Business Horizons (xxxx) xxx, xxx

Available online at www.sciencedirect.com

ScienceDirect
w w w. j o u r n a l s . e l s e v i e r. c o m / b u s i n e s s - h o r i z o n s

Collaborative intelligence: How human

and artificial intelligence create value
along the B2B sales funnel
Jeannette Paschen a,*, Matthew Wilson b, João J. Ferreira c

a
Royal Institute of Technology (KTH), Sweden Lindstedtsvägen 30, 114 28 Stockholm,
Sweden
b
College of Business Administration, Central Michigan University, Mount Pleasant, MI,
U.S.A.
c
NECE-UBI Research Center in Business Sciences, Universidade da Beira Interior (BI),
Covilhã, Portugal

KEYWORDS Abstract The B2B sales process is undergoing substantial transformations fueled
Artificial intelligence; by advances in information and communications technology, specifically in artificial
Natural language intelligence (AI). The premise of AI is to turn vast amounts of data into information
processing; for superior knowledge creation and knowledge management in B2B sales. In doing
Machine learning; so, AI can significantly alter the traditional human-centric sales process. In this
B2B sales; article, we describe how AI affects the B2B sales funnel. For each stage of the fun-
Predictive analytics nel, we describe key sales tasks, explain the specific contributions AI can bring, and
clarify the role humans play. We also outline managerial considerations to maxi-
mize the contributions from AI and people in the context of B2B sales.
ª 2020 Kelley School of Business, Indiana University. Published by Elsevier Inc. All
rights reserved.

remembered and loved and helped by so

“And when I saw that, I realized that selling many different people?”
was the greatest career a man could want.
dWilly Loman, Act II, in Arthur Miller’s Death of a
’Cause what could be more satisfying than to Salesman’
be able to go, at the age of 84, into 20 or 30
different cities, and pick up a phone, and be
1. Can AI contribute to sales
performance?
* Corresponding author
E-mail addresses: [email protected]
(J. Paschen), [email protected] (M. Wilson), [email protected] B2B sales managers have always been concerned
(J.J. Ferreira) with serving their clients at every stage of the

https://doi.org/10.1016/j.bushor.2020.01.003
0007-6813/ª 2020 Kelley School of Business, Indiana University. Published by Elsevier Inc. All rights reserved.
2 J. Paschen et al.

sales process. However, much has changed since market- and technology-driven changes, B2B sales
the days of stereotypical salesperson Willy Loman, organizations must become true knowledge orga-
the protagonist in Arthur Miller’s 1949 play Death nizations (Ingram, 2004). Superior knowledge
of a Salesman. As the opening quote reveals, about their marketdcustomers, users, competi-
Loman’s travels took him to many cities in which tors, and other market entities (Abrell,
he could rely on personal, emotional, and profes- Pihlajamaa, Kanto, vom Brocke, & Uebernickel,
sional connections with his clients. In the process, 2016; Kohli & Jaworski, 1990; Slater & Narver,
Loman and his fellow salespeople mainly relied on 1995)dcan help B2B sales organizations create a
manual, analog technologies (e.g., maps, landline sustainable competitive advantage. Here, AI is
telephones) to enable their B2B sales. expected to bring advanced information quality
The launch of the world’s first mobile telephone and quantity by which B2B firms can improve
certainly seemsdand wasda long time ago. Many personalization, customization, and data man-
technological innovations have shaped B2B sales agement (EverString, 2018).
since, especially profound advances in information However, there appears to be a discrepancy
and communications technology (Morkunas, between marketers’ enthusiasm for AI’s potential
Paschen, & Boon, 2019; Paschen, Pitt, & and their knowledge of AI and how to utilize it
Kietzmann, 2019), digitization, and artificial in- (Campbell, Sands, Ferraro, Tsao, & Mavrommatis,
telligence (AI) (Gupta, Keen, Shah, Verdier, & 2020; EverString, 2018). Understandably, mar-
Walutowy, 2017; Syam & Sharma, 2018). In keters are thus hesitant to incorporate AI into their
particular, AI systemsddefined as information marketing strategies and tactics. While it is clear
systems that act intelligently (Russell, 1997; that AI is poised to change the B2B sales process,
Russell & Norvig, 2016)dare substantially there is little guidance on the specific contribu-
different than technological advancements of the tions AI can bring to each stage of the traditional
past (Kietzmann, Paschen, & Treen, 2018). While human-centric process, and what type of changes
previous technological innovations, including the AI will bring to human contributions. This lack of
advent of the telephone, computers, or the managerial guidance is problematic (Kietzmann &
internet, helped with collecting, aggregating, Pitt, 2020). Moreover, scholars have recently
processing, or communicating data (Paschen, Pitt, recognized the need to further our understanding
& Kietzmann, 2020), decision making always res- of the role of AI in B2B sales, calling for scholarly
ted with humans. Today, emerging technologies work in this area (see Martı́nez-López & Casillas,
such as AI can significantly alter the interactions 2013; Singh et al., 2019; Syam & Sharma, 2018).
between humans and machines (Marr, 2016), In this article, we respond to this call for more
enabling computers to solve problems with mini- research by explicating how AI changes the B2B
mal or no human intervention (Syam & Sharma, sales process. Specifically, we rely on a seven-step
2018). sales process that has been used extensively in
The B2B sales field is not immune to the existing research and applies to most sales situa-
fundamental changes brought about by AI. In fact, tions (Dubinsky, 1981; Homburg, Müller, &
the B2B sector shows a growing interest in AI Klarmann, 2011). For each step of the process,
(Martı́nez-López & Casillas, 2013; Paschen, also known as the sales funnel (Syam & Sharma,
Kietzmann, & Kietzmann, 2019; Singh et al., 2018), we describe key sales tasks and explain
2019; Syam & Sharma, 2018). A recent survey of how AI adds value. In addition, we clarify the role
more than 1,400 B2B marketing executives con- that human contributions and human decision
ducted by MIT Technology Review and Google making play at each step of the AI-enabled sales
revealed that professional services rank among the funnel.
top sectors to embrace AI (MIT Technology Review In this article, we provide a brief overview of AI
Insights, 2018). AI can help B2B companies trans- systems and explain what contributions AI and
late vast amounts of data into information and, people can bring to each stage of sales funnel. We
ultimately, knowledge in order to develop effec- close with an outline of seven managerial consid-
tive sales strategies and tactics. This has been a erations to maximize the contributions from AI and
difficult undertaking for B2B due to rapidly people in the context of B2B sales.
changing customer preferences, a longer sales
process with multiple influencers making buying 2. Foundations of AI
decisions, and changes in the market that occur
more frequently and rapidly (Cotter, Guan, AI could be a misleading term as it suggests the
Mahdavian, Razzaq, & Schneider, 2018; Ingram, possibility that computers display human-like in-
2004). In successfully responding to these telligence (Kaplan & Haenlein, 2019; Russell &
Collaborative intelligence 3

