Sales
Can AI Really Help You Sell?
by Jim Dickie, Boris Groysberg, Benson P. Shapiro, and Barry Trailer
From the Magazine (November–December 2022)
Núria Madrid
Summary. Many salespeople today are struggling; only 57% of them make their
annual quotas, surveys show. One problem is that buying processes have evolved
faster than selling processes, and buyers today can access a wide range of online
resources that let them evaluate products before even meeting a salesperson. AI
tools can help organizations close the gap, but most don’t know how to use them
effectively. In this article the authors describe how sales AI has been a real game
changer at a few companies. They also provide a self-assessment tool, the Sales
Success Matrix, that will show sales leaders where to start or improve their AI
journeys.
The matrix has two dimensions: relationship level (which runs from transactional
vendor on the low end to trusted co-creator at the top) and process level (which
runs from ad hoc to customized). At the lower levels of both relationships and
processes, simple AI that decreases costs and improves efficiency works best. In
the mid levels, advanced AI increases sales effectiveness by analyzing
opportunities and customer needs. At the highest level, cutting-edge technologies
help firms generate deep insights about customers.
No matter where a firm falls on the matrix, AI can help it boost sales. And the
sooner and more broadly it applies AI tools, the better they work. close
Though more and more companies are
applying sophisticated technology to sales
processes, research suggests that most
aren’t using it effectively (and some don’t
even use it at all). Even customer-
relationship-management systems, which
digitally savvy sales organizations have had in place for decades,
aren’t being fully taken advantage of. In Sales Mastery’s 2022 Sales
Performance Scorecard survey of 332 sales managers, 15% of
respondents reported that their companies were not actively
using CRM, and 42% stated that they were using it only for storing
information about customers and prospects.
00:00 / 33:18
Listen to this article
To hear more, download Noa app for iPhone or Android.
No wonder salespeople have been struggling. According to global
surveys of nearly 1,000 sales leaders by CSO Insights, the
percentage of salespeople meeting their annual quotas fell from
63% to 57% from 2012 to 2019. When the leaders were asked to
evaluate their teams’ performance in 16 distinct sales activities,
they said their teams were less effective at 15 of those tasks than
they had been five years earlier. And according to the sales
executives we talk to, lately the performance of salespeople has
gotten even worse.
Part of the problem is that buying processes have been evolving
faster than selling processes. Buyers are better informed than
ever, with access to a wide range of online resources that help
them evaluate products before ever meeting with a salesperson.
Another factor may be that sales reps spend too much time doing
things that don’t directly involve selling. The Sales Mastery 2022
survey found that on average salespeople devoted only 32% of
their time to selling—and 68% to non-revenue-producing
activities.
The percentage of salespeople
meeting their annual quotas fell from
63% to 57% from 2012 to 2019,
according to global surveys.
For more than two decades, two of the authors of this article
(Barry and Jim) have studied processes and relationships in sales
organizations. A third author (Ben) has been teaching classes on
that subject since the 1970s. The fourth author (Boris) has
researched the organizational implications of AI and data
analytics for many years. Collectively we’ve published dozens of
articles and white papers on those topics. In 2021 we also
surveyed more than 500 sales organizations to assess the role that
AI plays in improving sales performance. In our research and
consulting work, we’ve observed a virtuous cycle: The more AI
tools are applied to a process, the more data is generated. Better
data leads to better algorithms. Better algorithms lead to better
service and greater success. Those, in turn, lead to more usage,
continuing the cycle. So we believe that the sooner an
organization implements AI solutions and the more broadly
they’re applied, the better they work. Success grows
exponentially. And the competitive risks of not adopting AI tools
grow as well.
Unfortunately for sales leaders, implementing AI-based sales
processes isn’t as simple as downloading new software.
In this article we’ll examine the ways in which AI is already used
to facilitate selling—and how it can be used to do more. We’ll
detail how sales leaders who have successfully adopted AI
optimize its performance. And we’ll provide sales leaders with a
self-assessment tool that is designed to help them start or improve
their AI-for-sales journeys.
Heightening Customer Engagement
Let’s begin by looking at an instance where the application of AI
in sales was a real game changer—one that McAfee Enterprise, a
leader in computer security solutions that later became Trellix,
shared with us in May 2021. AI in the company’s internally
developed platform analyzes a billion sensors across its
customers’ systems and identifies and prioritizes security threats.
