Category: Automation/RPA/AI

The Evolution of the Automation CoE Model – Why Many GBS Centers Are Adopting the Federated CoE Model | Blog

Automation CoEs in Global Business Services (GBS) centers or Shared Services Centers (SSCs) have evolved over time. Mature GBS adopters of automation have made conscious decisions around the structure and governance CoEs, evolving to extract maximum value from their automation initiatives. Some of the benefits they have hoped to gain from the evolution include:

  • Faster scaling
  • More efficient use of automation assets and components, such as licenses and reusable modules
  • Better talent leverage
  • Greater business impact

The typical CoE model evolution

CoE models generally evolve from siloed model to centralized and then to a federated:

Siloed model – kick starting the journey

Most GBS centers start their automation initiatives in silos or specific functions. In the early stages of their automation journeys, this approach enables them to gain a stronger understanding of capabilities and benefits of automation and also to achieve quick results.

However, this model has its limits, including suboptimal bot usage, low bargaining power with the vendor, lower reusability of modules and other IP, limited automation capabilities, and limited scale and scope.

The centralized model – building synergies

As automation initiatives evolve, enterprises and GBS organizations recognized the need to integrate these siloed efforts to realize more benefits, leading to the centralized model. This model enables benefits such introducing standard operating procedures (SOPs), better governance, higher reusability of automation assets and components, optimized usage of licenses and resources, and enforcement of best practices. This model also places a greater emphasis on a GBC-/enterprise-wide automation strategy, which is lacking in the siloed model.

However, this model, too, has limitations, suffering slow growth and rate of coverage across business units because the centralized model loses the flexibility, process knowledge, and ownership that individual business units bring to the bot development process.

The federated model – enabling faster scaling

The federated model addresses both of the other models’ limitations, enabling many best-in-class GBS centers to scale their automation initiatives rapidly. In this model, the CoE (the hub) handles support activities such as training resources, providing technology infrastructure and governance. Individual business units or functions (the spokes) are responsible for identifying and assessing opportunities and developing and maintaining bots. The model combines the benefits of decentralized bot-development with centralized governance.

The federated model has some limitations, such as reduced control for the CoE hub over the bot development and testing process, and, hence, over standardization, bot quality and module reusability. However, many believe the benefits outweigh the drawbacks.

The three CoE models are described in the figure below.

Automation Adoption in GBS centers and the Rise of the Federated CoE Model

The table shown below shows how the three models compare on various parameters.

Comparison of salient features benefits and limitations each CoE model

Why GBS organizations are migrating to the federated model

There are several reasons why GBS centers are moving to the federated model, as outlined below.

  • The federated model helps to better leverage subject matter expertise within a business unit. With bot development activity taking place within the BU, the federated model ensures better identification of automation opportunities, agile development, and reduced bot failures
  • The federated model leads to efficient resource usage. Centralization of support activities ensures: efficient use of resources, be they human, technology, reusable modules, licenses, etc.; standardization; and, clear guidance to individual business units
  • The federated model facilitates development and sharing of automation capabilities and best practices, which helps in the amassing of standardized IP and tacit knowledge important for rapid automation scaling

Federated model case study

A leading global hardware and technology firm’s GBS center adopted the federated CoE model, which houses the CoE hub, in 2017. In the three years since, it has grown to over 400 bots across more than 20 business units in a wide variety of locations, and saved more than $25 million from automation initiatives. The CoE hub has also successfully trained over 1,000 FTEs from technical and business backgrounds on bot development. As a result, firm-wide enthusiasm and involvement in the GBS center’s automation journey is high.

Transitioning to a federated CoE model has helped many GBS programs scale their automation initiatives rapidly. For more details, see our report, Scaling Up the Adoption of Automation Solutions – The Evolving Role of Global In-house Centers or reach out to Bharath M  or Param Dhar for more information on this topic.

The UK’s Perfect Storm: Brexit, EU Workers Returning Home, IR35 Changes, and Coronavirus | Blog

Businesses in the UK are facing a spate of challenges; there’s the specter of new Brexit-driven red tape on trade, a staffing shortage as some EU workers are returning to their home countries, and UK changes to the IR35 contract worker tax legislation, which is making it very difficult for companies to hire contractors. A Coronavirus pandemic could be the final straw that breaks businesses’ backs. Let’s face it – there is a perfect storm ahead.

