Automation Archives - Page 6 Of 10

Four Reasons Enterprises Aren’t Getting Full Value from their Automation CoEs | Blog

April 2, 2019

As we mentioned in our Enterprise RPA Pinnacle report, there are multiple benefits of having an automation CoE:

Facilitates best practices and skill development, which eventually help in faster scaling-up
Provides structure and governance to the automation program, e.g., clarifies roles and responsibilities of various teams involved
Enables optimization of software and license costs
Fosters sharing and pooling of resources
Develops strong cross-functional collaboration among stakeholders.

While many enterprises have established CoEs to overcome challenges and accelerate their automation journeys, not all have been able to extract value from them. Here are some common pitfalls that limit the true potential of an automation CoE.

No Enterprise-level Mandate to Drive all Automation Initiatives through the CoE

Automation CoEs work when there is a clear mandate from enterprise leadership. This directive helps ensure that the operating procedures and governance mechanisms are standard across the enterprise. Lack of it could result in the enterprise driving automation initiatives in pockets, leading to potential cost increases, lower license utilization, lower reusability of automation assets, and increased governance challenges.

Lack of Relevant Capabilities within the CoE

In its true sense, a CoE represents an entity with the capability to drive automation initiatives across the enterprise independently, with minimum oversight from outside. In many enterprises, however, automation CoEs lack the relevant skill sets – such as developers, project managers, solution architects, and infrastructure support staff – that are critical to driving these initiatives. Successful CoEs typically have three focus areas: day-to-day delivery, operational/tactical decision making, and strategic decision making/providing direction. Each layer, or focus area, has a unique set of roles and responsibilities that are critical to a smooth functioning CoE. Successful CoEs have an intentional focus on developing and nurturing in-house talent to strengthen the capabilities across the three layers. Many CoEs also bring in third-party specialists to accelerate learning. Our blog titled Driving Success in Your Automation Center of Excellence provides more details.

Loosely Defined CoE Roles and Responsibilities

The role of an automation CoE goes beyond just deploying bots into production. CoEs in best-in-class adopters of automation have evolved from executing solutions to empowering businesses across locations to drive initiatives on their own. For instance, the automation CoE in a financial services firm has established standard operating procedures (SOPs) for driving automation initiatives which include a well-defined approach for process selection, evaluation of ROI, talent impact, access to a library of reusable assets, etc. The CoE has created a platform through which business leaders can access these SOPs to evaluate opportunities on their own, and provides necessary governance and execution support, including talent and infrastructure.

As highlighted in Everest Group’s Smart RPA Playbook, the typical roles and responsibilities of automation CoEs include:

Providing training and education to develop talent
Approving all automation procedures before they are put into production
Assessing suitability of Smart RPA versus other Smart IT tools for use cases
Ensuring quality and compliance through well-defined standards, procedures, and guidelines owned and developed by the CoE
Driving the roll-out and implementation of Smart RPA projects, and ensuring coordinated communication with relevant stakeholders
Defining the roles, responsibilities, and skills sets required for driving automation across the enterprise, and regularly reviewing and optimizing them
Tracking success/outcomes in collaboration with operational teams so they can build, review, and refine the business case for scaling up.

Disconnect between Automation Strategy and Broader Digital Transformation Objectives

Outcomes achieved through automation initiatives are best realized when these investments are in line with the enterprise’s broader digital transformation objectives. Factors such as investments in other digital technologies (e.g., ERP transformation), changes in leverage of Global In-house Centers (GICs) or shared service centers, changes in vendor roles, etc., can have an impact on automation strategy. Automation CoEs need to closely align with enterprise strategy to realize maximum value.

Pinnacle Enterprises^TM – which we define as companies that are achieving superior business outcomes because of their advanced capabilities – have mastered each of these four requirements for successful automation CoEs. To learn more about how they’re maximizing value from their automation CoEs, please see our Enterprise RPA Pinnacle report.

Feel free to share your opinions and stories on how your automation CoE is evolving in its journey directly with me at [email protected].

