BLOG 1 “Use cases”: Data accessibility in AI-driven B2B sales

In B2B sales, which require negotiation and customization, both the seller and the customer benefit if the best possible offer can be reached as quickly as possible. In this blog series, we will highlight both technical and business aspects of accessing data for AI analysis that aims to improve the efficiency of the sales process.

The use of artificial intelligence to support sales is being studied in the European research project INNOSALE. Finnish companies Konecranes and Molok are identifying the most potential use cases for AI, while Wapice and LeadDesk are developing their technical platforms to support AI functionalities. Our research partner, VTT, is researching and developing AI models and data processing pathways.

In this blog series, we share the experiences of Finnish partners regarding the challenges and solutions we have encountered when a third party participates in the analysis of proprietary data. Although these experiences are gathered within a research project, they are generalizable to situations where a company seeks to commercially acquire third-party analytics services.

We will start the blog serie (Figure 1) by introducing use cases (part I) that are identified by Finnish companies, followed by a look at stakeholders that govern the data (part II), aspects of technical data wrangling (part III), the confidentiality of trained AI models (part IV). The final part of the series will explore the business benefits of an AI-driven business (part IV). In each blog, companies will share their experiences and lessons learned on the topic. Blogs can be found at https://www.innosale.eu/. Please also join our webinar 29.5.2024 14:00-15:30 Finnish time (13:00-14:40 CET), registration link.

Boosting the sales

In B2B sales, the products traded are often complex, customizable, and expensive compared to consumer products. They also frequently play a significant role in the success and efficiency of the customer’s business. The value of a product extends to its lifecycle services, such as after-sales services, spare parts, and refurbishment. This places great demands on even the most experienced salesperson—to be able to provide the right product to the customer quickly.

To improve the efficiency of a company’s sales activities, it would be useful to reduce negotiation time and the amount of back-and-forth information exchange required between the seller, the customer, and the internal support staff supporting the offer.

Harnessing artificial intelligence

As B2B sales are usually not a “shelf product,” the offer stage involves going through several offers. This is a challenge for both parties: the seller needs to understand the customer’s needs and environment. The customer, on the other hand, has to familiarize him-/herself with the products of each supplier, whose features and prices vary. Negotiations with the customer are often lengthy in order to find the product that best suits their needs and to select the exact characteristics of the product, such as the power source, color, size, etc., in the case of machinery.

Artificial intelligence has clear benefits here; it can help both experienced and new salespeople make more and better quality offers in less time. However, AI requires data to operate. Enterprise resource planning (ERP), customer relationship management (CRM), and service management systems such as IT Service Management (ITSM) platforms contain a wealth of data that can be used. Companies use data from these systems for analysis and forecasting, for example with Microsoft BI or with AI functionalities integrated to these systems. The results are usually company or departmental high level analyses, that provide only little value in each individual sales case. In order to bring the results of these analyses into the day-to-day sales work of individuals, more sophisticated AI and processes supporting sales information flows are needed such as developed in InnoSale project.

Examples of sales situations assisted by AI

The companies involved in the research project have identified a large number of uses for AI in the B2B sales process. These include the concrete support of salespeople during the sales negotiation with useful additional information and offer automation. AI can help sales in a number of ways, including:

USE CASE A: “Optimal offer recommendation” An offer is automatically generated for the customer, suggesting the most appropriate product configuration and price based on previous order history and local regulatory requirements. In addition to the reasoned configuration proposal, the salesperson receives from the AI a list of possible other products or additional services that customers in the same customer segment have ordered in the past. The price is dynamically estimated based on the customer history and segment, as well as the current price of the materials. The salesperson can use the proposal directly or refine it before sending it to the customer.

USE CASE B: “Telemarketing” The discussion with the client is based on a discussion matrix that can be followed during the negotiation, and different paths can be chosen based on the discussion. The AI is used to identify in advance what the customer or customer group values in products based on previous successful sales. In addition, the data helps to identify the right way and time to contact the customer. This is particularly useful for new salespeople or in telemarketing.

USE CASE C: “Product Customization” When a customer asks for product customization, the AI assists the salesperson by finding previous messages, offers, orders, and other internal information that can help the customer respond. In the best case, the company may already have product drawings or references to support sales.

In practice, the salesperson receives the AI-built proposal through the same system they use for their other sales activities. For example, if a salesperson uses his company’s product configurator to fill out a product description based on a customer’s requirements, the AI can enhance that description by suggesting relevant features, pricing, and additional services.

Molok's objectives and perspective

Molok is particularly interested in use case A, which proposes to provide the salesperson with a product configuration that is more suitable for the customer while also considering local waste management and sorting regulations. The AI-compiled proposal will accelerate the bidding process and support new sellers in creating high-quality quotations that meet the customer’s needs. Additionally, it offers insights into products that other similar customers have purchased. The seller retains the ability to modify the proposal generated by the AI.

Further processing of the company’s data will drive technological innovation in the sales process, resulting in faster quoting and a system that assists salespeople in creating quotations based on customer and quotation history.

From the company’s perspective, there are challenges to address. Data quality is one such challenge, stemming from both how the current configurator stores data and the diversity of local waste regulations. Another challenge is data availability, which necessitates enriching one’s own historical data for proper machine learning model development. The first challenge impacts the quality of outcomes and increases machine learning model errors when the AI model lacks high-quality training data. Data availability can also cause delays in the initial schedule if all necessary data is not readily accessible or takes time to obtain.

Looking ahead, future opportunities include leveraging the lessons learned from the project to innovate the B2B sales process for complex products. Additionally, developing the order-delivery process to harness AI capabilities, enhancing understanding of data quality and availability, and ensuring timely access to data are crucial steps.

Authors

Marko Jurvansuu (VTT), Sari Järvinen (VTT Project Manager) and Anna Räty (Molok).