Norvig, 2016). This is not the case. Rather, today’s distinguishes them from traditional information
AI systems are information systems that act systems. AI uses natural language understanding
rationally given what they know. In other words, to analyze and assign meaning to human language
the performance of an AI system is not measured in in spoken and written form (Paschen et al., 2019;
terms of its conformity to human intelligence, but Syam & Sharma, 2018). For example, AI systems
rather in terms of an ideal performance called can extract themes, keywords, sentiments, or
rationality (Gershman, Horvitz, & Tenenbaum, emotions from a piece of text, allowing marketers
2015; Paschen et al., 2019; Russell, 1997; Russell to paint a clearer picture of their customers’
& Norvig, 2016; Tecuci, 2012). An AI system is profiles, needs, and behaviors. AI systems can also
intelligent if it does the ‘right thing’ given what it process nonverbal unstructured data; for example,
knows; AI systems thus solve problems to achieve computer vision recognizes patterns and makes
the best outcome or, in the case of uncertainty, sense of still images, facial cues, or gestures
the best expected outcome. Our definition of AI in (Forsyth & Ponce, 2011; Paschen et al., 2019).
this article departs from the notion often adopted Natural language understanding and computer
in the popular media by which AI emulates human vision are two important preprocessing steps.
intelligence; we conceptualize AI as information These steps transform data via normalization,
systems that act rationally based on the informa- feature extraction, and selection. The resulting
tion available to them in order to solve problems. data can then be further processed in value-
While there is a large number of AI applications creating ways by AI’s three main processes
across many industries, all AI systems can be (Paschen et al., 2019): problem solving, reasoning,
explained using a common input-process-output and machine learning (ML). Problem solving and
model: AI information systems require data from reasoning define the problem AI should solve and
their environment (inputs), manipulate such data how to approach the analysis. These processes
in value-creating ways (processes), and feed in- result in the all-important identification of pat-
formation (outputs) back to the environment terns and conclusions from the data being
(Paschen et al., 2019). We explain each of these analyzed. For example, marketers may want to
three componentsdinputs, processes, and out- identify prospects (i.e., the problem to be solved)
putsdbriefly below. using a segmentation model based on customer’s
web browsing history, e-mail, and telephone in-
2.1. Inputs quiries and demographics (i.e., reasoning or how
to approach the analysis).
An AI information system requires data from its ML is an integral part of AI processes. If the
environment to feed its input-process-output foundation of AI information systems is to act
transformation. Data are values, describing an rationally based on existing data and information,
item or a person with respect to its qualitative or then they need to be able to learn from past at-
quantitative variables, but only when data are tempts. This is the essence of the third main pro-
analyzed and processed can they become useful cess, ML, which encompasses computational
for decision making (Bellinger, Castro, & Mills, procedures that enable AI systems to learn from
2004). For AI, data inputs come in two forms: experience. ML enables AI progressively to improve
structured data encompasses standardized data- its performance without being explicitly pro-
sets in numerical formdsuch as demographics, grammed by humans to do so (Paschen et al., 2019;
web clicks, or transaction recordsdwhile un- Russell & Norvig, 2016; Tecuci, 2012) as the in-
structured data is non-numerical and multifac- formation system stores previous data, problem
eted in the form of text, audio, or images (e.g., solving, and reasoning processes and outcomes in
comments, likes, reviews, inquiries, photos, its knowledge base.
videos). An estimated 80% of today’s data are There are three types of ML: supervised ML,
unstructured (Rizkallah, 2017) and they are unsupervised ML, and reinforcement learning.
growing 15 times faster than structured data (Nair With supervised ML, the AI system is given training
& Narayanan, 2012). data sets, including the inputs and the correct
outputs (i.e., correct answers) from which the
2.2. Processes computer learns patterns and develops rules to be
applied to future instances of the same problem.
With their enormous computing power, AI systems For example, marketers wanting to predict
can process large amounts of structured data very customer churn may collect past examples of the
efficiently, but it is their ability to process un- same problem and identify characteristics com-
structured data in value-creating ways that mon across customers who left for the
4 J. Paschen et al.