The platform predicts the impact each threat could have, alerts
the customer, and then prescribes corrective actions. In addition
to increasing the effectiveness of the security teams, the AI is a
valuable tool for the sales organization. According to Pilar
Schenk, the company’s former vice president of global sales
strategy and operations, its sales professionals harness the AI to
understand potential risks for noncustomers and for current
customers who aren’t yet using the platform. Aggregating the
sensor data, the AI gives salespeople targeted recommendations
about which firms in their territories they should proactively
contact and why. The salespeople then follow playbooks on the
sales organization’s High Velocity Sales (HVS) platform, which
describe how they should engage prospects and provide
supporting materials they need to do that.
This approach has altered the dynamic between salespeople and
buyers. Instead of asking potential customers to share
information from their systems, salespeople offer to share the
risks that the AI’s analysis has surfaced for their companies, as
well as advice on how to address them. Since incorporating
sensor-generated insights into HVS, in January 2020, the
company has been tracking the performance of the salespeople
who use them and has found a 10-fold improvement in their
ability to start conversations with prospects. The number of initial
conversations that they’ve converted into sales opportunities has
also increased three-fold. In addition, they’ve had a 5% increase
in renewal rates. Their managers have benefited too: Before HVS,
the managers were spending 9% to 10% of their time coaching
their team members. But since the platform now gives them a
continuous analysis of their salespeople’s activities, revealing
who needs what type of help on what type of opportunity, the
managers no longer have to spend hours trying to figure that out
—and the percentage of time they devote to coaching has jumped
to about 30%.
The Sales Success Matrix
To help companies determine what kinds of AI solutions they’re
ready to implement, we’ve developed a tool that we call the Sales
Success Matrix. It has two axes: relationship level and process
level. Sales organizations can identify their position on it, which
will point them to the kinds of AI tools that would best boost their
sales now and what steps they might take next. For most, the
ultimate goal will be to move up to the highest levels of
relationships and processes, where customer loyalty and
competitive advantage are the strongest.
Relationship Level
The matrix maps out five types of relationships that selling
organizations can have with customers: transactional vendor,
preferred supplier, solution consultant, strategic collaborator, and
trusted co-creator. AI can be useful with all five kinds, but in
different ways.
Transactional vendor. This is the lowest level of relationships.
The customer’s transactions are rapid, repetitive, and routine,
and usually involve self-service or online shopping. To remain
profitable and competitive at this level, companies need to
squeeze out costs, leverage automation, and minimize buyer-
seller interactions, and AI algorithms can help them do this. A
common example of AI at this level is e-commerce site
recommendations: Customers who bought this item also bought
these items. Recommendations can be based not only on the
activity of similar buyers but also on past purchasing history or
imported data, such as web searches, buyer demographics, and
paid placements.
See more HBR charts in Data & Visuals
Preferred supplier. At this level the organization has managed to
differentiate its offerings enough to create a measurable customer
preference. That differentiation gives sellers an opportunity to
gain customer information, which can then be used to win more
business from the customer, generate referrals, cross-sell other
products and services, and obtain still more information about
emerging needs or competitive activity. For example, many
preferred suppliers offer managed services (such as monitoring
printer ink levels and automatically sending new supplies), which
minimize outages and downtime for their customers while
increasing their own revenue and profits. AI can help sellers at
this level anticipate customers’ needs by analyzing historical
usage patterns, comparative user volumes, and maintenance
records.
Solution consultant. At the third level sellers offer a complicated
set of products and services that are integrated into one system.
To be competitive, a seller must get buyers to believe that the
integrated solution provides more value than assembling the
components on their own would. Typically, the seller makes a
profit on both the components and the integration. Software-as-a-
service companies fall into this category. Their sales teams
usually include someone in a customer success role who is
responsible for monitoring usage and encouraging the adoption
of other capabilities. AI applications can assist solution
consultants by offering suggestions, based on customer records
and on “like population” usage data, about how to increase
customer “stickiness” and minimize churn. Trellix’s use of AI to
improve relationships with customers and prospects is a perfect
example of a technology for solution consultants.
Strategic collaborator. At this level the connections between
buyer and seller are stronger and more numerous and intricate.
Relationships are usually regional or even global. And as both the
size of transactions and the duration of the relationship increase,
higher levels of management get involved. The sales approach
required is totally different from the traditional one salespeople
take with purchasing agents. Account management is truly
enterprise-wide and cross-functional and involves orchestrating
multiple varied conversations and marshaling both internal and
external resources. With interactions between buyer and seller
taking place on so many dimensions, complexity grows
exponentially. “Back of the envelope” or spreadsheet tracking of
opportunities is no longer adequate. Companies at this level can
harness AI to analyze a customer and compare its financial
performance with that of its closest competitors, identify and
prioritize gaps, and recommend possible solutions tailored to
both the customer’s needs and the supplier’s capabilities.