Process Mining for Automation Gold | Blog

January 28, 2019

The process automation market is evolving in more ways than one. Many organizations are taking the next step of complementing Robotic Process Automation (RPA) with Artificial Intelligence (AI) solutions such as virtual agents and intelligent document capture. Others are looking deeper into their business functions with process mining and discovery software to scale automation and capture more returns from them.

Process mining and discovery solutions automate a part of automation itself. This is effectively mining processes for elusive gold opportunities for automation.

Process Miners

Process mining software has been around for a while and can be used for many purposes, but several vendors have made a name for themselves in the automation space, e.g., Celonis and Minit. These types of solutions use application logs to reconstruct a virtual view of processes. They discover business process flows and models, and provide process intelligence analytics. They can even suggest how to change a process using smart capabilities. The result is information that allows organizations to decide what process to automate next.

Some service providers have developed their own capabilities in this space as well. An example is Accenture, which uses process mining for automation as a competitive differentiator.

Valuable as it is, however, process mining also has its drawbacks. For example, it requires a lot of data. And if you want to find opportunities among processes that go across enterprise systems, you need to integrate the logs from these systems, e.g., build a data warehouse. Those of you who have built data warehouses know what a massive pain this can be.

Process Discoverers

While process miners can also do process discovery, several RPA vendors – including EdgeVerve, Kryon, and Nice – are offering new solutions. They’re using their desktop automation and action recording capabilities, complemented with AI, to capture and reconstruct what the human worker does, and then map and analyze the actions to identify opportunities for automation. Process discoverers do not require a load of application data, but they do come with their own challenges. For example, a recording may not capture the full set of relevant steps. And employees may have concerns around privacy.

The Art of the Possible

So, is it worth it to use process mining and discovering solutions despite their downsides and flaws? Yes, absolutely. But curb your enthusiasm, set expectations at the right level, and go for the art of the possible.

For example, there are many opportunities for automation within individual applications, without having to include processes that go across systems. And, you can use human intelligence to manually fill in the gaps and augment the findings of an automation discovery tool, even though doing so is going out of fashion.

With yet another category of software coming to the fore, enterprises would be right to feel that they are on a technology investment hamster wheel – there is no end to the cycle. After all, in recent years we have had the huge wave of RPA adoption. And today, in addition to competitive pressure to invest in AI-based automation, enterprises are having to evaluate process mining and discovery as well.

The good news is that automation can generate significant returns on investment. Our research and interactions with enterprises have shown this to be the case time and again. Process mining is another piece of the jigsaw, and it can help you find more automation gold.

Everest Group will be publishing a detailed viewpoint on process mining and discovery very soon. Be sure to keep an eye out for it, so you can mine it for gold.

Thanks to RPA, “Integration” is No Longer a Dreaded Word | Blog

January 16, 2019

Many enterprises that have used Robotic Process Automation (RPA) have seen the power of digital transformation, even if only in a small way through a few automated processes. The transformational value they experience is often a tipping point that whets their appetite for even more automation and deeper levels of application integration. But, this creates a quandary about how to maintain the array of automations. Ultimately, their success depends on the scope of the centers of excellence (COEs) that maintain their automations. Let’s explore further.

Getting the Wheel Spinning – Getting that Old-time Integration Religion

I believe that RPA has helped companies that previously held back from adopting newer technology solutions see the value of a digital mindset. These converts are now finding more opportunities for automation, and greater conviction in moving to digital-first operating models.

In short, something comparatively simple like RPA helps inspire confidence and vision.

The Ironic Corner to Turn – Moving beyond what Initially Made RPA so Enticing

Once this passion is unleashed, organizations come to fully appreciate that RPA is only one tool for automating operations. Many desire to transform their high volume, fast processes, and must confront the reality that surface-level RPA integrations are often not sufficient. The next steps towards more powerful automations often include integration via connectors and APIs.

The following exhibit reflects the diversity of systems which may now need to be integrated in a digital-first operating model world. (Spoiler alert: we’ll be writing a lot more about the Digital Capability Platform in the upcoming months.) And there are many ways to go about creating the needed integrations.

Some enterprises have cast aside the promise of surface-level RPAs, and now use their RPAs more through APIs. This is a bit ironic and worthy of a discussion by itself, but let’s get back to what happens as the types of automations proliferate.