competition. The AI system would then use these systems can make at each stage of the sales pro-
training datasets showing the correct inputs and cess, in addition to clarifying how and where
correct outputs to learn the pattern and apply to human contributions add value to the AI-enabled
future instances of the same task (e.g., predicting sales process (see Table 1). To do this, we rely
customer churn among its current client base). on the seven stages of the B2B sales process:
With unsupervised ML, the computer is using
training data that are not labeled with the correct 1. Prospecting;
answers. Its task is to identify patterns or re-
lationships between the data points. For example, 2. Pre-approach;
unsupervised learning can be used to identify
products that are ordered together. To be effec- 3. Approach;
tive, both supervised and unsupervised approaches
require substantial amounts of data (i.e., big data) 4. Presentation;
and high computing power that today is relatively
accessible (Syam & Sharma, 2018). Finally, in 5. Overcoming objections;
reinforcement learning, an AI system learns from
its own experience. It differs from supervised 6. Close; and
learning in that the correct input/output pairs
need not be presented. Instead, the focus is 7. Follow-up.
finding a balance between exploration of un-
charted data territory and exploitation of current This model of the sales process is applicable to
knowledge based on past experiences (Kaelbling, most B2B sales situations (Sheth & Sharma, 2008;
Littman, & Moore, 1996). Syam & Sharma, 2018). It has been extensively
used in previous research (Homburg et al., 2011;
2.3. Outputs Järvinen & Taiminen, 2016) and traditionally has
been referred to as the sales funnel. More recent
studies suggest that, especially in the early stages,
The remaining component of an AI system is out-
the marketing and sales functions of B2B firms are
puts, the information resulting from the above
highly integrated, thus calling for a conceptuali-
value-creating processes that feeds into various
zation that includes both marketing and sales tasks
business applications (Paschen et al., 2019;
(Järvinen & Taiminen, 2016; Syam & Sharma,
Tecuci, 2012). In its basic form, an AI system may
2018). In line with this view, while we use the
produce information as simple as a list of topics
term sales funnel we also discuss marketing tasks
frequently mentioned in news articles about a
as they apply to different stages of the sales
competitor. This information may then require
funnel.
further actions by human decision makers such as
analysts using AI-generated information about a
competitor to create sales battlecards. In addi- 3.1. Prospecting
tion, some AI systems act independently of human
input. Consider how chatbots respond to customer The first task in prospectingdalso known as lead
inquiries, or how an AI system using natural lan- generationdis finding potential customers, which
guage generation can create ad copy or news aligns with the traditional segmentation task in
reports. marketing (Järvinen & Taiminen, 2016; Syam &
Sharma, 2018). For example, a prospect could
provide his/her contact information as part of a
3. The AI-enabled B2B sales funnel sales inquiry on a company’s website, via phone,
or while accessing digital content. In addition, a
In Section 2, we provided a summary of the foun- prospect’s web browsing and web search history
dational building blocks by which to explicate any can provide valuable information for managers
AI system. Now we turn to how the use of different about potential leads and their interests.
AI building blocks changes the traditional human- Next, firms will narrow the pool of prospects to
centered B2B sales process. Indeed, at the heart the ones that are likely to buy. This is the goal of
of our argument lies the idea that combinations of prospect scoring, the second task in prospecting,
different building blocks and the resulting AI ap- which evaluates whether the potential customer
plications lead to different opportunities for value can buy the firm’s offerings and how likely is it that
creation during the B2B sales process. Specifically, he/she will buy? The goal of scoring is to identify
we explain the specific value-add contributions AI high-quality leads (i.e., prospects that offer a high
 Collaborative intelligence
Table 1. Artificial and human intelligence for value creation in B2B sales
Stage of the B2B Marketing and Traditional Marketing and Sales AI Value-Add to Traditional Human Value-Add to AI-enabled
Sales Funnel Task Marketing and Sales Task Marketing and Sales Task
(1) Prospecting  Lead generation: Finding poten-  Build rich prospect profiles  Interpret prospect lists and
tial customers (structured and unstructured explain inconsistencies
 Lead qualification: Evaluating data)  Verify lead qualification and put
prospects’ propensity to buy  Predictive lead qualification into a business context
 Update lead generation and lead
qualification models via ML
(2) Pre-approach and (3)  Lead Nurturing: Acquiring more  Ad targeting and retargeting:  Monitor ad targeting, re-
Approach information about leads and Personalized and customized targeting, and content curation
making contact communication messages and  Take over from or delegate to AI-
channel digital agents during contact
 Content curation
 Making contact via digital agents
(e.g., chatbots)
(4) Presentation  Communicating the problem-  AI-enabled prototyping  Build rapport, trust, and
solving characteristics of the  Emotion AI credibility
offering (e.g., prototype, use  Address questions in person
cases, simulation)  Interpret emotions and respond
effectively
(5) Overcoming objections and  Negotiating sale and overcoming  Curate competitive intelligence  Communicate product benefits
(6) Closing objections (e.g., for sales battlecards)  Build rapport, trust, and
 Dynamic pricing credibility
 Interpret emotions and respond
effectively
 Persuasive communication
(7) Follow-up  Fulfilling the current order  Automate workflows  Oversee order processing and
 Follow up beyond the current  Automate post-order service fulfillment
order (e.g., chatbots)  Strengthening exchange
 Upsell, cross-sell  Build rich customer profiles relationship
(structured and unstructured  Personal post-service follow-up
data) to uncover new needs  Interpret new needs and explain
inconsistencies

5
6 J. Paschen et al.

