Making sense of AI in Sourcing at Keelvar Konnect and DPW

‍

This week's blog is from Diarmuid Cahalane, Data Science Manager at Keelvar:

Keelvar was in Amsterdam for Keelvar Konnect and DPW recently. We launched our new Autonomous Sourcing Bot Builder, and I discussed AI and automation’s emerging roles in sourcing and procurement. Here are my reflections on the week that was and the road ahead.

Collins Dictionary has just named “AI” the Word of the Year and wrote, “Considered to be the next great technological revolution, AI has seen rapid development and has been much talked about in 2023.” Whatever about the rest of 2023, that was undoubtedly true for the week of Keelvar Konnect and DPW in Amsterdam this October!

Starting with automation, a highlight of the week was the unveiling at Keelvar Konnect of Keelvar’s new Bot Builder. This no-code, drag-and-drop interface hands the baton (or mouse) to end users and empowers them to orchestrate and launch bespoke, automated procurement workflows with just a few clicks.

I’m always excited about tools that amplify people’s knowledge and ingenuity, and Bot Builder is a prime example of making powerful technology seamless for non-tech professionals to use. Unsurprisingly, Bot Builder demos wowed attendees at both KK and DPW.

"Bot Builder is a prime example of making powerful technology seamless for non-tech professionals to use"

‍

‍

To help attendees make sense of the “news and the noise” around AI capabilities and applications, I presented a session called “Demystifying AI” at Konnect. This offered three essential concepts for sizing up the potential value of applying AI in the sourcing process.

1. Some applications of AI will have a much higher ROI than others:

Conisder the steps in the sourcing process, some can be automated with (or without) AI speeding things up, but achieving the same outcomes as before.

By contrast, for other “core value” steps, applying AI will transform their outcomes and value.

Examples of the latter, as available in Keelvar today, include supplier discovery (by processing more data than any human could to find new, relevant suppliers and hence driving more bidder engagement) and scenario optimization (crunching massive numbers of possibilities to find the optimal assignment that best profits both buyers and vendors).

2. AI is a broad toolbox with many tools to choose from for a given application:

Much-discussed Generative AI will have a role in procurement, but not all AI in procurement will (nor should) be Gen AI.

"AI is a broad toolbox with many tools to choose from for a given application"

In fact, some older AI techniques have yet to shine, given that the data flows and integrations to fuel them are just now becoming commonplace. Many high-ROI, “core value” steps are better suited to the steady hand of these mature methods.

More peripheral tasks (like document preparation and vendor communications) could benefit from the creativity of Generative AI, but keep in mind the ROI may not be transformative in such applications.

There’s a potentially interesting intersection, though, wherein giving people GenAI-powered interfaces might make other (core) technologies more approachable and usable too.

3. AI is not always superior to humans (or to humans+AI):

‍
For tasks like the final stages of award negotiations, a human understanding of the nuances and full set of factors involved remains likely to exceed what AI tools could comprehend and leverage in negotiations. Other tasks such as analysis can meld the human’s intuition for what dimensions and questions to probe with the AI’s ability to crunch the scenarios.

For tasks like documentation and communication, the boundary is becoming more blurred, and it is likely to shift in AI’s favor over time.

In summary, when it comes to AI in sourcing, there are some questions to ask yourself:

• To what specific task within the process is it being applied?
• What type of AI is being applied?
• Why that type?
• What is the expected value-add over human performance?

This is how we prioritize AI R&D at Keelvar.

Looking to the road ahead, analysts have suggested, and we agree, that there is no shortage of potential business use cases for AI, but AI’s “next act” must be to prove its real and sustaining value in said use cases.

Our commitment is to focus on applications with potentially transformative impact for our customers, prove that value, and discover how to bring it to life.‍

See our new bot builder in action - request a demo today