Creating Intelligent Infrastructure for Extreme Automation

A TSIA Member Publication
Creating Intelligent Infrastructure for Extreme
Automation
Streamlining and Automating the Delivery of Outcome-Based Services
A Joint White Paper from Wise.io and TSIA
EXECUTIVE OVERVIEW
In order to profitably deliver outcome-based services and drive sales volume, extreme efficiency is
required. While overhauls to people and processes are needed, this paper focuses on the technology
infrastructure required for extreme automation. Service executives should evaluate current
technologies with an eye toward prioritizing investments, particularly in data analytics and automation
tools, in order to improve margins on every sales and service activity and allow the scale necessary
to succeed.
CHANGING BUSINESS MODELS REQUIRE SERVICE AUTOMATION TO SUCCEED
Technology companies have historically made their money by selling technology assets to customers
and attaching traditional services like implementation and customization, support contracts, and
technical training. With the advent of subscription-based consumption models, the traditional revenue
streams of technology companies are now under increased pressure. As customers switch from
buying to renting, product transactions disappear and traditional product support contracts designed to
secure uptime are not attached. These changes are impacting the economic engines of product
companies.
In the book B4B: How Technology and Big Data Are Reinventing the Customer-Supplier Relationship,
1
TSIA executives define four distinct levels of technology solution providers. For a comprehensive
definition of these four levels, please refer to the book. Here is a shorthand definition for the four types
of technology providers:
• Level 1: Sells products.
• Level 2: Sells products plus classic product-attached services to implement and support those
products.
• Level 3: Sells products, attached services, plus services designed to help customers operate
and optimize the usage of technology solutions.
• Level 4: Sells business outcomes to customers that are enabled by technology capabilities.
TECHNOLOGY INSIGHT
TSIA-02077
September 15, 2015
PROFESSIONAL
SERVICES
FIELD
SERVICES
SUPPORT
SERVICES
EDUCATION
SERVICES
SERVICE REVENUE
GENERATION
MANAGED
SERVICES

2
17065 Camino San Bernardo, Ste. 200 | San Diego, CA 92127 | Tel. 1.858.674.5491 | Fax. 1.858.430.3571
© 2015 Technology Services Industry Association | www.tsia.com | Wise.io | www.wise.io
TSIA is seeing more product companies that historically only offered traditional product-attached
services (Level 2 offers) now experimenting with services that help customers accelerate their
adoption of technology (Level 3 offers) or commit to helping customers achieve specific business
outcomes (Level 4 offers). This shift in the service portfolio is amplifying a change in the economic
engines of product companies.
Figure 1: The Business Model Impact of B4B
Source: TSIA.
As seen in Figure 1, to make the shift to delivering outcome-based services, and to do so profitably,
technology providers must dramatically increase the units sold, both through new sales approaches
and increasing adoption within the existing install base, and dramatically lower the cost of delivering
each unit, increasing margins on products and services.
In order to lower the cost of products and product delivery, tech firms must embrace extreme
automation. This report outlines the components of the extreme automation infrastructure required for
successful Level 4 companies.
EXTREME AUTOMATION: THE DATA-DRIVEN TECH SUPPLIER
There are three primary components to constructing the extreme automation platform for outcome-
based services. While the exact data and analysis will differ by company and industry, the high-level
components are the same:

3
• Real-time data feeds. A number of data streams are required for a full analysis of customer
sentiment, loyalty, consumption, etc., coming from remote monitoring and consumption
tracking tools, service and sales interaction, and other mechanisms and applications.
• Analysis engine. Using data warehouses and analytics, the next step is analyzing the
independent streams of data to define and model customer success.
• Dashboards and intervention. Armed with this analysis, technology suppliers can be
proactive in identifying customers needing assistance to better adopt, consume, and receive
value from purchases.
Figure 2: Components of Extreme Automation Infrastructure
Source: TSIA.
The following sections will provide additional detail on each of these components.
Real-Time Data Feeds Required for Extreme Automation
Today’s enterprises are awash with data, with one source claiming that 15 out of 17 sectors in the US,
including high-tech, have more data stored per company than the US Library of Congress.
2
At the core
of extreme automation are capabilities to capture, analyze, and leverage this data to drive customer
success. There are many sources of data relevant to this analysis, ideally including real-time feeds
from:

4
• Product data. According to the 2014 Global Technology Survey, 37% of support services
members currently use proactive support to remotely monitor customer equipment. These
systems not only notify support or managed services technicians when an error occurs, they
also can transmit a steady stream of information regarding response times, space allocation,
data throughput, security risks, etc. This data can be invaluable for researching failing
components, as well as creating profiles of conditions leading to error conditions or outages.
• Usage data. How are customers consuming product features? What process flows are the
most common, and how long does it take on average to complete each process? What are the
typical consumption models for customers in 30/60/90 days? For 6/12/24 months? The ability
to identify customers who are ahead of the adoption curve (so you can capture their best
practices) and those who are behind the curve (so you can provide additional adoption
services) is critical.
• Service data. What are customers struggling with? What features are hard to navigate or
learn to use? What product defects are most impacting customer success? What product
enhancements could benefit rapid product adoption most? Analyzing service data not only
helps define and prioritize features for future releases, but it also provides valuable “voice of
the customer” information to improve your customer-facing processes and policies.
• Customer data. Demographic and profile information, such as company size, organizational
structure, industry, geography, experience levels, etc., can all be used to help inform analysis
of adoption and time to value. What profiles tend to be the most successful, and which profiles
require additional assistance to be successful? Are adoption curves and customer success
heavily influenced by demographics or profile data? Also in this category are survey
responses and verbatims, and any information collected as part of voice of the customer
programs.
• Marketing data. How are prospects finding and choosing your services? What marketing
content is best for a particular prospect? Content engagement, customer reviews, campaign
responses, and social influence can all be used to seek out, attract, and acquire the right
customers and capitalize on your marketing spend.
• Sales history. Beyond what customers purchased which products and services, sales data
includes service levels, any value commitments, desired outcomes, contract
expiration/renewal dates, key account influencers, and other information gleaned during the
sales cycle. This information can help identify critical sales information, such as likelihood of
renewal.
• Employee data. Internal data can also be analyzed to help drive customer success. Which
account reps have the highest ROI? Which sales reps are best suited to close a particular

5
type of business? Information from daily interactions inside the organization provides valuable
information to influence and automate processes, procedures, and workflows.
The Brains of Extreme Automation: Analytics
Once you have all the applicable data streams identified, the next step is leveraging an analytics
platform, or platforms, to analyze the data and start identifying patterns, best and worst practices,
adoption models, success profiles, etc. As an example, analyzing the accounts with the most rapid
adoption, highest satisfaction and/or loyalty scores, most profitable accounts, and accounts receiving
the most value can help understand what these accounts have in common. If specific profiles emerge,
these can be used in the sales process to identify prospects with a high likelihood of success.
Additionally, understanding the people and processes your most successful customers have in
common will identify Pacesetter practices you can document and share with other customers.
While analytics tools are trending toward real-time analysis of data wherever it is stored, it may be
necessary to consolidate your various data streams into a single data warehouse to make analysis
faster and to include a wider array of content sources.
Constructing the analytics layer of the extreme automation platform will definitely require expertise
outside of services, and if your IT organization does not have a data scientist on staff, outside
expertise will likely be required to allow comprehensive analysis from multiple sources. As well, some
outside expertise on modeling customer success, i.e., “success science,” would be beneficial.
As a primer, success science is the process of understanding what activities must occur at the
customer site for the customer to achieve specific outcomes. At a high level, there are three main
steps:
• Step one in success science is identifying the target business outcome.
• Step two is understanding what performance metrics correlate to the outcome occurring so
they can be monitored to assess progress.
• Step three is identifying activities that must occur to move the performance metrics in the right
direction.
These steps typically depend on ad hoc analysis that focuses on what happened in the past. An
emerging final step of success science is to automatically and continuously identify patterns in the vast
amount of data to predict how customers will behave in the future. Machine learning is a form of
predictive analytics that learns continuously as new internal and customer data comes in,
automatically adapting to changing behavior to improve processes. This approach to automation can
integrate seamlessly into your existing data sources and scales on its own to proactively and
preemptively improve business outcomes.

6
The analytics layer of the extreme automation platform is critical to understanding the important
metrics related to success, as well as identifying Pacesetter practices in how to move those metrics in
the correct direction. Once these correlations are made, customers can be guided toward changes to
the way they adopt and use technology to achieve the desired outcomes. The next section will
describe how this guidance can be provided using automation.
Dashboards and Intervention
The final layer of the extreme automation platform is dashboards and intervention. This layer is
focused on taking the learnings gleaned from the analytics layer and putting them into action in order
to speed adoption and value realization by existing customers, and to better identify prospective
customers with the highest likelihood of success. Though the exact solution will differ by company and
industry, several components should be included. This section will describe three required
components: real-time dashboards, proactive notifications, and integrations to systems of record.
Real-Time Dashboards
Too many service organizations are running large, complex operations using nothing more than
spreadsheets. While spreadsheets are useful in calculating project costs and profitability, the
information is typically only available at the end of the project or maintenance agreement. With real-
time data streams populating dashboards, companies can begin to identify problem customers and
projects and intervene early enough to change the outcome and ensure customer success.
Analytic-based dashboards are being introduced into more applications, with CRM and professional
services automation (PSA) being two examples. Again, the challenge is that these dashboards are
assuming all of the necessary information required for analysis is within the database of one specific
application. Custom work is required to include data from multiple applications into the dashboard.
While dashboards are a great way to identify customers lagging on adoption or value realization, or
projects missing milestones or margin targets, a key point is that the dashboard information is of no
value if there is no one watching. Part of the shift to outcome-based services means having staff
dedicated to monitoring and acting on dashboard information.
Proactive Notifications
One approach to ensuring critical alerts or dashboard information is not ignored is to use proactive
notifications to prompt the appropriate technician, account manager, or service executive when certain
thresholds are met or missed. A common example today from the support services world relates to
customer satisfaction scores. If a customer’s satisfaction rating on a transactional survey falls below a
defined threshold, an email notification is automatically sent to a supervisor to contact the customer
immediately and seek a resolution.

