AI opportunities in procurement
Artificial intelligence (AI) is transforming how organizations operate across almost all major business functions. The sourcing, purchasing and management of goods that form the primary responsibilities of an organization’s procurement function is no exception to this trend.
AI is already helping procurement leaders reduce costs, optimize their processes and make more informed decisions. However, with AI technology advancing rapidly, how can you be sure which use cases best suit your organization?
AI implementation
Our recommended approach
Step 1
Define the AI vision and guiding principles
Step 2
Identify AI use cases and define a strategic blueprint
Step 3
Prioritize pilots and define an AI roadmap
Identifying potential AI use cases within procurement
Not all AI initiatives are equally appropriate for every organization. Before researching the specific applications that AI can have within a function, we recommend that an organization defines key guiding principles and a high-level AI strategy to ensure initiatives remain aligned with overall goals.
With these in place, it is useful to consider the relative ease of implementation compared to the potential business benefit any initiative can deliver. This will help you narrow down the list of appropriate use cases. Considering the specific procurement ‘pain points’ or ‘opportunity areas’ you want to address should help you identify a small number of initiatives that are the right fit for you.
Promising AI developments within procurement
AI has the potential to significantly improve the vendor selection process by both reducing the time it takes to complete and reducing human error and biases in decision-making. By analysing vast quantities of data such as available contracts, reviews and ratings, compliance records and performance metrics, AI can automatically recommend the suppliers that match your business’ selection criteria.
Natural language processing can also automatically generate RFI or RFP templates based on relatively small amounts of input data, significantly reducing the time spent writing these.
AI is already improving contract management in several ways, including automating the contract creation and review process based on predefined templates and organization-specific clauses. This includes scanning existing contracts for errors or clauses that may pose a risk.
AI can even optimize your contract negotiation process by analysing existing market data to determine the best terms and conditions or pricing model your selected vendor will likely accept. Several companies now use AI negotiation coaches to establish the strongest tools for each commercial scenario, e.g. using e-auctions versus full tender.
For organizations with frequent purchasing requests across a broad user base, implementing AI-enabled purchasing catalogues can reap significant rewards. By ‘Amazon-ifying’ purchasing through tailored user experiences, buyers can first see recommended and frequently purchased products, with AI analytics identifying which products resulted in the best outcomes.
AI-enabled chatbots can also assist users with potential purchase requests by offering clear guidance on purchasing recommendations. This application is particularly useful for industrial companies or those with large field services divisions who frequently purchase from the field.
Natural language processing can enhance the accuracy of spend categorization by extracting relevant information from unstructured data sources such as contracts, purchase orders and invoices, automatically assigning categories based on the criteria you require. Most of these tools will apply a ‘confidence level’ to their suggestions, allowing for the time-efficient review of potentially erroneous data. As these algorithms can learn over time, confidence levels will improve as more data is fed into the system, eventually enabling the system to review historic data and re-categorize potentially erroneous previous spend allocations.
A whole host of AI-enabled tooling can identify procurement opportunities which deliver cost savings, improve speed and/or quality, or decrease organizational risk. Some organizations are combining analytics tooling to create ‘AI savings radars’ that identify potential supplier and category-level savings along with actionable recommendations, triggering automated custom communication with suppliers based on their performance levels.
Evaluating feasibility and implementation risks
While the outlined use cases are becoming more popular and sophisticated by the day, feasibility limitations may still hinder deployment and user adoption. Before implementing your use cases, you should always consider any applicable technical and organizational limitations.
Some AI use cases are more straightforward to implement than others, and balancing business benefit with ease of use is critical to ensuring success. For example, while AI-powered, data-driven decision-making is already a reality, AI lacks the human intuition and judgment required for complex decision-making based on more subjective factors.
Common technical feasibility challenges for AI in procurement
Data availability and quality
Any decision-making tool relies on large, diverse data sets that are representative of the area it is attempting to analyse. Procurement data can often be fragmented, inconsistent, or simply false. Fake reviews are a particular threat to procurement decision-making and could easily skew an algorithm’s recommendation. The data cleanse activity required to produce good quality data may often be more challenging than implementing an AI tool itself.