probability of profitable sales). Prospect scoring, professionals to change their value-add contribu-
also known as lead qualification, is a challenging tions. While AI may be implemented to take charge
task that requires substantial human resources of identifying and evaluating leads, human judg-
(Järvinen & Taiminen, 2016) because of the diffi- ment is necessary in order to interpret and filter
culties in identifying the characteristics of high- the AI-generated information. Specifically, one key
quality prospects. Even when a firm is able to task of sales professionals is to interpret AI-
identify these characteristics, this information enabled prospect lists. While potential buyers
often becomes available only after a sales pro- may tick all the boxes on finding and qualifying
fessional has made contact with a potential client, prospects, sales professionals will need rely on
thus limiting its proactivity. their experience and intuition to check for in-
AI promises to benefit both of these key tasks at consistencies and put the AI-generated informa-
the prospecting stage by undertaking activities tion into their specific business context. They will
that previously relied heavily on sales pro- need to determine whether the prospect has a
fessionals and their support teams. First, AI is well specific need that the business can actually satisfy,
suited to analyze both structured and unstructured whether it is economically feasible and whether it
data used in segmenting consumers and generating is operationally achievable for the prospect to
lists of prospects. For example, natural language implement the firm’s specific offering. These
understanding can analyze text data (e.g., email judgments often tap into sales peoples’ tacit
inquiries, social media posts, news releases) and knowledge, experience, and intuitiondareas that
identify keywords, themes, or current interests AI is currently not capable of accessing or
from the text. AI’s capabilities are not limited to analyzing.
written content; rather, AI systems can also
analyze speech or image data originating from 3.2. Pre-approach and approach
photos, videos, or conversations. This, combined
with AI’s capability to analyze vast amounts of A prospect that has been identified and positively
structured data (e.g., web traffic and clicks) in evaluated as a quality lead moves on to the next
addition to offline structured data (e.g., tenure two stages: pre-approach and approach. These
and role in the client organization), can be used to two stages typically have been examined together
create a well-rounded profile of prospects. In in sales research and some suggest that these two
addition, ML is increasing the efficacy of these stages are merging (Syam & Sharma, 2018). Pre-
segmentation algorithms by updating and approach and approach mean acquiring more
improving their performance without humans detailed information about leads, including their
reprogramming them (Syam & Sharma, 2018). needs, habits, preferences, and other relevant
Second, AI is capable of developing and applying background information and then making contact.
predictive algorithms to engage in lead scoring. AI They include conducting more research about the
systems can analyze previous prospect data and prospective buyer and providing customized and
determine what actual online and offline character- personalized content related to a prospect’s
istics have the highest probability of resulting in interestsda tactic that is referred to as lead
conversions (Syam & Sharma, 2018). Through ML, nurturing. While pre-approach is the doing-your-
these predictive algorithms then can be used to homework part of the sales process, the
create new prospect lists. In addition, natural lan- approach stage is about building relationships,
guage understanding enables computers to identify with sellers aiming to establish rapport and build
sentiment and keywords in consumers’ text or speech trust with potential buyers in addition to gaining
to predict the probability that these prospects will more insight into whether and how a lead could
purchase. Finally, ML allows AI continuously to up- benefit from a firm’s offering.
date the rules for generating and scoring leads, AI influences the pre-approach and approach
storing this information in the AI knowledge base to stages of the sales funnel in several ways. First, AI
allow for continual improvement. Dell uses AI- can automate some of the routine tasks around
enabled analytics to identify the types of prospect making contact, such as scheduling meetings or
behavior most likely to lead to the purchase of sending follow-up emails. This frees up sales pro-
different products, thus helping its sales force to fessionals and allows them to spend time and en-
pinpoint high-quality leads (King, 2012). It has resul- ergy on other value-add contributions.
ted in a substantial increase in the productivity of Second, AI enables the creation and delivery of
Dell’s sales force as well as the efficiency of its sales. highly tailored and personalized communications
AI’s significant contributions to the prospecting to sales leads based on an analysis of the lead’s
stage also create new opportunities for sales behavior or profile. This is known as targeted
Collaborative intelligence 7