With Brexit and the EU trade negotiations still going on, there is little certainty about the red tape that businesses will face in order to trade with each other across the English Channel. Yet, with the transition period set to end on 31 December 2020, there is little time for businesses to prepare for whatever the new trade requirements may ultimately be.

Because adherence to the as yet unclear regulations will increase businesses’ workloads, a natural response would be to hire more staff. But unemployment is at record low, and many skilled EU workers are leaving the UK and returning to their home countries. Furthermore, the UK Office of National Statistics (ONS) reports that EU immigration to the UK is at an all-time low.

The HMRC’s new IR35 rules, which come into effect in April 2020, are exacerbating the problem. Many companies have had to adopt no-contractor hiring policies and cannot fill temporary vacancies. They are already feeling the impact of the regulation. If they can’t hire staff or contractors, where are companies going to find resources to handle the extra workload of trade red tape?

Additionally, widespread cases of the Coronavirus could lead to prolonged periods of sick leave, further reducing the number of staff who are available to help with the increased workload of trading with the EU. While cases are still far and few between in the UK, the impact of the spread of the disease in China has been great. Empty offices and factories in Chinese cities and manufacturing heartlands are already leading to a shortage of parts for cars and other products that are much in demand in the UK.

Clearly, UK businesses are facing a perfect storm.

Investing in digital and Intelligent Automation (IA) technologies can help them tackle some red tape issues, particularly if they use IA for what I call Red Tape Automation (RTA). This could be automation of compliance form-filling and reporting requirements, weights and measure conversions, or making changes to transaction or product-related data and synchronizing them across multiple systems such as those used for sales and revenue to record value added taxes and other duties. Companies that trade with both EU and non-EU countries could automate the red tape for all of those, using rules engines to fill in the right forms and apply the correct rates.

IA is not a perfect solution, as people will be needed to implement technology, and there is a growing talent shortage. Nonetheless, UK businesses will be well served by investing in learning the art of the possible with IA. While the final details of any trade deals with the EU, or new deals with the rest of the world, will not be known for a while, knowing how to implement the requirements quickly using IA can help them weather the impending storm.

For more information about IA, please check out Everest Group’s Service Optimization Technologies research.

Visit our COVID-19 resource center to access all our COVD-19 related insights.

Appian’s Jidoka Acquisition Sets the Scene for the RPA Market in 2020 | Blog

Appian announced its intent to acquire the Spanish vendor Novayre Solutions SL and its Jidoka RPA platform on January 7. With this acquisition, Appian, best known for its low-code process management and orchestration software, will be able to offer extensive automation capabilities natively, while it did so previously with partners’ software such as Blue Prism and UIPath.

So, what does the acquisition mean for the market?

Why the acquisition?

Our estimates show that the RPA third-party software market is expected to grow by 80 percent to reach $2.5 billion this year. With this phenomenal growth rate, it’s not surprising that non-RPA companies want a slice of the pie.

Appian has been active in this market for a while and has benefitted from many new clients thanks to its partnerships with RPA vendors. It is also a reseller for Blue Prism and has experienced growing demand for RPA first-hand through that channel.

In addition, technology giants are increasing their activities in this market. SAP acquired Contextor back in 2018. And most recently, Microsoft announced UI flows to add RPA capabilities to Microsoft Power Automate (previously Microsoft Flow). It combines digital process automation (DPA) via APIs with UI-based automation. Pega is another competitor that has also invested in this market; it took over OpenSpan back in 2016.

Why Jidoka, and what about the partners?

We have assessed Jidoka as part of our RPA Technologies PEAK Matrix for a number of years and most recently positioned it as a major contender in our 2019 assessment. Jidoka is a Java-based platform where robots are designed and managed by a web-based console. There is a design studio for workflow and orchestration of robot operations. A console centralizes monitoring, audit, and exception handling features along with secure user permission and authorization capabilities. It has proprietary image recognition technology, Hawk Eye, to support Citrix automation. The platform offers capabilities such as auto-scaling of robots, a secure credentials vault, roles-based access controls, execution logs, audit trail, robot performance analytics, and ROI calculator. It also offers a chatbot capability that is available from the console. Real-time human-robot collaboration is provided via chat interface from the console (and Google Home,) the Jidoka mobile app (voice and chat,) and via IoT devices.