Risk of biased decision-making
AI tools may introduce biases or risks that affect the fairness and transparency of the procurement process. Recommendations for appropriate vendors may be subject to bias based on individual vendor characteristics. It can also be challenging to explain why an algorithm made a particular recommendation, leading to potential audit challenges in future.
Common organizational feasibility challenges for AI in procurement
Resistance from users
Procurement professionals may express scepticism towards AI tools when they already have well-embedded processes and decision-making approaches. Explaining the drivers for change and managing employee expectations throughout the implementation is critical to a successful roll-out.
Aligning to wider AI goals
The AI use case you want to trial may address a specific procurement problem but does not align with the wider organizational AI strategy. This can result in strong opposition from leadership or potentially lead the function down the wrong path.
Lack of internal AI skills
The AI market is constantly changing, so developing the internal technical skills required to run or make changes to your initiative following its deployment can be challenging.
Organizational structure
You may need to consider changes to the overall organizational structure to make best use of AI benefits, e.g. freeing capacity within a role may result in requirements for new role definitions or appropriate changes to team structures.
Calculating the return on AI investment
Aside from potential feasibility factors, AI tooling often requires significant up-front investment. Procurement departments need to carefully evaluate the potential costs versus the benefits and build an appropriate case for change.
It is easy to be swept up in the current excitement over AI, but investment decisions should always be based on solid foundations. We recommend comparing the return on investment with other alternative investments or baseline scenarios to determine which AI tooling is worth pursuing.
Key benefit categories for a potential AI procurement business case
Many of the use cases we have examined come with direct time-saving benefits, such as the automated creation and review of contracts, or the reduced need for human effort over data analytics exercises. This frees procurement professionals from time intensive tasks, allowing them to focus on areas where they can add the most value.
Better supplier-related decision-making can have many benefits but for most organizations, cost will be a primary determining factor. Using the AI market analysis tools already outlined, organizations stand to save money by timing purchases at opportune moments e.g. when market conditions drive down costs for example, or by using AI-assisted negotiation to secure the best value for money contracts.
Understanding the sustainability impact of your procurement decisions can be challenging, but AI can assist by assessing potential human rights issues or environmental risks associated with specific vendors. With a wave of new ESG reporting requirements on the way, it is worth planning for the future and determining how AI tooling may be able to assist as part of your wider case for change.
AI can reduce procurement risks, such as supply chain disruptions, compliance issues, and fraud, using predictive analytics and anomaly detection to anticipate and mitigate potential threats before they happen.
Some AI tools may be reused or adapted by other organizational functions, allowing the investment to go further There is a lot of potential crossover with Legal, for example.
It is worth noting that most major procurement software suppliers are starting to develop AI add-ons to their existing systems. This may offer a more cost effective, and less risk adverse approach than implementing a brand-new tool. It would also help to mitigate the challenge of integrating AI tools into your existing technology landscape which can otherwise require significant effort.
However, ‘out of the box’ solutions may not be what they first appear. The maturity of these products remains in question and organizations may still need to invest in some degree of bespoke configuration and tailoring to ensure effectiveness.
It is worth noting that most major procurement software vendors are starting to develop AI add-ons to their existing systems. This may offer a more cost effective, and less risk adverse approach than implementing a brand-new tool. It would also help to mitigate the challenge of integrating AI tools into your existing technology landscape which can otherwise require significant effort.
A pragmatic approach for AI in procurement
New and existing AI use cases offer significant potential benefits to procurement functions, but a pragmatic approach to assessing and implementing them should always be adopted.
Consider the cost, compatibility, scalability, security and organizational context of your proposed initiatives before committing to significant investment. Trialling different use cases on a small scale e.g. procuring one category of products, or deploying tools within one region, may be an appropriate starting point for leaders looking to embark on their AI journey.
A risk averse approach can also be adopted when implementing your desired use cases. For example, use AI-assisted procurement approaches rather than opting for fully autonomous buyer selection, a prospect which many procurement professionals may be cautious to adopt.
As the technology matures in this space, becoming more accessible and affordable, various other use cases are certain to emerge.
Procurement leaders should make sure they stay informed of the latest developments to maintain a competitive edge and unlock the transformative potential that AI tooling looks set to bring.
The author
