Robotization and AI : the best of the two worlds for gaining in performance
Artificial intelligence

Robotization and AI : the best of the two worlds for gaining in performance

From simple robotization for automating low added value tasks to AI capable of furnishing predictable analyses, a field of possibilities exists for making gaining in performance Purchases and Financial functions inside companies.

Robotization, AI… new technologies are exploding inside companies and have only one goal : making gaining human time by automating low added value tasks and thus increasing productivity. Thus, accorded to Forrester’s cabinet’s survey, 58% of companies note that automation will create added value for Purchases and Finance’s transformation.

At Axys Consultants, we talk about “IPA (or intelligent process automation), technologies that rely on both RPA (robotic process automation) and AI (artificial intelligence) in order to develop Purchases and Finance’s performance”, explains Jeroon Pedurupillai, Director of the Purchases and Innovation business unit at Axys Consultants*.

From fraud detection to chatbots

If RPA (or simple robotization) can be used for fraud detection, erroneous or falsified invoices, many other uses cases exist. “Conversational robots capbale of get hold of system for validating a request or of intervention into the workflow exist”,enthuses Jeroon Pedurupillai from Axys Consultants. And to quote, all possibilities that IPA (coupling RPA and AI) offers: “the affectation of validators to invoices, the updating of suppliers’ datum with especially control of Siren number… Here again, we are not remplacing Human, but human’s action without added value. It is a sort of work hybridization between human and robot.

In the case of year-end discount rebates to suppliers: “Used during all the year, said “intelligent” robot permits to offload the buyer from the spendings’ analysis and to identify suppliers who are submitted to year-end discount to specifiers in order to maximize associated gains”, details Jeroon Perudupillai from Axys Consultants, who estimates a gain of time of three hours per supplier for the spendings’ analysis, claim and following of year-end discounts.

From task automation to predictable anaylsis

Accorded to Jeroon Pedurupillai, we can distinguish three robotization levels : simple, cognitive and intelligent ones. If cognitive robotization is a more thorough robotization with a little bit of AI, intelligent robotization is totally piloted by AI.

  • Simple robotization: basic processes automation with RPA (robotic process automation) type solutions. It is about coded software’s robots which work with structured datum and standardized rules, and which contains low exceptions. In that case, softwares robots imitate human’s action with applications. These solutions are easy to implement with a relatively low investment.
  • Cognitive robotization : At a superior level, automation of more complex processes exists with solutions of cognitive RPA. In that case, AI’s contribution with for example machine learning coupled with RPA permits to automize more complex processes with semi-structured (invoices, contracts, purchase order, etc) and/or unstructured datum (voice messages, emails, etc). Those softwares robots analyse the human’s action in order to capture knowledge of process and to understand its operational mode.
  • Intelligent robotization: Finally, at final stage, the automation piloted by AI with intelligent robot-type solutions, automation is piloted by auto-learning and permits to make predictable and advanced analyses. If those solutions permit to “low human’s error, they do not disclaim human’s factor into the decision-making”, specifies Axys Consultants’ Innovation and Purchases unit’s Director. “We can imagine a use of this technology with negotiator robots that relieve the buyer in some stages of the negotiation or as part of the call for tenders without added value”. This is an expensive and long-to-implement technology.
    “It is about well defining KPIs when framing of the project: with the researched performance, estimation of processed volume, ratio of expected productivity, ratio of to-be-eliminated errors, or ratio of tasks reallocation to collaborators”, concludes Fabien Richard, associate Director of the Financial and Purchases Practices at Axys Consultants.

*”RPA/IPA’s contrubutions to Purchases and Finances functions”, webinar organized as part of Rethink20 event from Poractis France.

Article published in “Décision Achats