advertising. Targeted advertising can take a num- prepared presentation should keep the audience
ber of forms, including retargeting in which AI can in minddwhat they want and do not wantdand be
discern the content that is likely to appeal to a clear and concise in tone and content.
lead based on previously consumed content. AI is capable of supporting major elements of
Ideally, these tailored approaches will lead to the presentation stage. AI-assisted prototyping
greater engagement and ultimately conversions already exists, whereby the AI algorithm takes
along the B2B sales funnel. AI can also be used to ideas from the drawing board and turns them into
curate promotional content. AI tools can be used actual products almost instantaneously (Mix,
to help discover similarities in the characteristics 2017). The AI system used by Airbnb, for
of top-ranking content. AI is sophisticated enough instance, is trained to recognize the company’s
to offer real-time recommendations on how to standard hand-drawn design sketches and render
improve content, what the best delivery vehicle them into actual computer programming code. The
might be, and the overall format of the commu- automation of these tasks allows sales pro-
nication. These types of targeted advertising are a fessionals to deliver tailored prototypes much
more effective way to pre-approach or approach faster than what was possible traditionally.
sales. For example, ServiceMax, a provider of field- In addition, AI systems can augment the pre-
technician management software for equipment sentation task. Slide bots (i.e., AI systems that
and infrastructure businesses, is using ML to pre- analyze existing presentation content) identify the
dict the best customer journey for its site visitors. main ideas and messages and subsequently
By showing prospects recommendations and snip- recommend and execute on the optimal layout and
pets for the next webpage they should visit, content. These bots can help sales professionals
ServiceMax decreased bounce rates to its site by deliver more compelling presentation materials.
70%, doubled the time visitors spent on their site, Natural language understanding, computer vision,
and substantially increased requests for product and speech recognitiondin combination with ML
demos (Fleming, 2019). algorithmsdcan recognize verbal or nonverbal
Finally, AI can automate contact through chat- cues in communication patterns of the presenter
bots (i.e., a computer program that can engage in or audience. For example, AI systems can run a
a conversation with a person). Sophisticated sentiment analysis on a presenter’s words, voice,
intelligent agents already make contact with and tone and compare the result against desired
leads, answer frequently asked questions, and norms, thus providing instantaneous feedback on a
refer individuals to sales professionals as needed. sales presentation. Moreover, AI can be used to
For all of these AI-enabled tools, sales pro- analyze the emotions of the audience to deter-
fessionals will be needed to monitor the adver- mine sentiment or themes that can render useful
tisement targeting, retargeting, and content insights about potential concerns or objections
curation completed by AI and engage in a personal from customers.
interaction with the prospect at the appropriate Although AI can be used to cover many tasks
time. For example, sales platform firm 6sense of- within the sales presentation phase, several
fers a service built on AI, ML, and predictive ana- remain with sales professionals. First, sales pro-
lytics. It provides visibility for B2B marketing and fessionals must continue to build rapport with
sales professionals into the buying journey of potential and current clients. As in all stages of the
prospects and flags a window of opportunity, sales funnel, a personal touch is important to build
indicating when it is most effective for salespeople trust and credibility, and sales professionals are
to engage with the prospect. In addition to moni- needed to address any potential ambiguity. In
toring AI results, sales professionals may also need addition, while AI may automate layout and
to take over from chatbots or delegate work that formatting tasks in creating presentation decks,
emerges through chatbot interactions in order to sales professionals will need to put the final touch
push potential sales further along the funnel. on these materials before implementing them in
the presentation stage.
3.3. Presentation
3.4. Overcoming objections and closing
At the presentation stage, the selling firm dem-
onstrates the problem-solving characteristics of At any point during a sales presentationdin fact,
the offering and also designs the offer specific to at any stage of the sales processda potential
the potential client’s needs, which often includes customer may object to one or more of the points
providing a prototype of the product or solution made by the selling firm. These objections can
being offered (Syam & Sharma, 2018). A well- come in the form of questions, statements
8 J. Paschen et al.

indicating disagreement, or even nonverbal ex- order is completed. Filling the current order
pressions such as facial expressions or body lan- broadly comprises recording the order, initiating
guage that may indicate a customer’s order processing, inventory management, and
unwillingness to buy. Potential clients may raise order fulfillment via supply chain and procurement
questions about the selling company, delivery systems. Following up after the completion of the
method, the competition, or the price, product, or initial order involves uncovering new needs the
service itself. A key task for sales professionals at client may have, which may result in the customer
this stage is to manage customer objections by entering the sales funnel again. In addition, there
becoming aware of the underlying reasons so that may be upsell or cross-sell opportunities during the
the selling company can close the deal. follow-up stage. Upselling is encouraging the pur-
AI changes the tasks involved in overcoming chase of anything that would make the primary
objections and enables sales professionals to purchase more expensive (e.g., an upgrade, add-
respond to the concerns faster, with more up-to- on or premium product), and cross-selling en-
date and more comprehensive information via courages the purchase of anything in conjunction
tools such as AI-enabled battlecards. Specifically, with the primary product.
the abilities of AI systems to analyze structured AI can contribute to this stage of the sales
and unstructured data can be used to strengthen a funnel in a number of ways. First, AI can automate
firm’s own value proposition and depose competi- the workflows required for order processing and
tors. The startup Klue offers an AI service based on the follow-up. This could include automated
natural language understanding and ML to curate paperwork, inventory management, and supply
competitive intelligence for B2B personal selling chain management and allows sales professionals
and sales management. This up-to-date informa- to be free from repetitive and time-consuming
tion is summarized in battlecards that sales pro- tasks. Second, with specific regard to the follow-
fessionals can use to oust competitors or handle up of the order, AI can be used to automate some
questions raised by clients. AI also influences the elements of the post-order service. Chatbots can
negotiation stage with dynamic pricing. Algorithms be used to initiate communication with the client
analyze historical data about pricing variables and begin discussing with customers about their
along with other information about leads to experience and future needs. Finally, AI analysis of
determine the best prices for different potential both structured and unstructured buyer data (e.g.,
buyers. Industrial manufacturing companies like behavioral information) can be undertaken in
Siemens or Honeywell leverage an AI-enabled order to uncover new needs and to re-enter the
pricing solution by Vendavo to manage price dif- funnel at stage one. Finding patterns in past pur-
ferentiation during their sales process. chase behaviors using ML, firms can identify which
For less complex products, AI systems are able products often are bought together and display the
to take over some of the tasks in the closing stage, complementary product to a customer, increasing
such as chatbots answering common questions. the overall size of purchase and improving the
However, more complex and personalized offer- customer experience. The Hyatt Hotels Group uses
ings require the closing stage to see the most sales ML to improve cross- and up-selling to its cus-
representative involvement. At this stage, sales tomers, resulting in a 60% increase in average in-
professionals still heavily rely on interpersonal cremental room revenue (Diaz, 2017). By
relationship skills such as building trust and identifying patterns based on guest history and
engaging in persuasive communication. In addi- past behaviors and comparing these to guests with
tion, perceiving client cuesdboth verbal and similar profiles, Hyatt is able to identify guests
nonverbaldis an important task that is best suited that are likely to upgrade their room or may be
to a sales professional. Despite AI’s ability to interested in the hotels’ amenities.
analyze emotion in some contexts, sales pro- The role of sales professionals in the follow-up
fessionals are needed to interpret AI-generated stage includes overseeing order processing and
analyses of emotions and use their discretion in fulfillment to ensure orders are processed accu-
adjusting closing approaches based on the infor- rately and in a timely fashion. In addition, sales
mation from AI systems as well as from their own professionals must continue to strengthen re-
observations. lationships with clients through personalized
follow-up communications. AI can provide more
3.5. Follow-up detailed information than a sales professional
might be able to gather on their own, but sales
Follow-up involves two different components: professionals remain uniquely and solely situated
filling the current order and following up after the to connect with clients on a personal level. Using
Collaborative intelligence 9