Holding it Together – not Firing and Forgetting

One thing that all integrations – surface, APIs, or connectors – have in common is that they need maintenance. With surface-level RPA, you need to do a lot of robot maintenance when application layouts change. But all integrations, RPA included, require maintenance for other reasons as well. The biggest is the need to resolve data ambiguities, e.g., common customer names (think Jane Smith) with similar account types requesting a temporary address change. Which record should be updated? How can this correctly propagate across all the relevant systems and processes?

This is why a COE should be responsible for all types of automations, whether through surface or other integration methods. By looking across all automations, a COE can not only more accurately maintain the automations, but also identify anomalies and conceive new ways to structure interdependent automations. Of course, adding AI-based tools into the mix adds even more API connections to manage. But AI connections are far from the only ones that will need to be managed; the landscape will become more complicated before it simplifies (yes, I’m trying to be optimistic here.)

I can hear some of you saying that the COE should be an overall digital center of excellence. My answer is a big “no.” Digital is a far broader field that often involves major legacy transformation projects. Automation is clearly a part of digital, but it is operationally focused on the practical realities that come from modernizing processes that still primarily run on legacy systems.

This is a different mindset and a different set of competencies. As a result, it is best to keep a separate automation COE focused on the details of operational processes, while separately working towards the corporate digital objectives in a broader digital office. And that automation COE’s remit should be bigger than just RPA – it must deal with the combination of all types of automations that are enabling the operating processes.

Stop Trivializing AI: It Is Not Just Automation | Blog

January 15, 2019

AI is certainly being used to attempt to solve many of the world’s big problems, such as health treatment, societal security, and the water shortage crisis. But Everest Group research suggests that 53 percent of enterprises do not – or are not able to – differentiate between AI and intelligent automation and what they can do to help them compete and grow. This trivialization of AI is both eye opening and frustrating.

While it’s true that automation of back-office services is one strong case for AI adoption, there are many more that can deliver considerable value to enterprises. Examples we’ve researched and written about in the past year include intelligent architecture, front-to-back office transformation, talent strategies, and AI in SDLC.

It’s been said that “audacious goals create progress.” So, how should enterprises think more creatively and aspirationally in their leverage of artificial intelligence to extract real value? There are three ingredients to success.

Think Beyond Efficiency

Enterprises are experimenting with AI-driven IT infrastructure, applications, and business services to enhance the operational efficiency of their internal operations. We have extensively written about how AI-led automation can drive 10-20 percent more savings over traditional models. But enterprises have far more to gain by experimenting with AI to fundamentally transform the entire landscape, including product design customer experience, employee engagement, and stakeholder management.

Think Beyond CX

Most enterprises are confusing putting bots in their contact center with AI adoption. We discussed in an earlier post that enterprises need to get over their CX fixation and drive an ecosystem experience with AI at the core. Our research suggests that while 63 percent of enterprises rank CX improvement as one of their top three expectations of artificial intelligence, only 43 percent put newer business model among their top three. We believe there are two factors behind this discouraging lack of aspiration: market hype-driven reality checks (which are largely untrue), and enterprises’ inability to truly grasp the power of AI.

Think Beyond Bots

While seemingly paradoxical, humans must be central to any AI adoption strategy. However, most enterprises believe bot adoption is core to their AI journey. Even within the “botsphere,” they narrow it down to Robotic Process Automation (RPA), which is just one small part of the broader ecosystem. At the same time, our research shows that 65 percent of enterprises believe that AI will not materially impact their employment numbers, and that bodes well for their realization of the importance of human involvement.

And, what do enterprises need to do?

Be Patient

Our research suggests that 84 percent of enterprises believe AI initiatives have a long gestation period, which undoubtedly leads to the business losing interest. However, given the nature of these technologies, enterprises need to become more patient in their ROI expectation from such initiatives. Though agility to drive quick business impact is welcome, a short-sighted approach may straight jacket initiatives to the lowest hanging fruits, where immediate ROI outweighs longer term business transformation.

