How RPA & AI/ Analytics can empower the Procurement functions (Use cases/ case studies and examples) - whitepaper

The Procurement and Supply Chain functions face at least seven main challenges

1) Disparate data obtained from multiple enterprise-wide systems 2) Various data types: Data on clients, spend, transactions, pricing, suppliers, and contracts from RFPs, POs, spend reviews, contract management, e-catalogues, SRM systems, and expense reports. 3) Every stage of the supply chain has incoming and outgoing data that affects the entire product journey, from upstream inventory planning to downstream demand management.4)  Quality of data: Manual data input in purchasing cycle can lead to inaccuracies. General Ledger coding, classification and item descriptions can all vary across the data. 5) High-level categories: Uncategorized spend at granular level (potentially defined with more focus on finance objectives, which can be suboptimal the purchasing process). Misclassified data due to lack of low-level detail required (Resulting reports are not entirely accurate and can result in less than 100% trust in the quality of output). 6)  Data sources and resulting volumes: Multiple ERPs (disconnected systems and IT support dependence). Multiple supplier and spend data sources and master data. Challenge to consolidate. Consequent data set can be huge — confounding accurate, comprehensive analysis. 7) Large time and resource investments are needed to resolve the above

The role of Procurement is changing and becoming more valuable

Procurement used to have the stigma of being just a cost centre/ back office and even an obstacle to driving value within an organization. Today Procurement has the opportunity to play a very strategic role within the organization. Procurement can help build and add value through operational efficiency and revenue benefits, generate ROI.

New emerging technologies are increasingly enabling procurement it its ability to add value. At the same time the expectations are also raised for Procurement. The ask is to address 100% spend globally, apply real-time marketing intelligence & analytics, continuously deliver savings, track compliance, support sustainability, reconcile each item-level spend against negotiated terms and verified by finance, continuous cost innovation, supplier performance improvements, manage/ mitigate risk, address supply chain disruption, drive overall supplier innovation, promote growth,…)

Procurement organization needs are evolving with clear targets

RPA & Cognitive and Analytics/ AI opportunities can add significant value to Procurement

There are numerous opportunities for RPA & Cognitive as well as analytics/AI to add value to the procurement functions. Below chart show the key functional procurement areas and respective efforts.

The darker the color of field above, the higher the automation potential. Pies indicated the level of involvement of cognitive technologies.

The entire Source to pay process shows key effort of the Procurement functions across the typical work journey end-to-end from identifying vendors managing categories, engaging with vendors, procuring goods & services, receiving & controlling receipt, processing respective invoices, paying and managing respective financials. Robots & Cognitive and Analytics/ AI technologies can add value to every level 2 process of the Source to Pay process.

AI/ Analytics value impact for Source to Contract (S2C) efforts

Below chart shows that AI/ Analytics can provide significant value to all key steps of S2C:

Seven specific value examples of AI/ Analytics to the key phases of Source to Contract:

1) Category Taxonomy Design:  Improving the structure and standardizing the nomenclature of categorization. 2) Data Visibility: Automate consolidation, cleansing and categorization of data. Identify patterns within the spend data. Unlock insights in cost opportunities and decrease spend. Enhance supplier, category and geographical knowledge with category teams who are better empowered in decision-making process. Deeper market intelligence from a wider range of sources. Dramatically advanced supplier risk management with improved visibility of industry news and formerly untapped resources. 3) Spend Categorization and management: Align disparate sources with different naming methodologies to the correct category. Leverage economies of scale. Elevate spend analytics and insights capabilities. 4) Real-time, automated analytics: Identify key themes, anomalies and trends. Extract value. Recommend best potential course of action for given situation. Enable procurement managers to take faster actions on more tasks. 5) Opportunity assessment:  Spend, demand, market and cost-component analytics before, during and after key events (ex: sourcing event). Faster and more targeted sourcing. 6)  Accurate report generation/ dash boarding. 7)  Contract administration, management and drafting: Review contracts at depth and speed. Tag terms and notification of key dates and clauses. Flag noncompliance by both parties to mitigate risk and ensure adherence.