Appian intends to rebrand the product as Appian RPA. It will turn it into a low-code environment and integrate it with its own solutions to be offered on the cloud on a competitively priced subscription basis. While growing in Spain and Latin America, Jidoka has limited presence in other geographies. This is something that Appian can address with its presence in major tech markets.

As for its partnerships, Appian is keen to keep them going and offer clients choices. It remains to be seen how partners such as Blue Prism and UiPath will react to this news. It is not unusual for partners to go for co-opetition. For example, last year Blue Prism announced an Intelligent Document Processing (IDP) solution called Decipher, but has maintained its partnerships in the IDP segment, e.g., with Abbyy.

What does it mean for the market?

We have been expecting M&A activity in this sector to increase with market maturity and as RPA becomes a key tool for process efficiency and productivity. RPA is also commoditizing, and the fact that Appian is acquiring a very small vendor shows that entry into the market is not expensive. The news of this acquisition could encourage other tech companies, particularly those in the process management and orchestration space, to act too. There are many small RPA vendors with good offerings. The big RPA players with their current large valuations could suffer if a wave of acquisitions materialized and bypassed them; but at the same time, they have an awful lot of customers and a huge global footprint among them. Furthermore, private equity investors continue to invest in the market, as evidenced by Automation Anywhere’s last round of funding. This market remains buoyant and dynamic.

With Microsoft getting into the RPA business, all vendors have to up their game to remain competitive.  As for the RPA scale challenge that many enterprises are facing, vendors are working on this with new, improved offerings in the areas of robot management and controls, ease of use, and increased robot resiliency. With its existing and new capabilities, Appian will be well placed to address the scale challenge to make RPA adoption and operations smoother and, in so doing, edge ahead of the competition.

Is Your Shared Services Center Driving Automation Across Your Enterprise? | Blog

Over the past few years, automation has become an integral part of Shared Services Centers’ (SSCs) growth and evolution. Whether large or small, whether onshore, nearshore, or offshore, SSCs – what we refer to as Global In-house Centers (GICs) – have made strong progress in adopting automation solutions.

Some have only dipped their toe into basic RPA. Others have moved ahead with more advanced automation technologies like machine learning and artificial intelligence. And a handful have started emerging as key strategic and revenue-generating entities for their parent companies. These GICs have built scaled delivery teams with strong domain knowledge around the implementation of automation solutions. There are multiple instances of GICs housing the global automation Center of Excellence (CoE) and driving initiatives across the enterprise. Aggressive adopters have moved beyond automating processes within the center and are now supporting process automation across locations and businesses. And they’re increasingly leading the design and execution of automation strategy, and are influencing decisions on go/no-go opportunities.

Everest Group’s recently published report Scaling Up the Adoption of Automation Solutions – The Evolving Role of Global In-house Centers discusses the key adoption trends and challenges in the GIC and automation space.

Let’s take a quick look at the four key trends.

4 trends GIC

Solutions and support

Some mature GICs have developed multiple offerings to support different businesses. Typical offerings include advisory support, platform or infrastructure support, and end-to-end implementation support. For instance, the India GIC of a leading European insurance firm provides bot infrastructure support to the company’s Singapore entity. With this set-up, the Singapore-based team didn’t have to invest in its own infrastructure to gain full access to the bots’ capabilities.

The talent ecosystem

From developing in-house automation talent to managing vendor resources, SSCs are making major strides in the talent management space. As part of their talent management strategy, many best-in-class GICs are investing heavily in building in-house talent, especially for AI-based solutions. This includes developers, data scientists, and project managers. These GICs are also investing significantly in upskilling/reskilling programs for their resources, and are strongly emphasizing education and awareness of automation’s capabilities and benefits. Some GICs are also training their business/operations resources on automation skills; this helps them scale-up faster.

CoE roles, governance mechanisms, and structures

Many GICs are upgrading their CoE model, roles, and responsibilities as they progress along their automation journey. Many successful centers are moving towards the federated hub and spoke CoE model, wherein the GIC houses the CoE hub and the functions have their own automation team (spokes.) The federated model enables rapid scalability and better opportunity identification than centralized CoEs. But, with either model, there are some pitfalls to avoid. Our blog titled Four Reasons Enterprises Aren’t Getting Full Value from Their Automation CoEs details what they are.

