AI and Sourcing – What Might the Future Process Look Like?

It is still not clear exactly how certain new technology will affect the world of procurement. For instance, while some blockchain applications are developing, and there is little doubt that it will affect the world of finance and payments, and be useful in tracking high-value items in the supply chain (like diamonds) it’s less obvious how it might affect the day-to-day work of procurement.

Other issues are emerging too; for instance, in the shipping industry, Shipping Watch website reported;   “Maersk's ambitions of developing a new technical platform built on blockchain technology, developed with IBM, for the container sector overall gets a muted reception from two of the liner company's closest rivals, German Hapag-Lloyd and France's CMA CGM”.

Questions about the merits of joint industry standards as opposed to proprietary solutions are likely to be prevalent in a number of blockchain application areas – just as they have arisen in electronic invoicing, for instance. However, when it comes to the role of machine learning and artificial intelligence in the sourcing process, it feels like we are getting closer to a clear vision of what the future might look like.

When I did my session recently for the University of Birmingham MBA students recently, we briefly discussed what an AI-informed sourcing process might look like. This isn’t a complete list by any means, but you could imagine “AI interventions” at different key stages in the CatMan / sourcing process, including maybe:

Initial brief and strategy – AI-informed analysis of whether you even need to do this sourcing exercise! Automated contact with existing supplier(s) to produce an initial offer to extend the contract or automated “negotiation” with current suppliers?

Understand requirements – gather specifications, look to build overall requirement – AI identifies where specs are out of line with market, suggests more cost-effective alternatives, looks for commonalities (both within the organisation and maybe even outside – networked group buying opportunities).

Understand market – research, market engagement – AI pulls together the market analysis, suggests suppliers based on performance ratings, other data, includes CSR and risk information etc. Analyses track record of potential suppliers, with both the buying firm and perhaps others.

Develop sourcing / category strategy – AI draws on best practice and experience to suggest optimum length of contract, analyses risk factors to suggest optimum number of suppliers, and suggests appropriate commercial models based on direct experience of organisation and other information from the wider network.

Supplier selection – automated sourcing (RFP etc.) process, AI sift of potential bidders, evaluates the offers, even goes back and “negotiates” (e.g. this supplier has best quality but is 10% more expensive… AI learns that the supplier may well reduce their price with some “persuasion”).

Contracting and negotiation – AI builds a model contract, or provides “guided” contract clause options and management, automated “negotiation” and agreement of key terms with supplier …

Now I’m not suggesting for a  moment that for more complex contracts this can be done in the foreseeable future with no human involvement at all. But for many more basic sourcing exercises … it doesn’t look too far off today, really. And even for those more complex services, the role of the procurement expert will be quite different from today. So we need to start thinking about what our human brains, creativity and experience can bring to bear if the machines are doing so much of what we currently do.

The other point – which we’ve made before – is whether it will be a “procurement function” bringing the human input to the party, even where that is needed. Or, in many cases, will it be someone from “the business”  budget-holding area who, with the help of the technology, can perfectly well manage even a pretty complex sourcing process?

Again, I’d like to think there still will be certain skills that best reside in “procurement” – but we better be clear ourselves exactly what they are if we are going to persuade our colleagues and bosses that we can still bring value in an AI-populated business world!

First Voice

  1. Dan:

    Knowing when not to blindly trust the AI would be a very valuable skill

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.