Companies can increase their skill levels and sophistication

Depending on maturity level, companies can move from descriptive and diagnostic analytics (what happened? How much spent/ by whom? Why did it happen? Where did leakage occur?) to predictive (Which contracts have the potential to create business risk? How can we benchmark pricing for complex services that are not exactly comparable?).

More advanced digital leaders have already moved to prescriptive analytics (What is the best sourcing strategy based on x, y, z requirements? If an auction is recommended, which format will be most successful? How can risk be minimized in achieving the best possible savings?).

Top leaders are exploring most advanced analytics (ex: Which RFI questions will give me the best ability to predict bidding behavior? What is the best bidding and negotiation strategy for each supplier? ) 

Better manage the Spend

Organizations have typically only visibility of about 60% of overall spend. Tail of smaller spend is particularly inaccessible. 80% of suppliers are typically not diligently sourced.

Often data is not available (uncategorized spend on not granular level). And it is not productive to spend costly time on analyzing small contract amounts (ex: $10,000). However, technology can automate much of the effort. There is a huge cost reduction/ efficiency opportunity. A systematic approach and use of full technology spectrum helps companies better manage and optimize the spend process:

AI/ Analytics and related technologies help generate additional benefits

1)  Compliance issues can be much better monitored. For example, flagging of a banned supplier that triggers a notification. Or unethical practices (in automatic check with online news) 2)  Identify suspicious behavior of employees, partners and suppliers through anomaly detection. 3) Identify price dispersion/ divergence of supplier charged rates. 4) Better manage working capital: notify buyer of early payments. Advanced payment discounts. Multiple payment terms for a single supplier. 5) Drive uniformity and savings across supply base. 6) Cost regression analysis to pinpoint irregularities (ex: unusual high vendor pricing. Provide lead argument for negotiation). Identify price differentiation / outliers across models. 7)  Identify win/ loss of bids. 8) Deliver automatically RFQ to vendor. Analyze quotes received. 8) Create negotiation script to category manager. Develop arguments for negotiator. (system learns from each approach, gains better understanding of which negotiation techniques are most successful in different conditions, improves its recommendation capability)

Supplier Management

AI/ technologies can draw on all possible data sources for a given supplier (previous scoring in a given category, logged commentary of sourcing manager knowledge, historical reporting, etc.). Crawler technology and NLP can scout supplier websites for key information. Build up comprehensive supplier network, provide summaries and recommendations (including advice on supplier assessments, performance and risk management, compliance).

Contract Management

 According to Harvard Business Review, inefficient contracting causes firms to lose between 5% and 40% of value on a given deal. Volume of contracts, variance in working, increasing level of complexity, compliance levels, expectations of value from across business.

Digital contracting is increasingly common, but still demands manual contract analysis, often too expensive in cost and time to extract all potential value. AI/ NLP can extract key data from documents, crate summaries, compare key terms, evaluate offers and support negotiation strategies.

Sometimes the automation or support of even a small task can generate significant benefit in terms of speed and outcome. Example: Review of NDAs may only take 10 minutes review time, but often the effort is delayed for a week.

P2P Process

RPA can have major impact on the processes/ tasks within P2P process as follow: Data entry, payment, PO creation, data management, vendor master data creation, reporting, order acknowledgement, invoice scanning and settlement. In addition, RPA & Cognitive and AI & Analytics together can add notable value in Category management, Guided Buying, and Help Desk.

Guided buying

"Guided Buying" uses intelligent technology to make back office more efficient. It consists of a series of questions by a system. Answers to the questions determines the appropriate buying channel based on category of goods, quantity, price, intended use, location and other key data built into algorithm. System will the prompt buyer to create a requisition or generate PO automatically. System may also initiate spot buy on sourcing event.

Benefit is also that a requester is engaged prior to PO creation, so that procurement has time to act on the need to buy, rather than reacting to purchase retrospectively. Procurement guidelines can be enforced.

Buyers are able to be compliant and self-serve the purchase of the correct items, using preferred suppliers, leveraging the product lists, suppliers and contract management that procurement has put in place. Such simple web-based or bot-fronted process may eventually become the single point for all purchasing, whether tactical or involving a strategic sourcing event. It creates a better user experience and helps mitigate risk that suppliers deviate excessively form agree or standard products in proposing unnecessary items at high prices . Chatbots can serve as the interface between systems and humans, trigger process automation and AI/ Analytics programs.