In-house automation platforms

Building on their understanding of automation capabilities, some mature GICs have started exploring the use of custom-built in-house platforms to run automations. While in most cases these are for attended RPA bots, some best-in-class SSCs have developed platforms using advanced technologies such as interactive virtual assistants (IVA) and machine learning. There are even a few examples of GICs adopting a 100 percent in-house development model, meaning no third-party vendor support. While we expect GICs to continue exploring in-house automation tools, we don’t expect that these will replace the use of third-party vendor products in the near future.

What GICs have accomplished over the past few years in scaling up the adoption of automation solutions across businesses and locations is just the tip of the iceberg. Going forward, they are likely to build on this foundation and penetrate deeper into the enterprise with ever more complex automations.

To learn more, please read our report — Scaling Up the Adoption of Automation Solutions – The Evolving Role of Global In-house Centers – or contact us directly at Bharath M or Param Dhar.

Smartphones and 5G are the Keys to AR/VR Success | Blog

A Goldman Sachs Research report published in January 2016 stated that venture capitalists had pumped US$3.5 billion into the augmented reality (AR)/virtual reality (VR) industry in the previous two years and that AR and VR have the potential to become the next big computing platform.

But a recent PwC MoneyTree report stated that funding for augmented and virtual reality startups plunged by 46 percent to US$809.9 million in 2018, as compared to 2017. Indeed, multiple startups in the space shut down in 2019 because they haven’t been able to materialize their claims and have been unsuccessful in making the technology economically viable to the masses.

It’s not just startups that are throwing in the towel on their investments. For example, a dwindling user base drove Google to shut down its Jump VR platform in June 2019, and Facebook-owned Oculus is closing its Rooms and Spaces services at the end of this month.

AR VR blog graphic

And the startups cited above sunk nearly $550 million in investments when they shuttered their doors.

So, what’s going wrong?

Problems with present-day AR/VR

New technologies, particularly those for the consumer market, invariably need hype to succeed. But, despite all the buzz around how AR/VR can change the way consumers interact with commercial and non-commercial entities (like healthcare providers and educational institutions), multiple problems are getting in the way of mass adoption.

  1. Cumbersome hardware: Despite 2-3 generational improvements, the hardware for these technologies remains bulky and difficult to set-up or use. More research is needed to bring advanced optics and computation of head-mounted displays (HMD) to a usable level
  2. High cost: Nearly all the standalone AR HMDs cost over $1000, and those for VR are over $150. At these price points, the vast majority of purchasers are technology enthusiasts and novelty buyers
  3. Poor content: While the premise of buying an HMD is to consume and interact with content in an engaging way, the flood of poorly designed experiences hardly makes the case for purchasing it, even for those who can afford it
  4. Selling an idea, instead of a product: This is perhaps the biggest reason for the slew of closures in recent months. While AR and VR both have compelling use cases, the entrepreneurs and enterprising providing the products and platforms promised the sky and underdelivered on expectations.

So, what should enterprises do to change the narrative behind and fate of AR/VR?

Here are our recommendations.

Focus on developing smartphone-based AR

AR adoption is far outpacing VR adoption, not only because it adds to users’ reality rather than replacing it, but also because smartphones make its cost much lower for consumers. Indeed, smartphone-based AR has gone mainstream in the retail and gaming spaces; examples are IKEA, Nike, Nintendo, and Sephora, all of which have deployed applications for interactive experiences. The buzz will stay alive, and the uptake will continue to grow as an ever-increasing number of developers incorporate AR elements into their applications.

Embrace 5G with open arms

Fifth-generation (5G) wireless promises to bring high bandwidth and reliable low latency in data communications. Along with the proliferation of edge computing, 5G will help move processing-intensive tasks closer to the edge of the network and content closer to the user. In the near future, telecom operators could provide dedicated network slices for AR/VR applications, greatly reducing network latency. By enabling faster processing and increased proximity to content, 5G will boost the overall user experience. And this will lead to increased adoption.

But, before going all in, enterprises should partner with communication service providers to test 5G PoCs for AR/VR. Doing so will help them better prepare for scaled adoption as HMDs become less cumbersome.

By placing hype before substance, AR/AV providers created the current low-growth environment. We believe that focusing on smartphone- and 5G-based AR/VR will increase both investor confidence and customer adoption.

What is your view on the AR/VR space and the emergence of 5G as a savior? Please share with us at [email protected], [email protected], and [email protected].