Trusted co-creator. At this level sellers go beyond helping
customers execute strategies and collaborate on formulating
them. Traditionally, this has been the crème de la crème of sales
approaches. But not every customer wants or will pay for the very
best. Furthermore, many companies lack the skills to be this type
of supplier. Because these relationships are so complicated, take
so many resources, and demand so much top-management
involvement, in the past companies could have only a small
number of them. They often require what’s known as “extraprise”
account management, in which multiple levels and functions on
both sides of the buy-sell equation communicate directly. CFOs
from the two parties, for instance, may talk about supply chain
issues and contingency planning. AI sales tools here would
involve multiple parties inside and outside the company.
Engineers at an engine manufacturer, for instance, might work
with the engineers at an aircraft maker to create a “digital twin” of
a jet engine that can predict maintenance needs, drawing into the
discussion a maintenance representative from an airline.
Interactions at this level are varied, intense, and forward-looking
—and exclusive.
Núria Madrid
Historically two-thirds of sellers have occupied the lowest three
relationship levels (transactional vendor, preferred supplier, and
solution consultant), while only 9% have reached the top level,
trusted co-creator. In general, the objective at the transactional
vendor and preferred supplier levels is to use AI to improve
efficiency and decrease costs while maintaining or improving
customer service. At the next levels, solution consultant and
strategic collaborator, the primary objective is more-effective
sales efforts, and AI tends to support more-sophisticated sales
professionals. At the trusted co-creator level, the goal is deep,
intensive collaboration with customers.
In addition to helping companies sell more, thoughtfully applied
AI can help move customer relationships to a higher level. But at
every level—and amid all the rapid changes in business,
technology, and how and where people work—some things
remain constant: Customers still will ask, What do you know
about me? What do you know about our business? What value do
you (singular and plural) add? Those are the questions sellers
need to answer to establish and elevate their relationships. AI
does not supplant that need or those questions; rather, it enables
better, richer answers.
Process Level
Advances in technology and access to new classes of information
and data have profoundly changed the way businesses need to
think about their sales processes. AI tools can improve
performance across the spectrum, which runs from ad hoc
process (at the low end) to informal process, formal process, agile
process, and customized process.
Ad hoc process. At this level each sales rep is allowed to do his or
her own thing. Reps receive relatively little or no sales training
beyond product or service information. Their feedback from the
field is not sought and often not even welcome. Sales support
means that the top marketing or sales executive or the CEO goes
out with the salesperson to close larger deals. Efficiency is the
watchword with this kind of process, and sales organizations can
improve it with simple AI that, say, scans emails and adds
prospects’ contact information into a CRM system.
Informal process. Here the company suggests a defined sales
approach to its reps and encourages them to use it but doesn’t
monitor whether they do or measure the results. In many cases
there is little sales training. More-experienced sales managers and
salespeople might even disparage the suggested process and
discourage its use: “That’s headquarters stuff that doesn’t work
here!” But CRM systems powered by AI can help salespeople work
more quickly, improve their client insights, measure results, and
ultimately better understand and refine their processes.
Formal process. At this level a company regularly enforces the
use of a defined sales process (sometimes religiously) and
conducts periodic reviews to see how effective it is. These
companies can see when win/loss rates and lead conversion ratios
begin to change, analyze the causes, and react. AI can help firms
do all this more accurately and with exponentially greater speed,
providing sales teams with immediate opportunities for
improvement.
Agile process. Companies at this level not only have a formal
sales process in place but also have CRM systems that continually
generate metrics on what’s happening in the marketplace. That
helps them react to external changes quickly. They can increase
their agility even more by leveraging analytics and business
intelligence and tapping the full capabilities of AI. These
companies may sense the winds of change at earlier and less
obvious stages (such as increases in the time various buy-cycle
steps take) and so can proactively minimize threats and take
advantage of opportunities.
Customized process. This is a level where companies build on
their agile AI experiences and begin to anticipate change rather
than just react nimbly to it. Predictive analytics are a key
capability needed to do that but cannot truly be implemented
until AI and machine learning are continually scanning sales,
marketing, and other data streams and identifying possible
threats and opportunities. The increased insights and flexibility
AI generates allow sellers to tailor messaging and proposals and
implement account-based marketing and account-based selling.
In addition to helping companies sell
more, thoughtfully applied AI can
help move customer relationships to a
higher level.
In the two lowest levels of processes—ad hoc and informal—AI
again can be used to increase efficiency. At the third level, formal,
it increases the effectiveness of activities such as coaching and
reporting. At the top two levels, agile and customized, AI mines
data to produce deep insights about customers’ realized and
unrealized needs.
