Back to blog

What opportunities and benefits does Artificial Intelligence bring in sourcing & procurement? [Part 2]

As mentioned in our previous blog, new Artificial Intelligence (AI) related business opportunities can be found by considering which tasks could be automated, in what way the business value drivers could be impacted or how the current main pain points could be minimized or better yet, removed.

Erik Kolppanen

Various AI based technologies are currently available, for example: Natural Language Processing (NLP), Speech to Text, Text to Speech, Optical Character Recognition, Machine Learning including Deep Learning (Neural Networks) and Linear and Logistic Regressions, Smart Workflows, Robotics (Process Automation), Image Recognition and Chatbots or Virtual Agents. Within sourcing & procurement, these technologies can help you to create business value-added by automating processes and providing predictive, cognitive and guiding capabilities. Read on and find out how.

16 ways to make AI work for Sourcing & Procurement

AI can enable guided and automated processes for sourcing and procurement experts, suppliers and internal customers from other functions. As a result, the availability of sourcing and procurement experts would no longer be a bottleneck, and guided processes would lead to better compliance and shorter process throughput times as well as a better user and customer experience:

  • A supplier onboarding and resource hiring process using smart workflows and virtual agents or chatbots - guiding the suppliers and resource candidates to provide an adequate amount of information for further assessment purposes. This reduces the time and effort of sourcing experts in supplier introduction and supplier assessment phase, enabling them to focus to the right aspects and candidates.

  • Self-source process in indirect sourcing using smart workflows and virtual agents or chatbots - guiding the end user to carry out sourcing activities on behalf of the sourcing expert in the right manner in desired categories, according to the company policies.

  • Guided purchasing requisition and order process with smart workflows and virtual agents or chatbots - minimizing or even eliminating maverick buying and guiding the end user to carry out the purchasing process end-to-end, according to the company policies.

  • Automated and guided source to pay routine processes for end users and suppliers with smart workflows, virtual agents or chatbots, image recognition, Robotics and Process Automation (RPA) and deep learning algorithms - releasing time for sourcing and procurement to define the game rules for the guided process and letting the process learn continuously. For instance, this could include supplier guided onboarding steps, invoice receiving and control process automation and exception handling procedures (untouched invoices and automatic payment of invoices).

  • Helping to interpret and clean historic transaction and master data quality with Machine Learning algorithms and NLP. The more data you get for training AI, the better the results. There are better possibilities to utilize external data sources and gain better visibility to the suppliers. Poor master data can be used more efficiently for supplier data consolidation.

  • Advanced capabilities, such as Procurement Social Networks with AI, would also enable joint technology and product technology roadmapping and bring closer together customers and suppliers – creating partnerships and joint innovations and possibly bringing new market opportunities and improved services and products.

AI can also provide more advanced spend and performance analytics with outlier and pattern identification, using technologies such as Machine Learning, Deep Learning, Logistic and Linear Regressions, Anomaly Detection as well as unstructured data analytics, such as NLP and Image Recognition:

  • Predicting category and supplier strategy performance based on product roadmaps, historical performance with selected suppliers and contracts and proposing alternative and perhaps a more optimal approach.

  • Spend analytics and patterns and outlier identification in desired dimensions. This could include certain undesirable risk countries and category combinations.

  • Purchase price cost breakdowns and cost analysis and outliers recognition for identifying saving opportunities. There should also be price and cost correlations to the specifications.

  • Design tools using cost breakdown and cost knowledge base for selecting the most optimal product and manufacturing technology. As a rule of thumb, 80% of the costs should be determined during the early phase of the product lifecycle.

  • Material and product quality improvements based on found deviations and patterns as well as correlation to the specifications.

  • Supplier manufacturing process improvement based on found deviations and correlations to the specifications.

  • Contract lifecycle management and compliancy checks against supplier performance. Using NLP for comparing contract terms against supplier performance.

  • Improved supplier lifecycle management by utilizing the supplier performance and assessments' results more efficiently. Predicting the supplier performance in new cases by utilizing data from past performances.

  • Supplier development – improved visibility to the audits and supplier performance correlations.

  • Supplier risk management and follow-up and helping the risk identification, mitigation and minimization activities based on supplier performance data received from inside and outside (including unstructured data).

Finally, most of the Strategic Sourcing and Procurement core solution suite providers (SAP Ariba, Ivalua, GEP, JAGGAER, Zycus, Coupa, and others) are in the middle of extending and embedding AI technologies into their solutions. If your digitalization journey is about to start and your intention is to select such a core Sourcing and Procurement solution suite beside your ERP solution, AI capabilities should be one of the main requirements to be evaluated. The AI capability roadmaps of these solution suite providers should be understood well before concluding and selecting such a core solution suite. If you are not extending your core solution suite, there are also some AI specific solution platforms available that are worthwhile to consider, including Google AI, IBM Watson, Microsoft AI, and Amazon AI. If you have your own AI and Machine Learning experts, recommendable tools to investigate include TensorFlow, R, Python, Caffe, Weka, Keras and CNTK. These should get you off to a good start on your AI journey.

More posts