Transforming tender management with AI – the big data game

Artificial intelligence (AI) has the potential to transform the way we do business. Many organizations are looking at how they can apply AI to their business, but it’s not a straight-forward process. A great deal of forethought and planning is required. Today’s AI works in specific ways, different algorithms can be applied to certain use cases and they must be structured appropriately. This can involve significantly altering the surrounding human processes to ensure that the business outcomes are achieving the highest value. It requires a concerted effort that pulls together business and technical requirements to ensure that solutions are effective and achievable.

Quantity, relevance, classification, formatting and accessibility of data can have significant impacts on the value that AI brings. Many organizations have plenty of data but it’s not right for the desired application. Data scientists spend high proportions of their time on data wrangling or data munging which involves pulling data together and getting them setup for running through AI algorithms.

Tender management

By its nature tendering is designed to be transparent. The process should be open, fair and free from bribery and nepotism. Applicants to the process can challenge results and the process will be audited to ensure that no corruption occurred. As such all parties receive similar information and the results of the tender are made public. This should be a great source of data with plenty of opportunities for applying machine learning to give tender teams recommendations for which tenders they should be bidding for, propensity to win, recommended price point, etc.

One of these portals, the EU’s Tenders Electronic Daily (TED) has over half a million tender notices each year. Unfortunately, these data are published in different media, formats and locations depending on the country, region, tender value, etc. Some regions have modern online systems for publishing, submitting, and online databases of contract award notices. Others require submissions to be sent in paper copy and the award notices are announced verbally at a public event. Having a good tender management system that facilitates the capture of these data in a consistent format and allows it to relate to the company's opportunities, configurations, pricing, dossiers, billing, invoices, rebates, and accruals (to name a few) enables a consistent and effective AI.

AI algorithms, tasks, and applications

Adding AI successfully to the process requires good analysis of the situation. A first step is to decide what applications would bring the greatest value to the organization. Some applications are:

• White Space Analysis
• Competitive Intelligence
• Tender Screening and Propensity to Win
• Pricing Intelligence
• Solution Guidance
• Rebate Program Recommendations
• Bid Response Recommendations
• Demand Planning and Jeopardy Management
• Accruals and Revenue Recognition Forecasting
• Bots

There are a number of AI tasks that exist that can be used to achieve these applications. Sometimes more than one of these tasks are needed to achieve an application:

• Regression
• Classification
• Anomaly Detection
• Clustering

Finally, there is a long list of AI algorithms that can be used for these tasks. Some of these algorithms can be used across multiple tasks. A few examples are:

• Linear Regression
• Logistic Regression
• Neural Networks
• Support Vector Machine
• Principle Component Analysis
• K-means Nearest Neighbor

Generally, the sophistication of the implementation influences the value that AI provides to the business. However, it is usually best to start with a system of low sophistication where it is easier to understand and changes are easier to make. Over time, as the AI is performing as expected, the sophistication can be iterated up to increase the value brought to the business.

The three pillars; technology, processes, and people

The previous paragraph identified the differing functional facets from the business requirements (applications) through to the technical requirements (algorithms and data). As with setting up or improving other business processes there are always the three pillars of technology, process and people to consider. An endeavor will fail if the business sponsor thinks they can ask the technical team to start implementing AI without first understanding it themselves. In the same way that the technical team will need to learn about how to structure data in the organization's systems, the business users need to understand the capabilities of AI and how it will change business processes. Businesses are systems of systems and changing the processes of one system affects the operation of other systems. The users involved with these changes will need training otherwise the applications will not be fully adopted. The people pillar is a regularly overlooked pillar.

Pricing intelligence

Let’s take the application of pricing intelligence. Pricing analytics alone is complex. It varies by industry and sales channel. There are whole university degree programs that teach pricing analytics. A report by Accenture suggests that integrating AI into pricing could move revenue and profit by high double-digit percentages. Pricing managers could have developed their business requirements effectively, redesigned processes appropriately and worked closely with the technical teams to implement correctly.

But the business may not be seeing the returns it was expecting because the sales executives are not engaging, they are still doing things the old way, they just don’t understand why the system is indicating prices or how to even engage with the system properly. It is important to bring the sales executives into the design process, to ensure they can, and want to use the system, and they understand concepts such as contribution margin that are driving the systems. They should be taught how to read a price waterfall to understand how the changes they make to structuring the deal affects profitability for the business.

Check out this article launched as part of the European Pricing Platform’s thought leadership series. It provides insights and guidance for applying AI to the middle office process of tender management.

About the European Pricing Platform (EPP)

The European Pricing Platform (EPP) is, since 2006, on the front of supporting pricing and monetization leaders with on- and offline knowledge sharing initiatives and events. If you are looking to improve your organizational pricing maturity, need help bringing your pricing journey to the next level, or if you are looking to learn how to better champion your next price management initiative, EPP is here to help.

From the start, the EPP is designed around three important principles:

• not-for-profit
• give-and-take
• open, neutral and unbiased