As a rule, if your organization wants to have more-sophisticated
relationships with customers, it should also have higher-level
processes. For instance, to be strategic collaborators or trusted co-
creators, companies almost always need agile or customized
processes. (They also need the right data and analytics.) Again,
effective applications of AI can get them to the process level
where they want to be.
In the following section, we’ll explore three AI-for-sales
implementations. The first example, which offers lessons to
companies that would like to become strategic collaborators, is
about how Accenture built its own proprietary AI sales solution
from scratch. The second example, which provides a road map to
organizations that want to make their processes more agile, shows
how Honeywell worked with Aviso, a provider of AI-powered CRM
tools, to completely redesign its sales forecasting and pipeline
management processes and systems. The final example, which
outlines an approach that could help preferred suppliers become
solution consultants, describes SAP’s partnership with
Grapevine6 (now Seismic LiveSocial) to apply AI to client social
media data.
AI-Powered Sales Intelligence
Accenture’s Value Insights Platform (VIP) is a digital research
assistant powered by AI and machine learning. It assesses the
business imperatives of the firm’s current and potential clients by
deeply analyzing transcripts of their earnings calls. Before VIP’s
development, skilled analysts would spend a minimum of six
hours a call reviewing and extracting insights into companies’ key
priorities. VIP, in contrast, can process about 7,500 earnings call
transcripts in 480 minutes, or 3.8 seconds a call.
Because these tools’ efficacy increases
over time, first movers will gain a
meaningful and sustainable lead.
VIP’s benefits go far beyond time savings. After identifying
specific client priorities, such as a goal of realizing $22 billion in
savings over five years, the platform matches them with specific
solutions. Whether the client wants to optimize revenues, drive
operational efficiencies, or reduce its carbon footprint, VIP
reviews its performance relative to its peers, identifies solutions,
and calculates the financial impact of successfully achieving
those goals.
The AI’s predictive capabilities allow Accenture’s sales teams and
technical solution architects to take an unequaled value-based
selling approach, and the speed at which VIP provides supporting
insights gives Accenture a key advantage in the marketplace. VIP
now has roughly 20,000 users worldwide and has helped
Accenture generate more than $1 billion in sales over the past two
years.
AI for Forecast Management
In 2018, the technology company Honeywell began looking for a
tool that would help its business units improve the accuracy of
their sales forecasts, make more-informed decisions, and predict
short- and long-term performance. Honeywell, which builds
solutions for the aerospace, performance materials, and safety
and productivity industries, has multiple global teams and uses a
mix of sales tools and spreadsheets. It wanted an automated
system that would bring all its sales processes onto a single
platform, offer real-time insights, and boost efficiency.
After evaluating several partners and vendors, Honeywell chose
Aviso. In addition to building a unified global forecasting process,
Aviso customized the solution by giving Honeywell the ability to
analyze its deal portfolio and forecasts by geography, team,
product line, and business model (new or recurring). Honeywell
later added deal intelligence capabilities to the system to gain
insights on its sales pipeline, scoring its health and identifying
the top deals that could help reps beat their quotas.
Núria Madrid
Aviso’s tools allowed Honeywell’s sales teams to provide their
individual perspectives and predictions for success and get both a
combined forecast and forecasts for the discrete components they
were responsible for. Aviso also created a customized dashboard
for each salesperson providing a view of his or her own forecast,
accounts, communications, and more. Aviso’s tools helped sales
managers spot trends and deal opportunities and identify
obstacles to closing sales and take action to mitigate them.
Later Aviso introduced conversational intelligence, which
captures information from calls, web meetings, and email, as a
solution on its platform, giving Honeywell’s sales management
deeper insights into the status of all forecast deals. The results
have been overwhelmingly positive. Average yearly
improvements for Honeywell include $150 million in total
estimated revenue won and more than $1 million in CRM cost
savings at some divisions. Additionally, pipeline activity and
online interactions between reps and customers have grown by
more than 80% each, and the number of new deals has increased
by more than 70%.
AI-Powered Social Selling
Working with its sales and marketing teams, the German software
company SAP began using Grapevine6, a social engagement
platform. Acquired by Seismic in 2020, Grapevine6 has since been
renamed Seismic LiveSocial. Sales professionals connect their
social media accounts to it, and it produces two profiles for each.
The first covers the salesperson’s professional interests (goals,
markets, challenges), and the second covers the salesperson’s
personal interests (sports teams, hobbies, and the like).
