BLOG 5 ”Business Benefit”: Data accessibility in AI-driven B2B sales

The integration of AI into B2B sales processes holds great promise for reaping business benefits. One potential approach is to use AI-based recommendation system that speeds up the sales process and improves offer quality.

Companies can derive significant advantages from AI-driven B2B sales. For example, more accurate sales predictions, improved understanding of customer behavior, higher success rates in offers, and dynamic pricing. To provide a concrete example, we will showcase how Finnish companies Molok Oy and Wapice Oy expect the results of InnoSale research project to improve their sales processes and products.

The current blog marks the conclusion of our series, where we explored various facets of AI-driven sales, including use cases (Part I), stakeholder roles (Part II), data wrangling (Part III), and the confidentiality of analysis results (Part IV). In May 2024, the InnoSale project will host a webinar presenting the highlights from this blog series. During the webinar, companies will engage in a roundtable discussion to share their perspectives on AI-driven B2B sales. Blogs are published 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.

B2B sales challenge originates from area specific customer needs

Molok is a leading provider of waste management products in Finland and has been extending its market area to other Nordic countries and Europe. The company provides waste containers and accessories (see Figure 1). In waste management business, regulations play a strong role in shaping customer demand. These regulations typically specify certain waste fractions to be collected based on the size and location (urban or countryside) of the households producing the waste. There are also variations in how regulations are implemented locally. For example, indicative colours for specific waste type may differ from an area to another or local contractors use different waste collection vehicles and equipment which must be taken into account when creating an offer.

Figure 1: Molok waste management products.

For successful offer creation, B2B salespeople need to have detailed and timely information on local customer requirements. Inexperienced salespeople may not possess such tacit information, and the process to reach a valid offer requires information to be collected from the customer and from other sources.

AI-based recommendation system for new offers

The envisioned solution is a B2B sales recommendation system that proposes to salespersons a previous offer with similar product configurations based on local customer needs. Salespeople can choose whether to use the recommended product configuration as a basis for a new offer or not. Additionally, they receive reasoning behind the recommendation, such as the number of orders on which it is based.

In the InnoSale project, VTT conducted a research on AI methodologies and data pre-processing, which led to the analysis and subsequent recommendations. Molok provided the use case, and together with Wapice, they supplied data and assisted researchers with industry-specific knowledge. The outcome will be a demonstration of an AI-assisted recommendation system (see Figure 2). This system provides salespersons with typical offers tailored to specific geographical areas, such as a country or a city. In the demonstrator, the recommendations are represented by IDs of previous offers with similar product configurations that resulted to an order. These recommended offers can be visualized to salespersons using the Wapice’s Summium® CPQ sales configurator.

Figure 2: AI assisted offer recommendation demonstration.

Research continues by collecting user feedback from the salespersons using the demonstrator. This includes analysis of user acceptance, accuracy of recommendation, and estimation of improvements in sales process efficiency. Algorithm-related research will include studying trend analysis to identify how customer needs will develop locally. This information can be useful for salespersons during customer discussions as well as for marketing and product portfolio management. A topic for future research is to study whether it is possible to create entirely new product configurations and offers based on analyzed customer needs.

Business benefit

The main benefit of a recommendation system is that an inexperienced salesperson can provide an offer that takes into account local customer needs. It requires less manual work and time to create an offer, resulting in fewer iteration rounds with the customer. Additionally, it reduces the need to consult more experienced salespersons or scout earlier offers from the same region in the CRM.

If the offer quality improves, there will be fewer product returns and associated costs. The higher quality and faster negotiation process may also enhance the win probability of leads. Furthermore, implementing this solution can potentially free up more time from routine work, which can then be allocated to customer interactions.

The solution can be integrated into the existing sales configurator used by Molok’s salespersons. As a result, this new AI-assisted feature seamlessly becomes part of the existing sales process with minimal effort and training. Given its integration, there is an expectation that salespeople will be more willing to use this new feature.

Challenges

Some critical viewpoints can be raised. The expected benefits rely on the assumption that the analysis has been conducted correctly and that the recommendations align with customer needs. There are uncertainties on both fronts. It is also unclear whether salespersons are willing to accept such a new AI tool as part of their work process. If accepted, is there a danger that recommendations will be trusted too much without careful inspection? Will salespeople provide feedback to the system so that the quality of recommendations can be improved, making the system more useful in their work? Finally, what is the responsibility of the party that provides the AI system in the first place? These questions are addressed in later research.

Wapice business perspective

Even though the awareness of and interest towards AI assisted solutions and use cases in the industrial domain is increasing, AI solutions in commercial CPQ (Configure, Price, Quote) tools are still in their early stage. It will most likely be just a matter of time when AI solutions will become more common also in commercial CPQ tools. Wapice targets to be a forerunner in this field and is thus aiming to integrate AI as part of its Summium® CPQ sales configurator system. This use case of a recommendation system is one of many potential use cases in which AI can bring significant benefit in B2B sales processes and CPQ tools.

One purpose of the demonstrator within InnoSale project from Wapice’s point of view is to validate and test how AI assisted recommendation system could be implemented, what kind of data would be needed, and how it would be integrated into Summium® CPQ. InnoSale project is very fruitful setting for this kind of research where the work is done together in close collaboration with VTT, Wapice, and Wapice’s current customer Molok. Depending on the results and the feedback from Molok, Wapice is willing to continue the development after the InnoSale project and integrate the solution into Summium® CPQ, to commercialize it and have it available also for other Wapice’s customers.

If such a solution will be commercialized and made available in Summium® CPQ tool, it will give Wapice a head start in AI field compared to its closest competitors and strengthen its market share as a CPQ market leader in Finland. Also other Wapice’s customers, whose needs may vary between markets or geographical locations, would receive similar benefits of the AI assisted CPQ solution. When the feedback and needs of other customers can be taken into consideration in the AI development, the original solution can evolve and be iterated to include additional features which are not necessarily seen at this stage of the research and development phase.

Lessons learned

During the project, the following lessons learned were revealed related to business benefits:

  • Accuracy and data volume: The amount of data plays a critical role in achieving reliable and accurate analysis results, which is essential for business success. For instance, a dataset containing thousands of offers from the main market area can yield more precise analysis outcomes than deduced from a smaller or new market area with less accumulated data.
  • Data structure for AI analysis: The data format and contents in existing configurators and CRMs may not be best suited for AI-assisted analysis. For example, in the data structure, there is no field for certain data that is relevant for analysis. Another challenge can be that configurator rules—i.e., which product parts fit with each other—are not easily decoded from the data structure. As a result, if the analysis is based only on data, it may recommend a configuration that is not feasible.
  • Changing customer need: Relying on historical data can lead to non-optimal recommendations if customer needs change rapidly. For instance, when regulations shift in the waste management business, these changes quickly impact customer requirements. Trend analysis alone may not be sufficient to identify the problem promptly. A potential solution involves enabling salespersons to provide regulation-related feedback to the AI system. Alternatively, another AI entity could collect public regulation information and provide updates to the recommendation AI and to salespersons.

Authors

Marko Jurvansuu (VTT), Kai Huittinen (Wapice), Marjaana Tyven-Jokinen (Molok) and Anna Räty (Molok).