Have Dedicated AI Teams

Enterprises need AI champions within each working unit, in appropriate size alignment. These champions should be tech savvy people who understand where the AI market is going, and are able to contextualize the impact to their business. This team needs to have evangelization experts in who can talk the language of technology as well as business.

Hold Technology Partners Accountable

Our research suggests that ~80 percent of enterprises believe their service partners lack the capabilities to truly leverage artificial intelligence for transformation. Most of the companies complained about the disconnect between the rapid development of AI technologies and the slowness of their service partners to adopt. Indeed, most of these partners sit on the fence waiting for the technologies to mature and become enterprise-grade. And by then, it is too late to help their clients gain first-mover advantage.

As AI technologies span their wings across different facets of our lives, enterprises will have to become more aspirational and demanding. They need to ask their service partners tough questions around AI initiatives. These questions need to go far beyond leveraging AI for automating mundane human tasks, and should focus on fundamentally transforming the business and even creating newer business models.

Let’s create audacious goals for artificial intelligence in enterprises.

What has been your experience adopting AI beyond mundane automation? Please share with me at [email protected].

Can Indian Tier-2/3 Cities Fit the Bill for Digital Services Delivery? | Sherpas in Blue Shirts

December 18, 2018

India continues to offer an attractive service delivery location proposition for global companies, given its unique combination of a low-cost, scalable English-speaking talent pool, and the breadth and depth of available skills.

As the global digital services industry matures, and with increasing competition in the tier-1 cities, companies are looking to reduce the costs of talent and access additional untapped talent pools for digital services delivery.

Can tier-2/3 cities in India fit the bill? Let’s start by looking at the current state of digital services delivery in these cities.

Existing Landscape

Today, India is the largest destination for digital services delivery, with 75 percent of the market. Tier-2/3 cities in the country currently hold 14-16 percent of the market share, and we expect this proportion to grow by 15-20 percent in the next couple of years. Ahmedabad, Chandigarh, Coimbatore, Indore, Jaipur, Kochi, Lucknow, T-puram, and Vadodara are the top nine tier-2/3 locations, accounting for 55-60 percent of the digital services headcount in tier-2/3 cities.

Tier-2/3 cities are mostly leveraged to provide social & interactive (41-43 percent), cloud (21-23 percent), analytics (16-18 percent), and automation (10-12 percent) related services. When it comes to sophisticated digital technology services, such as cybersecurity, mobility, and Artificial Intelligence (AI), service providers still prefer tier-1 locations such as Bengaluru.

Now, let’s evaluate how tier-2/3 Indian cities’ value proposition stacks up against tier-1 cities.

What’s ahead for India’s Tier-2/3 Cities?

Here are some of the key findings from our recently published report, “Will Tier-2/3 Indian Cities Carve a Niche in the Digital Story?”

Tier-2/3 cities will continue to be leveraged predominantly as spokes to major hubs in tier-1 cities for the next two to three years
Because of a lack of skilled talent, delivery of advanced digital services such as machine learning, cyber security, and mobility from tier-2/3 cities will remain a distant dream for the next few years
An increasing number of enterprises will set up global in-house centers (GICs) or shared services centers for delivery of digital operations, due to increasing confidence and improvements in infrastructure quality
Reskilling/upskilling for digital capabilities will be paramount for companies operating in these cities
A few large service providers will invest in training talent, and benefit from early mover advantage by becoming distinguished employers in a less competitive market

To learn more – including the metrics around availability of talent, market maturity, cost of operations, business and operating risk environment, and implications for market participants including buyers, service providers, investment promotion councils, and industry bodies – please read our recently published report, “Will Tier-2/3 Indian Cities Carve a Niche in the Digital Story?.” We developed the report based on deep-dive discussions with leading shared services centers, service providers, recruitment agencies, and other market participants.

Have RPA Vendors been MARVELous? | Sherpas in Blue Shirts

December 7, 2018

The relationship between RPA vendors and their clients isn’t so different from the relationship between Marvel Studios and its fans.

Since the movie Iron Man hit the big screen in 2008, fans’ expectations of superhero films have skyrocketed. Despite the rising and evolving expectations, Marvel has satisfied its audience and has made a little pocket change in the process.