the detailed information that AI can provide and quantity (Archer-Brown & Kietzmann, 2018). An
integrating this information with their personal important step in knowledge management is in
experience and knowledge, sales professionals making these assets tangible and accessible for
have a fine-tuned understanding of client needs managerial decision making. AI is a key enabler to
and can integrate this knowledge into their follow- making intangible assets accessible by capturing,
up strategies. In particular, using this detailed in- organizing, and sharing information. To maximize
formation can help sales professionals to identify the benefits, managers should integrate AI-
inconsistencies and to interpret and anticipate enabled knowledge management activities with a
new client needs, ultimately strengthening the B2B firm’s enterprise knowledge management
client relationship. strategies and tactics.

4.3. Leave insights and social/emotional

4. Managerial considerations for
tasks to human intelligence
maximizing collaborative intelligence
AI systems can analyze big data, particularly un-
The potential for combining AI and human intelli- structured data, often in real-time and transform
gence to maximize value throughout the B2B sales these data pieces into useful information. How-
funnel is significant. The importance of employing ever, it is human intelligence that is critical in
highly skilled sales professionals remains high, yet deriving the implications of the AI analysis and
there is potential for added value in the use of AI translating AI information into knowledge.
throughout all stages of the sales process. In this Answering the ‘so what’ question and deciding on
section, we provide guidance for managers to an appropriate course of action is a task for which
maximize value from human intelligence and AI in human intelligence outperforms AI. Human
B2B sales. These managerial considerations are reasoning is able to make use of a wide context of
intended to help managers understand some of the human experiences, backgrounds, and skills and
critical issues they may face when integrating AI bring this to bear in solving business problems; in
into their sales process. contrast, AI systems typically have a very narrow
focus. In addition, AI systems are limited in dis-
4.1. Training is essential playing emotional or social competencies them-
selves (Canhoto & Clear, 2020; Kaplan &
Managers must train both sales professionals and Haenlein, 2019). However, these social and
support staff to use AI and interpreting AI- emotional competencies are particularly impor-
generated information. One broad trend tant in B2B sales and will continue to be critical
throughout all stages of the AI-enabled sales fun- tasks performed by humans in the AI-enabled
nel is that AI can help create, organize, and even sales funnel.
use a wealth of information, but it is up to humans
to check this information for weird output or in- 4.4. Support customers through the
consistencies. It is necessary to train salespeople transition
and staff so they can correctly interpret the output
of AI and identify information that does not align Similar to achieving employee buy-in on the
with the specific business context. Moreover, as importance and role of AI in the sales process,
the AI output is contingent upon the input and managers need to acknowledge and prepare for
training data (particularly in the case of supervised the transition customers will go through. AI will
ML algorithms), sales professionals need to be change the customer experience, and customers
aware of the limitations with these training data or will need to be brought up to speed in order to
inputs. Indeed, employees will need to develop help them embrace the AI-enabled sales funnel.
new skills in order to extract value from AI systems Managers should ensure that customers do not
(Kaplan & Haenlein, 2019), and training will be perceive AI as a way companies to do less for
essential in helping employees adapt. them, but rather as a way for firms to provide
customers with better offerings and more effec-
4.2. Link AI to enterprise knowledge tive sales interactions than ever before. Some
management strategies and tactics customers may be unwilling to embrace this tran-
sition or distrust AI-enabled tasks. In these situa-
Researchers suggest that knowledge management tions, managers must identify these hesitant
is a process by which firms create value from customers and then possibly take a traditional
intangible assets such as information quality and approach to serving them.
10 J. Paschen et al.