7
While notifications are helpful in escalating issues, they can also be used to notify employees who
routinely touch the customer regarding pending problems. For example, when a customer is nearing
the end of a cloud subscription or a maintenance agreement, every employee who touches that
account should receive a screen notification of the issue. Having call center and field support
personnel remind the customer of the pending expiration can help prompt the customer, but also gives
the employee an opportunity to transfer the call to an account manager who can discuss renewal
options.
Rules-based approaches are a step forward in automation, but as manual processes, they suffer from
a few common limitations. Specifically, they are:
• Inherently one-size-fits-all. All customers that look alike are treated alike, while in reality no
one customer or customer engagement path is ever exactly the same.
• Based on incomplete data. It’s nearly impossible to use the full range of available customer
data when rules are set up because of either human or system limitations.
• Costly and time-consuming to set up. It takes a lot of time and manpower to develop,
complete, and maintain rules-based approaches.
• Static over time. Assumptions typically stay static until the next refresh can be budgeted,
approved, and executed, which often happens annually, at best. Customer behavior likely
changes more frequently than once a year.
Self-learning automation approaches like machine learning can remove many of these limitations. By
continuously adapting to new data and behaviors and by taking into account a more complete view of
that data, machine learning enables notification and action to happen even more quickly and
accurately to help customer-facing teams generate better outcomes.
Integrations to Systems of Record
CRM systems introduced the concept of the “360-degree view of the customer” in the 1990s. The idea
is that every interaction, every detail, every issue facing a particular account should be visible from a
single location. TSIA surveys show that an average of 13 separate systems are used by TSIA member
companies to store customer information, and rarely are these systems integrated. With customer data
stored in multiple locations, the onus is on each employee to know where to look for the most current
information.
With so many data streams being analyzed, and so many details about customers emerging,
integration is a critical part of building an automated platform for outcome-based services. Consider
these integration benefits:
Customer service agents typically sort through dozens, if not hundreds, of templates when responding
to customer requests. In most cases, these responses and their modifications are only tracked as an
activity on the support ticket, rarely captured and analyzed to share within the Customer Service

8
organization or with other customer-facing teams like Sales, Marketing, and Product. Knowing which
templates are used most successfully to resolve customer concerns can help the Customer Service
team scale more effectively, especially as new agents ramp up during onboarding. In addition, Sales
and Marketing can benefit not only by knowing which responses are well-received by customers and
folding those into content marketing efforts, but also by getting insights into the types of customers
that demand higher levels of support and factoring that into the messaging and pricing conversations
with those types of customers.
A tight integration between support software, marketing automation, and CRM systems can automate
the engagements as they happen. Customer-facing teams will have access to a more complete
picture of the customers’ needs, and individual customer engagements and journeys can be
automated to deliver better outcomes for the customer and the company.
CUSTOMER USE CASE SPOTLIGHT
Integrating data across systems can ultimately increase revenue, like in the case of a large online
gaming company that uncovered hidden revenue by speeding up their support response time. Tens of
thousands of tickets come into the company’s support team each month. A majority of the tickets are
from non-paying customers. However, buried within the deluge are customers wanting to pay, but
experiencing purchase issues. By not quickly getting back to these customers, revenue was walking
out of the door.
To address this, the company integrated a machine learning application from Wise.io into their existing
support system. The application automatically used past support tickets to learn what a billing issue
ticket looks like, and then started predicting whether each new ticket represented a billing issue or
some other technical or customer issue. As a result, the company’s support team is able to prioritize
and address billing issues much faster, resolving thousands of purchase issues each month that
would have otherwise turned into missed revenue.
CONTINUAL AUTOMATION ACROSS SERVICE DISCIPLINES
This paper has described a framework for automating the infrastructure for outcome-based services.
But automation does not stop there. Commercial technology is available to automate many activities
within service disciplines, and support executives must understand what tools are available, and how
investing in automation can boost productivity and cut the costs of delivering services. Some
technology can be leveraged across disciplines, such as knowledge management and enterprise
collaboration, but there are many discipline-specific examples, such as:
• Support Services. Support has many avenues of opportunity to embrace automation with
intelligent infrastructure. Automatic ticket routing to the best agent can reduce response time.
Automatic template or macro recommendations can improve agent productivity. Support can
even reduce ticket volume by implementing automated responses to designated requests.

