learns from feedback: Its progressive learning ability enables it to automatically learn from reviewer-corrections, whereby new ML models get auto-deployed when a threshold-criteria is met. Thus, its accuracy of prediction progressively increases over time. Most of the other available AI solutions lack the ability to learn from feedback.
- Its extractive, editable summary visually highlights key entities within the intents.
- It enables end-users to query contracts-information in Natural Language
- It provides for straight-through-processing when the confidence level of prediction is high. This further reduces the turnaround time and increases productivity. Most of the other available tools lack this ability.
- Contract Redlining: It can auto-detect discrepancies and redline contracts by highlighting differences in between the contract-elements and their amendments, and between the contracts and the corresponding linked documents, such as between Supplier Contracts and Purchase Orders.
- Risk Management: Its workbench allows configuring different risk-levels for different intents and entities and satisfying different user-configurable criteria. This enables effective clause comparison and real-time risk-scoring at intent, entity, document and portfolio levels. Automated redlining combined with dynamic risk-scoring with associated actionable insights make the contract review process very streamlined and effective.
- Tabular Extraction: It can accurately extract data from tables including tables spanning across pages, tables with merged cells, nested tables etc.
- Its intuitive workbench makes it easy for business users to configure the tool without any IT intervention
- It can identify and classify objects, align their orientation, look into nested objects, identifying fakes. It can extract textual content within identified object with high accuracy
- It has a Bio-inspired architecture, which uses Deep Learning to read and process the contracts in the way humans would, keeping its context and semantics intact. ML algorithms have parallel neural pathways for vision and text processing to enhance prediction accuracy
- It is language independent, and supports contracts in any language
How XtractEdge Contract Analysis is helping enterprises in their contract analysis journey
XtractEdge Contract Analysis has been helping organizations realize significant business values in their contract analysis and review journey, through improved compliance rates, prevention of losses, reduced cost of operations, risk mitigation, elimination of contracts’ inefficiencies, improved staff productivity and employee satisfaction. Some of the actual examples are as follows:
- XtractEdge Contract Analysis automated multi-language contract extraction and querying for one of Europe’s largest multi-national pharmaceutical companies. It ingested labour agreements in French, spanning different topics and entities to create a neural network-based knowledge model. A natural language querying interface enabled functions like payroll processing, new employee on-boarding, HR & those working on employee connect programs to accurately answer queries and compare variations in contract-terms across sites. This resulted in significant increase in productivity, reduction in query-time, harmonization, risk reduction, and compliance with Labour Laws.
- One of the largest multi-national conglomerate companies headquartered in Japan was dealing with a large number of vendor contracts (buyer suppliers) and spending huge effort and time in reviewing these contract documents. The enterprise deployed XtractEdge Contract Analysis, which automated contract-extraction and summarization of business-critical information, comparison of intents and entities across multiple vendor-contracts and contract-versions, identification of risks and risk scoring. This resulted in reduced legal costs and manual effort, early risk identification and mitigation, and enhanced staff productivity.
- XtractEdge Contract Analysis was deployed at a leading software services company with over 11 billion dollars in revenue. It processed Master Service Agreements (MSA), Statement of Work (SOW) in over 1400 different formats and worth several billions of USD. XtractEdge Contract Analysis ensured more than 30,000 person hours of savings every year, arresting of revenue leakage by finding discrepancies between contract-elements across versions and upstream/downstream systems, and improved compliance.
There are many more success stories on how XtractEdge Contract Analysis has been helping enterprises in their contract analysis journey. Its well-defined yet flexible roadmap, and periodic releases signify that its new upcoming features will continue to help enterprises in their value realization and stay ahead of the curve.
Santanu Saha
Product Line Manager, Infosys
More blogs from Santanu Saha >