In a similar way, RPA buyers are expecting increasingly more from their RPA vendors. So, have RPA technology vendors been MARVELous in their customers’ eyes?

The Drivers

Our recent research study among 50 enterprise RPA buyers makes it clear that vendors have excelled in addressing their primary drivers, which are cost reduction and process optimization.

However, vendors didn’t score as high on secondary drivers such as improved customer experience, governance, and top-line growth. With increasing awareness about the potential impact of RPA beyond immediate cost and efficiency benefits, enterprises have started to view RPA as a primary contributor to their digital strategy, rather than a tactical measure.

Consequently, technology vendors should focus on continuously evolving their RPA solutions with a host of capabilities to help enterprise buyers achieve their strategic business outcomes.

The Capabilities

As to be expected, the buyers in our research study found their RPA vendors excelled in certain areas and had work to do in others.

The key strengths for those vendors who were identified as the Leaders as per our PEAK Matrix™ assessment on RPA included:

Customer support and service
Ease of use and robot development
Vision and strategy

Key improvement areas for Leaders included:

Responsiveness
Product vision and strategy
Product training and support

The X Factors

As there are so many RPA tools available in the market, each with its own strengths and weaknesses, it can be daunting for enterprises to select the right vendor for their unique needs. One critical part of the decision-making process is to focus on the X factors that are most important to their strategic agendas.

Our study found that factors including “ease of use and robot maintenance” and “scalability” highly correlate to buyers’ overall satisfaction levels. This is not surprising, as these are factors that buyers typically face issues with during RPA adoption. “Product vision and strategy” – and in some cases vendor expertise in a specific vertical industry or function – are also important buyer X factors.

While it’s clear that RPA vendors can do more to satisfy the needs of their customers – and that they’ll need to continually evolve their solutions – they have indeed been relatively MARVELous in delivering value and overall satisfaction to their buyers.

To learn more, please read our report “Buyer Satisfaction with RPA – How Far or Close is Reality From Hype.”

AI for Experience: From Customers to Stakeholders | Sherpas in Blue Shirts

December 5, 2018

Everest Group’s digital services research indicates that 89 percent of enterprises consider customer experience (CX) to be their prime digital adoption driver. But we believe the digital experience needs to address all stakeholders an enterprise touches, not just its customers. We touched on this topic in our Digital Services – Annual Report 2018, which focuses on digital operating models.

Indeed, SAP’s recent acquisition of Qualtrics and LinkedIn’s acquisition of Glint indicates the growing importance of managing not only CX, but also the digital experience of employees, partners, and the society at large.

AI Will Usher in the New Era of the Digital Experience Economy

Given the deluge of data from all these stakeholders and the number of parameters that must be addressed to deliver a superior experience, AI will have to be the core engine powering this digital experience economy. It will allow enterprises to build engaging ecosystems that evolve, learn, implement continuous feedback, and make real time decisions.

AI’s Potential in Transforming CX is Vast

Today, most enterprises narrowly view the role of AI in CX as implementing chatbots for customer query resolution or building ML algorithms on top of existing applications to enable a basic level of intelligence. However, AI can be leveraged to deliver very powerful experiences including: predictive analytics to pre-empt behaviors; virtual agents that can respond to emotions; advanced conversational systems to drive human-like interactions with machines; and even to deliver completely new experiences by using AI in conjunction with other technologies such as AR/VR, IoT, etc.

Digital natives are already demonstrating these capabilities. Netflix delivers hyper personalization by providing seemingly as many versions as its number of users. Amazon Go retail stores use AI, computer vision, and cameras to deliver a checkout free experience. And the start-up ecosystem is rampant with examples of cutting-edge innovations. For instance, HyperSurfaces is designing next-gen user experiences by using AI to transform any object to user interfaces.

But focusing just on the customer experience is missing the point, and the opportunity.

AI in the Employee Experience

AI can, and should, play a central role in reimagining the employee journey to promote engagement, productivity, and safety. For example, software company Workday analyzes 60 data points to predict attrition risk. Humanyze enables enterprises to ascertain if a particular office layout supports teamwork. If meticulously designed and tested, AI algorithms can assist in employee hiring and performance management. With video analytics and advanced algorithms, AI systems can ensure worker safety; combined with automation, they can free up humans to work on more strategic tasks.

