The Future of AI in Procurement:
Quick wins using the power of passive AI
Generative artificial intelligence (AI) is revolutionizing the nature of work across almost every industry, around the globe. We see it in creative work, where large language and image AI models generate automated content. In manufacturing, generative AI augments human-based product design efforts, optimizing key processes, and improving quality control. For example, it’s changing how healthcare is delivered, improving patient outcomes through better diagnosis and information sharing.
To better understand AI’s role in transforming the procurement landscape, ProcureAbility and our AI partner, dSilo, present our co-authored Insights series, “The Future of AI in Procurement.” In the first installment of this series, “The transformative impact of generative AI,” we discussed how generative AI is creating new value in procurement. In this second blog, “Quick wins driven by the power of passive AI,” we’ll explain more about the “passive” AI use cases (where the technology is presenting insights but not acting on them) that can drive immediate benefits.
Deploying the capabilities of AI is a journey that looks different depending on where the organization starts. But we have identified trends that are already emerging. Not surprisingly, the focus is on EBITDA improvement, but not exclusively. Buyers are looking for a quantifiable ROI out of the box, especially in North America. In Europe, AI adopters are navigating a balance between cost savings and risk management driven by regulations.
With this focus on ROI, our customers are geared towards near-term savings and other benefits, which means reconciling unstructured information and generating new insights that can be executed by the category leads, buyers, and A/P clerks. Customers are discussing use cases that will employ the generative capabilities of our platform, but they see that as the next phase of development and deployment after realizing quick-hit successes.
In this blog, we’ll explain more about the “passive” AI use cases (where the technology is presenting insights but not acting on them) that can drive immediate benefits. In our next installment, we’ll talk more about using the generative AI capabilities to create work products.
These are several use cases where AI is already having an impact on procurement processes – filling in the gaps in legacy systems and processes and reconciling data at a speed and scale that would be impossible without this ground-breaking technology.
Spend analytics
Most legacy spend tools are using payments data to provide visibility into spend. This has the advantage of accuracy as the payables data comes from the system of records.
This is a good start, and AI can go several steps further to deliver deeper insights. For example, AI can leverage line-item level information from the invoices and other payment documents such as credit card statements, providing visibility at the lowest level and enabling more granular categorization of the spend. With this detailed information, it’s possible to separate out charges such as freight, taxes, duties, and overhead fees. Coupling proven traditional spend analysis methodologies with the power of AI elevates the value of spend analytics to achieve full spectrum data visibility and spotlight opportunities for savings and process improvements.
Furthermore, AI enhances purchase price variance analysis for the same items across different suppliers, locations, and seasons. For example, one of dSilo’s customers, which operates over 340 hospitality facilities across the US, can compare costs for common items such as linens to see the best prices available. Another example is where one of ProcureAbility’s clients is using a static AI to assess spend/usage patterns from the last five years to produce recommendations such as quarter-end, volume-based discounts, or long-term capacity lock ins. This information is then fed into the quarterly strategic sourcing and contracting pipeline to capture value.
Category management
The detailed spend analytics powered by AI also enables a deeper understanding of what is being purchased and from whom. The category manager can now see the quantities bought at the line-item level, in preparation for negotiations with a supplier. This can help identify areas of maverick spend, such as in the case of items that are being purchased from someone other than the preferred supplier. A company might have a contract with negotiated pricing for PCs and peripherals from one supplier and a contract for office supplies with another supplier, but users are buying peripherals from the office supplies company and paying list price rather than the negotiated discount. AI can do this by reconciling items on an invoice, to the respective contract and flagging where the documents don’t align; without the technology the best alternative would be manual spot audits.
One of ProcureAbility’s clients is using static AI to review all past category plans, spend profiles, sourcing and contracting databases to have a recommended category plan that gets refined based on the stakeholder interviews. Historically, this approach was imperfect science where category managers had to spend a significant amount of time to analyze and synthesize information before putting the category plan together and now with AI, they are able to spend time on gathering inputs and doing what-if analyses vs. synthesizing data.
Contract management
Typically, executed contracts sit in a repository that provides access, but analytics is only possible on the limited set of meta data fields that are captured when the contract becomes active.
AI has the power to extract many more meta data fields, making it possible to analyze contracts in aggregate, as well as individually. dSilo has 140 data elements they’ve extracted for various customers depending on their business needs, including for a European customer that wants to understand which of their contracts already include a GDPR clause and which do not. Once their contracts are ingested onto the platform, dSilo can report this information back, sorted by filters such as geography, category, or type of agreement.
One of ProcureAbility’s clients is using AI in clause library management of Contract Lifecycle where the AI recommends the fallback positions based on the seventeen input fields and historical approach and gold standard approach. The category manager then reviews the options and decides the best possible language within the clause library. The AI continues to learn as the category manager makes decisions in these clauses and fine tunes language for future. AI has changed the way category managers manage risk reviews and acceptable variations with the language while finalizing the contract.
Invoice processing
Another even more powerful passive application of AI is in invoice processing. If contracts and invoices are uploaded to the AI platform, it’s possible to reconcile between the documents at scale. This enables purchase price verification, including from complex rate tables. A customer receives monthly invoices for professional services that can contain hundreds of line items, and the platform is able to extract the rate tables and reconcile to the details in the invoice, flagging any discrepancies as part of the invoice approval workflow. Such benefits go a long way toward improving accuracy and productivity for the A/P team.
AI can also verify payment terms – we find about one percent of spend where invoices are paid before the contracted due date, often because the invoice states different terms, and the payment system is updated but not fully reconciled to the contract. Being solved for by AI, it can sit in the invoice approval workflow rather than coming through the after-the-fact audit, eliminating the need for follow ups with the supplier and accounting for credits.
Undoubtedly your organization has its own gaps and opportunities to create new value by closing these holes with AI. Start to collaborate with your colleagues and a trusted third-party partner who is experienced in planning out the digital roadmap for procurement, to identify where an AI solution can replicate human processes and do it at scale.
The emphasis here isn’t on hitting homeruns but rather singles and doubles in bunches. If you’re able to do that, you will have implemented AI with a strong tangible ROI that will also allow your team to focus on more complex issues.
In our next blog of this Insights series, “The Future of AI in Procurement: Transforming theory into practice,” we’ll look at taking the next step and utilizing the generative capabilities of AI to create work products and drive even greater value.
