In the past decade, Software Robots (aka Bots) have been rapidly influencing procurement processes, making them more streamlined and quicker. While allowing humans to work smarter on human-centric strategic processes, such as planning, innovation, or relationship management, various cognitive systems are steadily taking over approx. half of the organization’s actual routine activities.
Difference between Bots, RPA, NLP, ML & AI?
A Bot (short for Software Robot) is a program created to perform a task that usually involves interaction (e.g. verifying an Invoice, or placing a Purchase Order). Bots can be programmed to do almost anything where information is digitalized. Their biggest upside is the they can be applied across a couple of system environments. RPA is a type of software that imitates a task-related activity of a human within a process (e.g. log-in, copy/paste, field entry etc.). NLP represents algorithms that can interpret, transform and generate human language. Machine Learning (ML) is more complex and its goal is to acquire knowledge (e.g. image/speech recognition, medical diagnosis, etc.). On the other hand, AI is driven by data, and it actually aims to become more intelligent, to simulate intelligence.
What are Chatbots?
These types of bots are designed for “human” interaction. Their main task is to handle an initial conversation with a human, and then forward their input to the back-end in order to be accomplished.
There are different versions of chat-bots:
Menu-based: Most common type of Chat-bots available, used in the form of menus or buttons. Most of the time, they cover FAQ sections.
Rule-based (linguistic): Predetermined by the language, wording, synonyms, and various conditions, their purpose is to create a scenario closest to a conversation. These chatbots are guided by a decision scheme where the user is given a set of predefined options/questions that will take him to the answer. Currently, it is limited to typos and wrong keywords. The most famous one was FB messenger.
Keyboard recognition: This type of chat-bots can follow your typing in order to respond correctly. They utilize NLP to decide on the actual response to be sent to the user. Their biggest downside is the utilization of similar sentences and wording.
Self-learning: By using ML and AI, these contextual chat-bots can remember your previous conversation and learn from it over time. They can automate more questions while requiring less data. The most common example is buying something on a daily basis, like ordering the food.
Significance of Bots in Procurement
Bots can have a substantial impact on all Procurement areas and business processes in general.
(a) Strategic Procurement Benefits
- Better resource management
- Enhancing performance by simplifying it
- Saving money that can be redirected on hiring staff for more important roles
(b) Tactical Procurement Benefits
- Better purchasing decisions (procurement-related questions, information, guided buying, creating tickets)
- Empowering experts to work smarter, by providing them with volumes of accurate intel
(c) Operational Procurement Benefits
- Operational support with available knowledge
- 24/7 instant support (users, interactions, processes)
- Cost reduction (covering volumes of boring and repetitive tasks)
- Improving efficiency with tailored & context-aware interactions
- Improving reliability, by taking away cognitive bias and human errors from the workflow
- Reducing time for each task cycle
- FAQ sections
- Easier onboarding
- Assistance in different jurisdictions with the language benefit
Some of the risks with bot implementation include several reasons why they are introduced in the first place, such as:
- Removing cognitive bias, that can be helpful for speed and consistency, but also offers drawbacks in specific requirements that are not predetermined (wording errors in voice-based conversations, specific local issues, or additional requirements that demand human input – the conversational paradox).
- The necessity to having a real person to take over at your disposal
- Errors that can be heavily multiplied in a short period of time if not monitored correctly.
- Gathered information and knowledge that can become too centric.
- Relationship management, where lack of human input can interfere “against” user’s intention, damaging the overall relationship.
- Level of professionalism, user acceptance testing (validating end-to-end business flow), and the importance of quality guidance for the organization during the actual implementation.
Overall, bots have a great future in Procurement and strategic sourcing in general, and the mass adoption is already here. Their personalization, availability, reliability, and cost-efficiency are something that can not be ignored, as their advantages definitely outweigh the disadvantages. All we need is to determine the exact areas where we can implement them.