AI in the Supplier and Partner Experience

Enterprises also need to include suppliers and other partners in their experience management strategy. Using predictive analytics to automate inventory replenishment, gauge supplier performance, and build channels for two-way feedback are just a few examples. AI will play a key role in designing systems that not only pre-empt behaviors/performance but also ensure automated course correction.

AI in the Society Experience

Last but not least, enterprises cannot consider themselves islands in the environment in which they operate. They must realize that experience is as much about reality as about perception. Someone who has never engaged with an enterprise may have an “experience” perception about that organization. Some organizations’ use of AI is clearly for “social good.” Think smart cities, health monitoring, and disaster management systems. But even organizations that don’t have products or services that are “good” for society must view the general public as an important stakeholder. For example, employees at Google vetoed the company’s decision to engage with the Pentagon for use of ML algorithms for military applications. Similarly, employees at Microsoft raised concerns over the company’s involvement with Immigration and Customs Enforcement in the U.S. AI can be leveraged to predict any such moves by pre-empting the impact that a company’s initiatives might have on society at large.

Moving from Customer to Stakeholder Experience

As organizations make the transition to an AI-enabled stakeholder experience, they must bear in mind that a piecemeal approach will not work. This futuristic vision will have to be supported by an enterprise-wide commitment, rigorous and meticulous preparation of data, ongoing monitoring of algorithms, and significant investment. They will have to cover a lot of ground in reimagining the application and infrastructure architecture to make this vision a distinctive reality.

What has been your experience leveraging AI for different stakeholders’ experiences? Please share with us at [email protected] and [email protected].

Upskilling and Reskilling: Is It Just Good L&D or Something Different? | Sherpas in Blue Shirts

December 4, 2018

Is upskilling and reskilling little more than a thinly disguised attempt by HR departments to rebrand Learning and Development (L&D)? The answer, as one practitioner pointed out at a conference in Poland, is “no.”

I recently presented to the Association of Business Services Leaders (ABSL) Chapter in Krakow, Poland about the talent acquisition challenges that digitization poses to Shared Services Centers (SSCs.) The argument runs roughly like this:

Robotic Process Automation (RPA) is replacing human agents in transactional roles, freeing up capacity in the workforce. This can mean lay-offs at worst, or unqualified internal candidates reapplying for roles at best
There is greater demand for people with new skills both technological (design thinking, robotics, autonomics, analytics) and soft (pattern-recognition, complex problem solving, leadership, intuition) than can be met by simply recruiting externally
As automation takes care of transactional processes, organizations can enhance the value of their brands and the service they provide by having more people in roles which emphasize first contact resolution, emotional intelligence, listening, etc.
This new value chain focuses on outcomes: people are measured against quality of outcome rather than throughput (for instance, a shift from average handling time to CSat), which in turn requires new management thinking around staff incentives, culture, and business model.

The data in the presentation was based on the Everest Group survey of 81 SSC leaders in Poland, the Philippines, and India, published earlier this year (see “Building a Workforce of the Future – Upskilling/Reskilling in Global In-house Centers.”)

So obvious was the message that emerged from the survey that one or two skeptics in the audience questioned why retraining that part of the workforce most affected by the trend of automation was even worthy of discussion. Is it not just good L&D practice? And surely survey respondents would not admit to anything other than good practice when asked the question?

Not quite true: there were survey respondents, albeit no more than 10 percent of them, who said that they were not planning to undertake upskilling and reskilling as a means of addressing talent shortages. A small majority, 58 percent, said upskilling/reskilling was the highest priority in addressing this same problem, while 10 percent, possibly the same nagging 10 percent, said it was a low priority.

The discussion continued after the presentation. Without experience as a practitioner, I wrestled with an explanation as to why this 10 percent stubbornly refused to fit the theory. Thankfully, the HR head of a Krakow-based SSC rode to my rescue and gave the answer.