Other Support functions:

Reporting: AI/ NLP with RPA can also create reports automatically from Excel files and mathematical calculations.

Use Cases

The following list exemplifies some processes and where RPA & Cognitive as well as AI & Analytics can add value (green: high impact; blue medium impact; red: less impact)

Please notice that the increase of technology capabilities will help companies increasingly explore new use cases, or re-consider use cases that were put on hold.

Specific RPA use cases:

P2P Process: RPA can have major impact on the processes/ tasks as follow: Data entry, payment, PO creation, data management, vendor master data creation, reporting, order acknowledgement, invoice scanning and settlement.

Case Example: Vendor Selection and procurement: A food producer used to have a highly manual process for procuring produce vendors and transport materials. Process before: Employees prepared request for quotation (RFQ) package, communicated with vendors, performed preliminary analysis of vendor documents, evaluated running and credit check, finalized vendor selection. Automated process: Robot automates most of the steps. Only human interaction required for preliminary work (specify the project for sourcing, generate list of potential vendors, engage in face-to-face site visits and negotiation). Result: Food producer improved cycle time by 25-50% and processing time by 15-45%.

Specific AI use cases:

AI/ analytics can have strong impact on upstream S2C processes: Spend classification, opportunity analysis, market intelligence, contract authoring, RFx management, supplier risk mitigation & mgmt. Plus select P2P activities: Guided buy. Help desk.

Case Example: Contract Analytics: Authoring, vetting and analysis of contracts is a many-stage process. NLP technology enables data extraction from contracts (contract expiry, pricing sheet, pricing conditions, SLAs, key legal causes) that may take weeks to assemble. Before: Core team of 25 people spread across three departments (procurement, sales and HR) extracts and tracks data from 225,000 contracts. Each member responsible for extracting and tracking relevant data from 9,000 contracts. 20 minutes review time by contract equates to 3000 hours per review. Average salary of $63/ hour. Annual contract review $4,725,000. Technology solution: Company leverages AI-driven cloud solution and conducts review in half of time. Savings using AI technology: $4,725,000 x 1/2 (average time saved using software) - $404,000 (cost of software) = $1,958,500. Team focuses now on key value activities (build ideal contractual terms with suppliers. Plan negotiation strategies ahead of renewals).

Case Example: Use advanced sourcing optimization to negotiate contract: Supplier of automotive parts leveraged technology to assist in annual contract negotiation for carbon/ stainless steel. Company looked for efficient means to evaluate and award contracts to minimize price and supplier risk; place more spend under management to solidify margins (and improving EBITDA). Also sought to rationalize supply base and reduce supply risk for subcategories. Supply base of 30 companies was fragmented, spread over three locations with too few / too many suppliers across alloys. North American spend was $75 million.

Used eSourcing SaaS backed by sourcing optimization engine and scenario analysis capability (intelligent bid sheet design, RFxs, eAuctions and analytics). Benchmarking software helped understand hidden risks and sourcing opportunities, and instantly identify alternative sources of supply). Used input of metal price forecasting on specific buying behavior to match underlying market trends). Result: Complex bids with more than a dozen suppliers. $2 Million identified and $1.5 million implemented saving, 69% supplier rationalization, streamlined/ faster contract award process, bid optimization.

Conclusion

If you have any questions, please reach out to me personally. You can communicate with me through email at [email protected] or WeChat (ID: alexwsteinberg2 ).

About the author: Alex drives digital transformation, innovation and intelligent automation efforts for the largest companies in China. He has helped build a world-leading RPA & Intelligent Automation practice recognized/ awarded by third parties. He has been some of the first to develop/ support some of the largest automation programs in the industry.

Legal disclaimer: This article represents my personal opinion and does not reflect that of my current/ previous employer(s) or clients. The article intends to increase awareness, understanding and dialog about key issues to serve the industry. It does not present any offer or advice in a legal sense. Markets and technology change quickly and information gets out-of-date. The reader is advised to get individual analysis & consultation.