Most of us will be familiar with the pop-up chat bots that often greet us on e-commerce websites: virtual assistants designed to handle very specific interactions with customers and often also to ensure a smooth hand-off from automated messaging flows to a human for more personalized and qualitative assistance. But it’s important to remember that these chat bots are not fueled by optimization, and while sourcing bots -- such as the ones developed by Keelvar -- bear some similarities (both are primarily designed to save people time), they are also very different.
Chat bots are designed to mimic human conversations, and while some can do so in a very convincing and natural-sounding manner, they are limited in the information they can gather and share. Additionally, the needs of businesses can be quite different; there must be clarity and no ambiguity so voice or natural language instructions are not ideal. Precision and speed are more important attributes.
By comparison, AI-powered sourcing bots are essentially agents designed to execute several tasks and reason about inputs they receive, such as identifying suppliers to invite to a bidding event, managing supplier communications and bid analysis, and recommending award decisions. They also get smarter about that work overtime. This allows the human buyer to offload a significant portion of tedious, repetitive task work to the bot so the employee can focus on more strategic projects. These intelligent bots also provide an abstraction layer for complex work, making it ever-easier for more human users to benefit from things like sourcing optimization without needing specialized training.
Bots With Autonomy
Keelvar’s bots are currently considered Level 4 “high automation” on the continuum of automation technology. They can perform multi-phased tasks once initiated by a human. For some of those tasks, a human must approve a decision for the bot before the bot proceeds with additional actions. Overall, the bots operate with a high degree of autonomy but are not (yet) fully self-driving.
Thanks to the AI, sourcing bots at this Level 4 stage can reason about the information being received, suggest intelligent actions that should be performed, and react to new information – all while keeping their interaction with employees streamlined and user-friendly. For this reason, it’s crucial to have a simple interface that allows for seamless user adoption, even if a more complex process lives under the hood.
Also notable is that intelligent sourcing bots can be “category smart.” For example, Keelvar’s Ocean Bot identifies and understands things about ocean freight sourcing that are different from the Air Bot’s understanding of air freight sourcing.
Despite the immediate benefits of using intelligent bots as listed here, some question whether bots can really take on the role of human workers in procurement and do the job as effectively, or whether these AI bots will replace humans altogether. For the first question about effectiveness, bots can ensure consistency of process, reduce errors, and execute tasks to specifications at scale.
With regard to fears of the impact on human jobs: Many tasks require empathy, creativity and strategic thinking that humans are clearly better at than machines, but most personnel working in procurement are deprived of the time needed to engage in these activities such as supplier relationship management, risk management and innovation discovery. By automating tactical buying activities, this time is released for investing in more strategically important work that the machines cannot do on a company’s behalf. This has been the story of human progress since time immemorial. The winners in industry are those that embrace more efficient tools fastest.
Not Just Another RPA Bot, Either.
Keelvar’s approach to our intelligent sourcing automation is different from RPA (robotic process automation), which is built with a high degree of customization to automate predictable, repeatable keystrokes and cursor movements through an existing app’s user interface. Our automation is built within the app to understand the context of the activity that is taking place and can adapt to real-world changes to help guide humans to the best decision possible with speed and precision.
Let’s take the example of the business challenges facing logistics, which include ensuring speed, simplifying decisions, and bringing compliance into higher-frequency spot bidding activities that occur outside of annual strategically negotiated contracts. RPA can help automate things like manual administrative processes, drafting emails and reports, checking inventories and reordering products. But while RPA can automate repetitive, predictable tasks, RPA lacks knowledge representation and reasoning that are essential for automating more complex processes.
By comparison, Keelvar’s AI-powered sourcing bots can establish spot bidding or mini-tender events within a matter of minutes, managing and automating mundane tasks such as inviting carriers, collecting and validating bid data, messaging bid status updates, conducting necessary rate card and lane information lookups, and generating award recommendations. The other thing to remember is that intelligent systems display antifragility and get better and better over time.
In summary, by adopting intelligent bots, companies can boost competition and savings dramatically, improve decision quality, increase process compliance, traceability, and ensure operational continuity with lower risk. We see a future that is not too far away from where the best buyers of key commodities in the world are sourcing bots that are unmatched in their understanding of the markets in which they play.