Against the AI Robots Taking Over Procurement by 2020

An Argument Against the AI Robots Taking Over Procurement By 2020 (Presented in London, 13 July 2017)

Author's note: I originally presented this essay as the opposing argument in a "Pub Debate" following the Oxford/Cambridge Union debating structure in London on 13 July 2017. The motion, which I was instructed to argue against, was that "The Robots" would run procurement by 2020. Peter Smith , the debate organizer and the gentleman arguing in favor of the motion, can share more details, as I was likely a pint or two of bitter into the festivities by the time I presented these words, and all I remember from the evening is winning (against the motion) while adding some additional off color jokes into the actual delivery to further take the piss from Mr. Smith.

Enjoy!

When I think of robots, what comes to mind? For me, it’s the good guys. Robots, after all, can do no harm. It’s Michael Jackson in the movie, Moonwalker. You know, when the junior member of the Jackson Five transforms into a robot and goes on a dark-side killing rampage.

May the moonwalker rest in peace.

But robots aren’t all good.

Here are examples of a number of bad robots that come to mind.

Who can forget HAL in A Space Odyssey?

Or what about the fat, squat shooting robot ED-209 in RoboCop.

I was somewhat partial to Vanessa Kensington in Austin Powers, until she took off her human mask. But I’ll leave it at that.

There are truly creepy robots too: Ash (from Alien) and the creepy-crawly sentinels from the Matrix.

And who can forget Auto from Wall-E – probably the best example of gently trained AI gone wrong.

Finally, I would be remiss, given the venue, if I did not mention the androids in The World’s End, a movie as metaphor for the death of the independent pub and real ale in this wonderful country – at least how I read into it.?

But now, let me tie this back around to the motion: “This House Believes that Robots will Run (and Rule) Procurement by 2020”.?

I believe that the general direction of this argument is not in and of itself wrong. But there are a number of flaws in the nuance of how the motion has been proposed. And we are, after all, asking you to judge the merits of the proposal on its own, as it stands.

Let me present you with three arguments against it.

First, I have a serious issue with the timing. 2020 is not happening.

Artificial Intelligence, the central component of robotics as we define it, is not ready for prime today outside of highly targeted applications. We’ll come back to this in a couple of minutes with some real facts, but there’s no way robots are running purchasing if in 2017, AI is a novelty of a novelty.

Second, procurement is not purchasing.

Transactional purchasing can (and should be automated) to the greatest degree possible. It already has been in multiple places:

• Multi-way matches in P2P to avoid unnecessary paper and human oversight
• Guided buying approaches in E-Procurement that cut out “buyers” from the process of shopping and approvals and guide users to the best decision for the business while making them feel good about what they’re doing
• And of course the overall digitization of all sorts of back-office functions such as AP ?

People are bad when they touch these areas. But even here, we’re not talking about robots. We’re talking about basic cloud procurement apps 101.

Third, let me argue that robots lack empathy. And procurement (not purchasing) is about putting yourself in the shoes of everyone else in the business and in our supply chains.

Ladies and gentlemen, as I just noted, our jobs are in procurement today, not purchasing. But please allow me the chance to expand on this argument and why it matters. In short, a transactional orientation to buying is only part of the world we live in.

Let me give you some examples …

Do you honestly think a robot, even in the next decade, will be able to …

• Convince the business to work with a secondary key potential supplier in a category based on a small (and accurate) set of old KPI data that shows the vendor did not perform as some expected at the time
• Get an internal working group together or consortia of like companies to agree on data collection and validation standards for the EU General Data Protection Regulation?(GDPR)
• Charm a CMO to consider a new approach to bidding out and managing agencies of record?
• Agree on what savings really should count with finance?
• Get to the bottom of an accounting treatment for a rebate in a supply chain finance program (which could push it over the edge or not) in different geographies?
• Work with developing local suppliers in emerging markets – either on the ground (already) or bringing them into a region
• Deal with the sensitivities of post-merger synergies in an acquisition situation bringing two procurement organizations and two supply bases together

Ask yourself these and similar question and I suspect you’ll see arrive at the same conclusion I have.

Now, let me return to a prior point.

If what I have said is not enough to convince you to vote no, let me argue the robots are not ready for prime time. Here are three examples:

1)???In spend analysis, 95+% of the deployments in the market today still rely on rules-based cleansing models that work. And even AI-based models still require humans behind the scenes to do QA on the data once a system is trained in a given area. There’s speed advantages, yes, but only if the model is trained correctly. And believe me, you don’t want to go first. ?

2)???AI simply doesn’t exist in P2P yet. We’ve been shown demonstrations of AI by more than one vendor that simply did not work in practice as planned. No harm – it’s beta. But when you ask for supplier guidance in one area and Siri says “let me think about that” you quickly lose your audience in a demo

3)???Look at those scrambling to embed robotic process automation. It’s those firms that have already chased labor cost arbitrage as far as they can. Now they’re looking to replace bodies with bots. If the “push” in the market comes from BPO firms we know they’re just trying to save their margins. True AI and robotics will be a tech play, by technology firms with scale, SAP Ariba as one of them. It won’t be a last ditch margin preserving effort for services oriented firms.

In Conclusion ...

Finally, please consider the first gentlemen making the argument in favor of robots (a one "Mr Smith). He has had to resort to music to win your affection (Editor's note: Mr. Smith played music and danced like a robot during his argument to win the audience's affection).

Robots (true robots) don’t understand music. If they dance, it’s based on linear programming. A bit like Mr Smith in fact …

My esteemed Spend Matters colleague is playing on a particular human emotion to charm you into voting in favor of the motion. And don’t feel sorry for his dancing. It’s why he went into procurement after all.

And in regards to Mr. Marland ( James Marland ), robots do not wear hats, however charming and dapper they can make one participating in a pub debate come across.

So please, focus on the reality of where robotic process automation is today as well as the scope of the argument.

• On a timeline perspective, 2020 is unrealistic
• And on a scope perspective, robots will not yet have empathy to support broader procurement outside of transactional purchasing
• And finally, the kit simply is not ready for prime time. It will be. But it’s not there yet

Thank you for your time. And let us raise a glass not only against the motion, but against Mr. Smith’s two left feet.

?