This is the group, she said, which understands that reskilling and upskilling is indeed good L&D practice but remains wedded to external hiring of permanent and temporary staff. It is the group that fails to see that existing employees must be recognized as the key pool to meet scarce talent requirements in SSCs.

Her explanation, thankfully, echoed our contention that successful application of reskilling/upskilling to talent acquisition needs:

Senior leadership backing to ensure adequate resource and profile within the organization
Implementation of a skill-specific talent acquisition strategy to identify both critical areas of shortage and those most suitable for reskilling/upskilling
Quick roll-out of pilots in critical areas of shortage to build confidence and to learn
Breaking down of functional barriers and giving employees wider exposure to functional roles
A combination of effective duration and appropriate method (job rotation, on-the-job training, mentoring, peer-to-peer learning, and specialist external providers)
Clear communication of career paths, internal opportunity, incentive, and compensation
Patience and persistence!

She explained further. In her experience, the real difference between reskilling/upskilling as good L&D practice and reskilling/upskilling as a talent acquisition solution is simple. The talent acquisition solution approach is not considered aspirational, “something that HR does,” or nice to have. Rather, it is a strategic imperative.

How nice to have somebody who really knows what they are talking about answer a difficult question on my behalf!

Using AI to Build, Test, and Fight AI: It’s Disturbing BUT Essential | Sherpas in Blue Shirts

November 15, 2018

Experts and enterprises around the world have talked a lot about the disturbing concept of AI being used to build and test AI systems, and challenge decisions made by those systems. I wrote a blog on this topic a while back.

Disquieting as it is, our AI research makes it clear that AI for AI with increasingly minimal human intervention has moved from concept to reality.

Here are four key reasons this is the case.

Software is Becoming Non-deterministic and Intelligent

Before AI emerged, organizations focused on production support to optimize the environment after the software was released. But those days are going to be over soon, if they aren’t already. The reality is that today’s increasingly dynamic software and Agile/DevOps-oriented environments require tremendous automation and feedback loops from the trenches. Developers and operations teams simply cannot capture and analyze the enormous volume of needed insights. They must leverage AI intelligence to do so, and to enable an ongoing interaction channel with the operating environment.

Testing AI Biases and Outcomes is not Easy

Unlike traditional software with defined boundary conditions, AI systems have very different edge scenarios. And AI systems need to negate/test all edge scenarios to make sense of their environment. But, as there can be millions of permutations and combinations, it’s extremely difficult to manually assure or use traditional automation to test AI systems for data biases and outcomes. Uncomfortable as it may be, AI-layered systems must be used to test AI systems.

The Autonomous Vehicle Framework is Being Mirrored in Technology Systems

The L0-L5 autonomous vehicle framework proposed by SAE International is becoming an inspiration for technology developers. Not surprisingly, they want to leverage AI to build intelligent applications that can have autonomous environments and release. Some are even pushing AI to build the software itself. While this is still in its infancy, our research suggests that developers’ productivity will improve by 40 percent if AI systems are meaningfully leveraged to build software.

The Open Source Ecosystem is Becoming Indispensable

Although enterprises used to take pride in building boundary walls to protect their IP and using best of breed tools, open source changed all that. Most enterprises realize that their developers cannot build an AI system on their own, and now rely on open source repositories. And our research shows that 20-30 percent of an AI system can be developed by leveraging already available code. However, scanning these repositories and zeroing in on the needed pieces of code aren’t tasks for the faint hearted given their massive size. Indeed, even the smartest developers need help from an AI intelligent system.

There’s little question that using AI systems to build, test, and fight AI systems is disconcerting. That’s one of the key reasons that enterprises that have already adopted AI systems haven’t yet adopted AI to build, test, and secure them. But it’s an inevitability that’s already knocking at their doors. And they will quickly realize that reliance on a “human in the loop” model, though well intentioned, has severe limitations not only around the cost of governance, but also around the sheer intelligence, bandwidth, and foresight required by humans to govern AI systems.

Rather than debating its merit or becoming overwhelmed with the associated risks, enterprises need to build a governing framework for this new reality. They must work closely with technology vendors, cloud providers, and AI companies to ensure their business does not suffer in this new, albeit uncomfortable, environment.