LiveSocial’s AI engine then uses this information to search
through millions of media articles from more than 10,000 sources
a day and identify content that might be relevant for the
salesperson. Salespeople review the flagged content and, when
appropriate, share it with their clients. LiveSocial tracks the posts
that the customers interact with, providing insights into their
interests. The AI helps SAP position its salespeople as subject
matter experts who are on top of the latest changes in the market
—a tactic that other companies could use to raise their
relationships to the solution consultant level.
More than 10,000 sales professionals within SAP and its partners
are on LiveSocial. Many of them use it in conjunction with
LinkedIn Sales Navigator, a tool for finding and engaging leads.
Account executives who’ve harnessed both report significantly
higher closing rates and higher average deal sizes. They also see a
20% increase in their LinkedIn Social Selling Index (SSI) scores, a
measure of sales effectiveness. SAP account executives with the
highest SSI scores, in turn, have seen a 55% increase in their sales
performance, close 3.6 times more deals than their peers do, and
make deals that are 516% larger than those of their peers. In
addition, they’re 3.4 times more likely to achieve or exceed their
sales quotas. Overall, SAP has attributed €2 billion of its pipeline
and €1 billion of its closed deals to this social selling program.
Getting Started with AI for Sales
To get the most out of AI solutions, organizations must have the
necessary hardware, software, and processes in place. They also
need high-quality data to feed into AI tools and the right people to
leverage them.
What steps should organizations take to successfully implement
AI sales solutions? First, they should clearly articulate an AI
strategy: What are they trying to achieve? Second, they should
examine whether their structures support that strategy: Are teams
set up to achieve the AI goals? The AI tools will be part of an
integrated framework that includes people, processes, traditional
technologies, and knowledge: Are those components aligned?
Next, organizations must ensure that they have the right systems
—for data collection, performance management, training, and
communications.
AI-for-Sales Dos and Don’ts
To reap the highest benefits from initiatives that apply
artificial intelligence to sales… Do Optimize your
processes ...
The right culture is also key to success. People throughout the
organization need the skills to understand and apply AI tools—
starting at the top. AI initiatives must begin with senior
executives, but buy-in by involved employees is critical for full
adoption. The culture needs to support experimentation and
learning. The rollout process should be managed carefully,
employing the change management strategies that are necessary
for the success of any new initiative. It must involve goal setting,
benchmarking, and accountability. (See “AI-for-Sales Dos and
Don’ts” for additional tips for success.)
...
As more firms begin to implement AI-for-sales solutions, those
that don’t will find themselves losing ground. Because these tools’
efficacy increases over time, first movers will gain a meaningful
and sustainable lead. Eighty-one percent of our 2022 survey
respondents said that organizations without AI sales tools would
be at a “significant competitive disadvantage” or missing “an
important/key addition to their CRM.” However, that percentage
increased to 94% among respondents who had already
implemented or were currently implementing AI solutions. In
other words, people who have firsthand experience with AI-for-
sales solutions feel even more strongly about their importance.
Some companies have successfully applied AI and automation to
sales processes for almost a decade. Others have done it for even
longer. Those sales organizations that haven’t gotten started, or
have tried and failed, may get left behind for good.
The AI-for-sales landscape is packed with promise and challenge.
The technology is evolving rapidly, reshaping how sellers sell and
buyers buy. Trellix, Accenture, Honeywell, and SAP prove that
combining AI with well-defined sales processes and strong
customer relationships leads to high sales. No matter how deep
your bonds with your customers are or how complex your sales
processes, AI can help your company maximize its profitability.
Disclosure: SAP and Accenture have both been clients of Sales Mastery.
AHarvard
versionBusiness
of this article appeared in the November–December 2022 issue of
Review.
JD
Jim Dickie is a research fellow for Sales
Mastery, an independent research firm that
focuses on sales and AI-for-sales solutions.
BG
Boris Groysberg is a professor of business
administration in the Organizational Behavior
unit at Harvard Business School and a faculty
affiliate at the school’s Race, Gender & Equity
Initiative. He is the coauthor, with Colleen
Ammerman, of Glass Half-Broken: Shattering
the Barriers That Still Hold Women Back at
Work (Harvard Business Review Press, 2021).
BS
Benson P. Shapiro is the Malcolm P. McNair
Professor of Marketing Emeritus at Harvard
Business School.
BT
Barry Trailer is a cofounder of Sales Mastery.
Recommended For You
How Generative AI Will Change Sales
Using AI to Adjust Your Marketing and Sales in a Volatile World
How AI Is Changing Sales
PODCAST
How Generative AI Changes Creativity