4.5. Expect resistance (Singh et al., 2019). Managers need to consider the
potential impact of AI on their departmental and
Many of the employees involved in the sales process organizational structure and devise processes that
will be hesitant to embrace AI over concerns that AI allow AI to be integrated into organizational
will make their roles obsolete or simply because workflows and practices.
they are resistant to change. In either case, those in
positions of leadership must follow best practices
for change management (Seijts & Gandz, 2018) in 5. AI will enhance, not replace
order to smooth the transition to an AI-enabled
sales funnel. In particular, managers must AI is poised to set off a series of fundamental
approach change management with an eye to changes to the B2B sales process. These profound
managing rapid change via digitization (Crittenden, changes will affect the nature of human work,
Crittenden, & Crittenden, 2019). One point lead- specifically human intelligence and decision making
ership can make clear is that personal, human (Syam & Sharma, 2018). B2B sales professionals will
contributions remain a critical need in the sales be supported by AI and may even relinquish some of
process. Indeed, it would be incorrect to assume the tasks previously performed by people. The
that the need for personal touches will become less trouble is that executives eager to adopt these new
important in an AI-enabled world. It may very well technologies are unclear about the contributions
become more important to clients that there is a that AI may bring to the B2B sales process and the
personal connection, and reiterating this point to role that sales professionals play in the AI-enabled
hesitant employees may be important. sales funnel. This understanding is even more crit-
ical when adopting AI means relinquishing some
4.6. Information security is paramount degree of human decision making and control.
Against this backdrop, our article explicates the
As more information is collected, stored, and relied value-add contributions of AI systems at each
upon, information security becomes increasingly stage of the sales funnel, in addition to clarifying
important (Santanen, 2019). Managers should the role that human intelligence and decision
revisit their information security practices to making play at each stage of the AI-enabled sales
ensure that this valuable resourcedand their cli- funnel. We hope that our article provides a
ents’ privacydis protected. One of the key pre- comprehensive perspective on the complemen-
mises of AI is to paint more well-rounded profiles of tarity (Jarrahi, 2018) of humans and AI. AI, with its
customers by identifying and analyzing customer enormous information processing capacity, can
behavior across multiple devices and many situa- augment human intelligence or even replace well-
tions. This means that even if one’s personal data is defined and repeatable human tasks in B2B sales.
anonymized once they become part of AI’s input On the other hand, humans still hold an upper
processes, an AI system could de-anonymize this hand in using intuition to deal with contradictory
data based on inferences from other devices. This or uncertain information and to derive insights and
blurs the distinction between identifiable and un- implications related to their business and in the
identifiable data and poses challenges for firms to relationship-building aspects that are paramount
adhere to legislated privacy requirements. in B2B sales. In essence, even in times of profound
transformation, the ultimate goal of sales pro-
fessionals is to help customers satisfy their needs,
4.7. Build a sales force structure and as the character Willy Loman suggested. What has
processes supportive of AI changed is how the end goal is achieved and how
emerging technologies, specifically AI, can support
Traditionally, salespeople relied heavily on support salespeople in achieving this goal by turning big
staff such as sales enablement staff, sales assis- data into information and, ultimately, knowledge.
tants, or marketing research specialists for gath-
ering and curating information about customers,
References
competitors, or other market forces. AI’s ability to
analyze vast amounts of data and develop and
Abrell, T., Pihlajamaa, M., Kanto, L., vom Brocke, J., &
adjust segmentation or pricing models in near- Uebernickel, F. (2016). The role of users and customers in
real-time may change the structure and organiza- digital innovation: Insights from B2B manufacturing firms.
tional processes of these sales-related functions Information Management, 53(3), 324e335.
Collaborative intelligence 11