Has your enterprise started leveraging AI to build, test, or fight AI systems? If so, please share your experiences with me at [email protected].

Telematics in Insurance – A Big Opportunity yet to be Fully Explored | Sherpas in Blue Shirts

November 14, 2018

Price competition used to define the competitive dynamics of the P&C insurance industry. However, as margins started squeezing with low interest rates and rising claims costs, it became imperative for insurers to focus on product differentiation in order to attract new customers and drive premium growth.

This is when usage-based insurance (UBI), an insurance product model where the premium varies according to the risk of claims that the insured’s policy-related behavior poses, started gaining traction. UBI is noteworthy as it offers a remarkable opportunity for insurers to deliver hyper-personalization and evolve from a product-centric to a customer-centric business mindset.

To date, the auto insurance segment has been the most aggressive adopter of the UBI model, which is enabled by the underlying telematics infrastructure. Telematics technology enables insurers to capture each customer’s driving data, which is then used to continually update the customer’s risk profile and compute the payable premium. Data collection devices have evolved from black-box to OBD-II dongles to in-built telematics units in automobiles and smartphones.

UBI’s Business Case is Strong; however, Sourcing Gets Complicated for Insurers

We expect the market for UBI to grow substantially at a CAGR of ~40 percent during 2018-2020, with an estimated 35-40 million UBI policies in force by the end of 2020. This is certainly impressive growth.

However, to launch UBI products, insurers must make substantial investments in connected devices and data infrastructure. Moreover, not all insurers have the scale, risk-appetite, investable capital, or technology expertise to make significant inroads into UBI. Thus, insurers are leveraging third-party vendors to support their telematics journey.

Yet, the vendor ecosystem is fragmented, making it challenging for insurers to determine what organization to partner with.

Here’s the breakdown of the three major categories of telematics vendors:

Telematics Service Providers (TSPs)

These have the capability to manage the entire value-chain, from telematics device sourcing to device deployment and maintenance to end-customer engagement to telematics data management. However, as a single TSP might not be able to provide access to all the underlying connected devices, insurers must pre-strategize their requirements for data depth and breadth. There have been cases where insurers have entered into partnerships with multiple vendors with varying competency to leverage connected devices and technology maturity.

Data exchanges

The core value proposition of this class of vendors lies in their access to huge volume of data and their data handling capabilities, which reduces the burden of data management at the insurer’s end. Players that have entered this market also have developed a modest understanding of the insurance sector, which enables them to provide risk assessment support to insurers. However, while data exchanges typically can augment insurers’ telematics journey, they cannot provide end-to-end support.

OEMs

OEMs have emerged as significant competitors to the other classes of vendors due to their direct control of the point-of-sale. As the telematics unit is prebuilt into the automobile, insurers do not have to worry about the entire infrastructure management of telematics devices. However, partnering with an OEM could also mean loss of revenue from value-added services.

Service Providers as the Orchestrator – Big Opportunity Waiting to be Capitalized

With each of the categories of vendors specializing in specific parts of the telematics value-chain, insurers face a big challenge in connecting with different parties for different values, and in managing the multi-vendor ecosystem.

This is where IT/BP service providers can enter the picture. To date, they have failed to establish a competitive differentiation for themselves in this market. However, considering they have a sound understanding of insurers’ businesses, operations, and IT systems, they could provide significant value as the orchestrator of this branched ecosystem.

They could look to source the best value from different classes of vendors by tying partnerships with select technology vendors across the ecosystem. Then, they could serve as a specialist to help insurer wrap their operations around telematics technology to drive product differentiation.

In this model, service providers could – potentially – offer an integrated value proposition that would involve: owning the implementation risk; providing value-added services such as risk assessment and customer management support; managing the complexity involved in coordinating with multiple classes of vendors; and assuming responsibility for the risks (e.g., business risk, technology lock-in, etc.) associated with engaging with niche firms.

This could be a win-win-win scenario, for insurers, end-customers, and providers.

How service providers ultimately decide to capitalize on the telematics opportunity remains to be seen. However, they should be cognizant of not frivolously trying to compete where their expertise does not lie, and instead leverage their strengths to make themselves most relev

Category: Automation