9
• Managed Services. Proactive monitoring tools are allowing managed service teams to
anticipate and react to equipment errors at a customer site, ideally before the customer even
notices there is a problem. Though the majority of companies tend to build this themselves,
commercial tools come with certified security mechanisms and support any and all operating
systems and hardware environments. These tools have a dramatic effect on customer
downtime and related satisfaction and loyalty scores.
• Professional Services. While PS teams have long used project management tools, many
companies, including large organizations, have yet to adopt the resource management
modules available in professional services automation suites. These tools can automate
scheduling consultants with an eye toward hitting goals for utilization rates and rate realization.
And, with an integration to talent management tools, you can even predict futures skills and
certifications required, and begin recruiting for new talent with these requirements in mind.
• Service Revenue Generation. One of the downsides to embracing cloud CRM solutions is
the lack of sophistication around complex entitlements and automating contract management
and renewals. A new breed of recurring revenue management products are now available to
fill this void, offering rich dashboards for expiring contracts and likelihood of renewal, allowing
renewal teams to identify struggling customers early and help them find value.
While these examples illustrate automation used to reduce labor, companies should always be on the
lookout for automation plays that can remove labor entirely. Some emerging examples include:
• Automatic response. Support organizations experience continued pressure to cut costs and
improve efficiency in the face of rising ticket volumes and customer expectations. By using
automation to respond directly to customers, support teams can often take a significant portion
of the ticket volume out of their agents’ queues. In turn, support teams can often flatten out
their headcount growth and have their existing agents spend time working on more
complicated customer issues rather than triaging and resolving the large volume of simple
requests.
• Intelligent scheduling. Many areas of service operations today require considerable manual
activity to schedule service projects and appointments, particularly for field services
technicians and professional services consultants. More sophisticated scheduling tools are
emerging that leverage analytics as part of scheduling to reduce or eliminate manual
intervention. As an example, innovative field service scheduling tools now allow field techs to
accept the next best appointment based on location, experience, and available parts, with no
call center agent required for dispatch. Additionally, self-service scheduling tools are available,
allowing customers to pick the best appointment time, with the information “drip-fed” directly to
the technician’s calendar with no human intervention.

10
• On-demand learning. A big trend in education services is shifting away from labor-intensive
classroom training toward on-demand learning using learning management systems.
Customers can subscribe to particular modules or an “all you can eat” consumption model for
a specified period of time. Instead of requiring customers to fly across the country to attend a
five-day class, they can watch instructional, interactive videos designed in short chunks on
targeted topics or concepts. Moving toward this just-in-time approach, more customers will
have access to critical training to speed consumption, paying subscription fees for on-demand
content instead of live instructors.
CONCLUSION
As John Donne wrote, “No man is an island,” and the same can be said for technology providers.
While new profiles are emerging, the largest percentage of technology firms today are still Level 2
providers. Enormous effort will be required to shift corporate people, processes, and technology
toward Level 3 and Level 4. Companies will not be able to make this shift successfully, and in
sufficient time, without outside expertise. TSIA Research is always challenging member companies to
reevaluate core vs. context, meaning any part of your operation that does not provide market
differentiation might be better outsourced to a service provider who can perform the function as well or
better than you can, often for a much lower cost.
Wise.io is transforming the business of customer success by applying the power of machine learning
across the full customer life cycle. Driven by a cloud-based machine learning engine, the company’s
predictive applications automatically learn from past patterns in order to predict future behavior,
helping users to optimize how the acquire, monetize, support, and retain customers. Companies
ranging from Silicon Valley start-ups to the largest Fortune 500 corporations rely on Wise applications
to better engage with their customers.
2150 Shattuck, Suite 525
Berkeley CA 94704
www.wise.io
ENDNOTES
1
Wood, J.B., Todd Hewlin, and Thomas Lah. 2013. B4B: How Technology and Big Data Are Reinventing the
Customer-Supplier Relationship. San Diego, CA: Point B, Inc.
2
“Big Data: The Next Frontier for Innovation, Competition, and Productivity.” May 2011. McKinsey & Company.

Creating Intelligent Infrastructure for Extreme Automation

More Related Content

What's hot (20)

Viewers also liked (18)

Similar to Creating Intelligent Infrastructure for Extreme Automation (20)

Recently uploaded (20)

Creating Intelligent Infrastructure for Extreme Automation