Archer-Brown, C., & Kietzmann, J. (2018). Strategic knowledge Kietzmann, J., Paschen, J., & Treen, E. (2018). Artificial intel-
management and enterprise social media. Journal of ligence in advertising: How marketers can leverage artificial
Knowledge Management, 22(6), 1288e1309. intelligence along the consumer journey. Journal of Adver-
Bellinger, G., Castro, D., & Mills, A. (2004). Data, information, tising Research, 58(3), 263e267.
knowledge, and wisdom. Available at https://homepages. Kietzmann, J. H., & Pitt, L. F. (2020). AI and machine learning:
dcc.ufmg.br/wamendes/SistemasInformacaoTP/ What general managers need to know. Business Horizons,
TextosBasicos/Data-Information-Knowledge.pdf 63(2). XXXeXXX.
Campbell, C., Sands, S., Ferraro, C., Tsao, J., & King, R. (2012, December 5). How Dell predicts which customers
Mavrommatis, A. (2020). From data to action: How mar- are most likely to buy. The Wall Street Journal. Available at
keters can leverage AI. Business Horizons, 63(2). XXXeXXX. https://blogs.wsj.com/cio/2012/12/05/how-dell-predicts-
Canhoto, A., & Clear, F. (2020). Artificial intelligence and machine which-customers-are-most-likely-to-buy/
learning as business tools: A framework for diagnosing value Kohli, A. K., & Jaworski, B. J. (1990). Market orientation: The
destruction potential. Business Horizons, 63(2). XXXeXXX. construct, research propositions, and managerial implica-
Cotter, T., Guan, M., Mahdavian, M., Razzaq, S., & tions. Journal of Marketing, 54(2), 1e18.
Schneider, J. D. (2018, January). What the future science of Marr, B. (2016, April 5). Why everyone must get ready for the 4th
B2B sales growth looks like. McKinsey & Company. Available industrial revolution. Forbes. Available at https://www.
at https://www.mckinsey.com/business-functions/ forbes.com/sites/bernardmarr/2016/04/05/why-everyone-
marketing-and-sales/our-insights/what-the-future-science- must-get-ready-for-4th-industrial-revolution/#6c7cce403f90
of-b2b-sales-growth-looks-like Martı́nez-López, F. J., & Casillas, J. (2013). Artificial
Crittenden, A. B., Crittenden, V. L., & Crittenden, W. F. (2019). intelligence-based systems applied in industrial marketing:
The digitalization triumvirate: How incumbents survive. An historical overview, current and future insights. Indus-
Business Horizons, 62(2), 259e266. trial Marketing Management, 42(4), 489e495.
Diaz, J. (2017, March 16). How marketers are using machine Mit Technology Review Insights. (2018, November 2). Profes-
learning to cross-sell and upsell. Principa. Available at sional services firms see huge potential in machine learning.
https://insights.principa.co.za/you-want-fries-with-that- Available at https://www.technologyreview.com/s/612374/
using-machine-learning-to-cross-sell-and-up-sell professional-services-firms-see-huge-potential-in-machine-
Dubinsky, A. J. (1981). Perceptions of the sales job: How stu- learning/
dents compare with industrial salespeople. Journal of the Mix, D. (2017, October 25). Airbnb built an AI that turns design
Academy of Marketing Science, 9(4), 352e367. sketches into product source code. The Next Web. Available
EverString. (2018). The state of artificial intelligence in B2B at https://thenextweb.com/artificial-intelligence/2017/
marketing. Available at http://xd07g309cu21kppg63qolt4t- 10/25/airbnb-ai-sketches-design-code/
wpengine.netdna-ssl.com/wp-content/uploads/2018/04/ Morkunas, V., Paschen, J., & Boon, E. (2019). How blockchain
AI-in-B2B-marketing_ES_FINAL.pdf technologies impact your business model. Business Horizons,
Fleming, J. (2019, March 7). Four real-world stories of B2B 62(3), 295e306.
marketers unlocking growth through AI. LinkedIn. Available Nair, R., & Narayanan, A. (2012). Benefitting from big data:
at https://business.linkedin.com/en-uk/marketing- Leveraging unstructured data capabilities for competitive
solutions/blog/posts/B2B-Marketing/2019/Four-real-world- advantage. New York, NY: Booze&Co/PricewaterhouseCoopers.
stories-of-B2B-marketers-unlocking-growth-through-AI Paschen, J., Kietzmann, J., & Kietzmann, T. (2019). Artificial
Forsyth, D., & Ponce, J. (2011). Computer vision: A modern intelligence (AI) and its implications for market knowledge
approach. Upper Saddle River, NJ: Prentice Hall. in B2B marketing. The Journal of Business and Industrial
Gershman, S. J., Horvitz, E. J., & Tenenbaum, J. B. (2015). Marketing, 34(7), 1410e1419.
Computational rationality: A converging paradigm for intel- Paschen, J., Pitt, L., & Kietzmann, J. (2019). Guest editorial.
ligence in brains, minds, and machines. Science, 349(6245), Journal of Business & Industrial Marketing, 34(7),
273e278. 1401e1402.
Gupta, S., Keen, M., Shah, A., Verdier, G., & Walutowy, M. F. Paschen, U., Pitt, C., & Kietzmann, J. H. (2020). Artificial in-
(2017). Digital revolutions in public finance. Washington, telligence: Building blocks and an innovation typology.
DC: International Monetary Fund. Business Horizons, 63(2). XXXeXXX.
Homburg, C., Müller, M., & Klarmann, M. (2011). When should Rizkallah, J. (2017, June 5). The big (unstructured) data prob-
the customer really be king? On the optimum level of lem. Forbes. Available at https://www.forbes.com/sites/
salesperson customer orientation in sales encounters. forbestechcouncil/2017/06/05/the-big-unstructured-data-
Journal of Marketing, 75(2), 55e74. problem/#57541ca2493a
Ingram, T. (2004). Future themes in sales and sales manage- Russell, S. J. (1997). Rationality and intelligence. Artificial In-
ment: Complexity, collaboration, and accountability. Jour- telligence, 94(1/2), 57e77.
nal of Marketing Theory and Practice, 12(4), 18e28. Russell, S. J., & Norvig, P. (2016). Artificial intelligence: A
Jarrahi, M. H. (2018). Artificial intelligence and the future of modern approach (3rd ed.). Upper Saddle River, NJ: Pearson
work: Human-AI symbiosis in organizational decision making. Education Limited.
Business Horizons, 61(4), 577e586. Santanen, E. (2019). The value of protecting privacy. Business
Järvinen, J., & Taiminen, H. (2016). Harnessing marketing Horizons, 62(1), 5e14.
automation for B2B content marketing. Industrial Marketing Seijts, G. H., & Gandz, J. (2018). Transformational change and
Management, 54, 164e175. leader character. Business Horizons, 61(2), 239e249.
Kaelbling, L. P., Littman, M. L., & Moore, A. W. (1996). Rein- Sheth, J. N., & Sharma, A. (2008). The impact of the product to
forcement learning: A survey. Journal of Artificial Intelli- service shift in industrial markets and the evolution of the
gence Research, 4, 237e285. sales organization. Industrial Marketing Management, 37(3),
Kaplan, A., & Haenlein, M. (2019). Siri, Siri, in my hand: Who’s 260e269.
the fairest in the land? On the interpretations, illustrations, Singh, J., Flaherty, K., Sohi, R. S., Deeter-Schmelz, D., Habel, J., Le
and implications of artificial intelligence. Business Horizons, Meunier-FitzHugh, K., et al. (2019). Sales profession and pro-
62(1), 15e25. fessionals in the age of digitization and artificial intelligence
12 J. Paschen et al.

technologies: Concepts, priorities, and questions. Journal of artificial intelligence in sales research and practice. Indus-
Personal Selling and Sales Management, 39(1), 2e22. trial Marketing Management, 69, 135e146.
Slater, S. F., & Narver, J. C. (1995). Market orientation and the Tecuci, G. (2012). Artificial intelligence. WIREs: Computational
learning organization. Journal of Marketing, 59(3), 63e74. Statistics, 4(2), 168e180.
Syam, N., & Sharma, A. (2018). Waiting for a sales renaissance
in the fourth industrial revolution: Machine learning and