AI and the practice of Procurement – Contracting and Legal
03/25/2024
Complementing our recent research into AI in procurement solution use cases and our wider theme around AI in general, Spend Matters has been undertaking a series of interviews with senior procurement practitioners to gauge the level of interest in, and infiltration of, AI and generative AI (GenAI) in the practical lives of the procurement organization today.
We recently spoke with Stuart Brock, founder and President of startup iKinetiq Innovation Solutions which helps procurement functions design the right strategy to optimize processes, gain deeper insights, and enable data-driven risk management with a focus on using AI technologies responsibly.
Stuart formed this firm after spending more than 30 years in third-party risk management, procurement, data privacy and compliance for numerous industries including financial services, life sciences and legal. Having practised law at a top national law firm, he went on to serve numerous roles for a global financial services company in its legal, compliance and procurement organizations where he also gained deep experience in contract analytics and contract lifecycle management, which he has used to help his clients remediate some of the most critical issues raised by regulators.
Having led multiple projects to select and deploy technologies and worked both as a practitioner and a consultant, he knows the amount of time and effort procurement puts into finding a practical solution to help them run more efficiently, and that spurred his move to form iKinetiq.
“A common problem,” he tells us, “is the reuse of the original use case’s governance and risk template for another use case that has different requirements, increasing risk. This addition of unintended use cases can snowball and lead to serious compliance gaps. So we help our customers establish appropriate guardrails for governance and compliance to prevent these serious gaps.”
It’s enlightening to talk to a procurement professional who has seen the issues from both sides of the fence. We were interested in his view on GenAI use cases, particularly in the contracting world where we see a lot of uptake by tech vendors.
GenAI use cases in contracts and negotiations
One use case of AI in procurement is abundantly evident in contracting and legal. GenAI can be used to summarize a contract and create a risk profile. This allows laypeople to negotiate lower-risk contract terms while ensuring higher-risk terms are routed to the legal department. While this has been transformative, he sees even more exciting developments from an operational and process standpoint.
“What excites me,” he says, “is the ability of GenAI to ‘farm’ your existing contract portfolio. It can review the wealth of data in your contract repositories and provide insight into how you historically negotiated similar provisions. Thus, GenAI can become your team’s copilot, standardizing how you negotiate contract terms while ensuring risks are managed within your company’s risk framework. That, for me, revolutionizes how we leverage prior negotiations to obtain better financial and risk provisions, resulting in a tremendous competitive edge.
“Historically, there has been value leakage from siloed negotiations of contracts with the same vendor. There is typically no reason to have different contracts with the same vendor. Our goal should be a master agreement containing the overall terms of the vendor relationship with sub agreements limited to the project’s specific terms. Gen AI provides insights into this historical data that can be used to leverage better positions in current and future negotiations.
“GenAI also helps you determine which of your supplier contracts have performed the best and whether you can leverage these performance metrics against similar providers. This is particularly useful where you have multiple vendors providing similar services.”
However, Stuart doesn’t see the biggest use case in actual spend, such as rate card comparisons, or how much time invoice-to-contract takes.
(For actual current use cases, read our analyst team’s survey of generative AI use in the procurement tech market.)
“For the near term, I don’t see GenAI being disruptive in ‘how’ and ‘where’ we spend. But I do see it being transformative in due diligence, in transforming contract negotiation and, more importantly, in transforming vendor management. I believe that in 3-5 years’ time, we won’t be looking at ‘Contract Lifecycle Management,’ we’ll be actually doing ‘relationship management.’ That’s where the industry is going and where we need to be, because that’s where the greater risk is. It’s not just about the paper contract, it’s about the totality of the relationship.”
CPO requirements are changing
Coming from a highly regulated procurement background, Stuart notes that there are more requirements now than ever on the CPO, including reporting requirements. He has seen this in his work with large global companies dealing with multinational regulatory changes and technology silos and with smaller regional firms facing increased regulatory oversight and staffing shortages. The CPOs have seen the performance targets set by their CEO and Board continue to increase. These targets shape the transformation that occurs within the procurement organization. Given the CPO role changes every few years, it is difficult to achieve real transformation over time. “I think GenAI’s biggest impact here is the ability to analyze vast amounts of data that can uncover insights and opportunities that previously were elusive. The resulting paradigm shift will change forever how we think about business transformation and how teams will collaborate to drive that transformation.
“One of the most rewarding moments is when I see the AI-powered analysis of a customer’s documents and it’s an eye opener for them. Even where customers believe they have the best templates, processes, risk management, etc., we see patterns where template provisions are routinely negotiated to the same alternative provisions. This means the procurement and legal teams are wasting valuable time negotiating these terms. It’s this lack of visibility that AI is going to completely transform. We can use AI to find these provisions and create AI powered playbooks that provide fallback provisions to be used within the conditions set out by legal.”
And that brings us nicely onto the topic of security and provenance.
The thorny subject of security
Potential legal conditions could limit current implementation of GenAI. Namely, the need for data security and clear provenance of the algorithm’s output. On that subject, Stuart says: “GenAI is not ready for unsupervised application right now. And the reason for that is simple: we currently do not have consistent standards to monitor output and authenticate the underlying models and data sources. Every day I see output that is false or otherwise not the desired output. That’s an inherent risk in all things GenAI.
“The good news is that legislatures, industry and trade associations are trying to fix this. In the US, we have the National AI Advisory Committee that is tasked with ensuring the US leads in safe, secure and trustworthy AI innovation to harness the opportunities of AI while mitigating its risks. Another example is the newly formed group called RAILS (Responsible AI in Legal Services) which is assembling an AI cross-industry group of leaders (judiciary, corporations, law firms, tech providers, access to justice organizations, etc.) to support the responsible, ethical and safe use of AI to advance the practice of law and delivery of legal services to all.
“Most people using Gen AI experience challenges with the outputs. There are hallucinations where the output contains false or misleading information presented as fact. Also, GenAI is trained on tremendously large data sets. We have to monitor the tools we use because some of that training may have been carried out on unlicensed data protected by copyright or trained on fake or misleading information.
“Another area of concern we’ve witnessed is that companies don’t always know or ask how the tool will use their data. Some tools may anonymize the data and use it for other data processing. For procurement organizations the data in their contracts programs can be their business differentiator providing them a tremendous competitive edge. Companies need to ensure their data is protected and used only as permitted.”
So what does the future hold ?
“I think that organizations that are large enough to use GenAI at the enterprise level will have no interest in having their data mixed with others, even if it is anonymized or otherwise protected. And that will limit the tool in the short term — but not in the long term. It will force the technology providers to develop processes that address the issues of data protection, copyright and intellectual property rights, and authentication. I think this is a direction that all of us should accept and plan accordingly.
“For the next five to ten years, GenAI likely will be used in limited ways owing to all of the concerns mentioned. Plus, the media hype and the legislative reaction will establish additional boundaries.
“But I don’t think this is necessarily a bad thing. It will slow the pace and allow us time to address these concerns. It might even be the thing that actually accelerates us, although it may be difficult to see this from where we’re sitting today.”
Many thanks to Stuart Brock for being part of this practitioner-focused series.
To learn more about how procurement is navigating the cutting edge of AI technology, check out our dedicated AI in Procurement page.
And as always, if you’d like to share thoughts/opinions on any of the above please reach out.
-
-
CLM ANALYTICS07/07/2022
-
-
-
-
-
CLM ANALYTICS07/07/2022
-
-
-