Demand for Next-Gen Services Defining Location Strategies | Blog

Regulatory uncertainty, technological disruption, talent challenges, and a host of other issues have all played significant roles in enterprises’ and service providers’ location strategies for global services delivery over the past couple of years.

The deep-dive analysis we conducted on enormous volumes of 2018 data to develop our Global Locations Annual Report 2019 made it clear that five key trends came into play in 2019, and will continue into 2020:

  • Increased focus on digital and R&D/engineering services
  • Increase in nearshoring
  • Slowdown in headcount growth
  • Increase in onshoring by service providers
  • Growth in emerging locations.

Here’s a quick look at each of these trends.

Digital and R&D/engineering services continue to dominate

Enterprise demand for digital services and the associated R&D/engineering services compelled most global service providers to set up innovation centers and COEs to keep up with the changes in the digital landscape. And there was a significant rise in the number of R&D/engineering and digital service delivery centers – especially in APAC and nearshore Europe – as providers vie to develop data-driven, intelligent, and robust systems using automation, cloud, and AI-based capabilities.

DC3 1

Global services delivery is increasingly being characterized by nearshoring

In a move to rebalance and optimize their existing locations portfolio and comply with data protection mandates, both enterprises and service providers are marginally shifting from offshore to nearshore locations. Nearshore Europe experienced the greatest increase in headcount and new center setups in 2018 due to the availability of complex skills, proximity to customers in Western Europe, increased regulatory oversight, and demand for multi-lingual support.

Poland, Ireland, and Scotland will continue to dominate the global services landscape in nearshore Europe, followed by Ukraine, the Czech Republic, and Romania.

DC2 1

Global services headcount continues to grow, but modestly

Increasing use of automation for low complexity, high volume services is having a considerable impact on the talent landscape. While growth in digital services will lead to newer job and skill profiles, the headcount required for newer digital jobs will be significantly lower than that required for low complexity jobs, and the growth will be slower due to technological advances and the shortage of talent for new-age technologies.

DC1 1

Service providers continue to grow in onshore geographies

Leading service providers have been continuously growing their presence in onshore geographies. This is in large part due to increasingly stringent data protection laws and mounting pressure from clients to have local delivery centers. The United States and continental Europe continue to remain the destination of choice for setup activity across onshore locations. The lion’s share of the work delivered from these onshore centers is in IT services.

We expect the United States to continue to grow in the wake of uncertainty around visa regulations and increased pressure from clients to have local delivery centers for ease of coordination, better alignment/training, and promoting customer intimacy. And, we also expect growth in digital services to push providers to continue to expand in other onshore locations – such as Belgium and Switzerland – due to availability of skilled talent and the ability for extensive collaboration with Europe-based clients.

Growth in emerging locations for global services delivery

While use of the traditional delivery locations continues to grow, other locations are picking up steam, including:

  • Jamaica continues to grow in setups for voice services
  • Ghana and Kenya are being leveraged to support the East and West Africa regions
  • Israel is growing significantly for delivery of R&D/engineering and high-end IT services
  • Lithuania is also growing as a destination for delivery of IT (largely digital) and R&D/engineering services.

To learn more about the dynamics shaping the global services locations landscape, please read our recently published report, “Global Locations Annual Report 2019: Demand for Next-Gen Services Defining Locations Strategies.” We developed the report based on deep-dive discussions with the regional investment promotion bodies, leading shared services centers, service providers, recruitment agencies, and other market participants.

What’s Your Company’s Digital Ethics Score? | Blog

I marvelled at the passion demonstrated by the London Extinction Rebellion activists while I attempted to make my way to the Digital Agenda Power & Responsibility Summit at the British Library on 9 October.

During the Summit itself – while listening to presentations delivered by eminent speakers including Matt Warman MP, Minister for Digital and Broadband at DCMS; Sana Khareghani, Head of UK Government Office for AI; Russell Haworth, CEO, Nominet; Cheryl Stevens MBE, Deputy Director for Trust & Identity at DWP; Jacqueline de Rojas CBE, President, techUK; and Caroline Criado Perez OBE, award-winning author of Invisible Woman and activist – it struck me that consumer disillusionment with unethical applications of technology could lead to its own type of activism in the form of product and service boycotts.

Read my blog on Digital Agenda

Digital Experience Platforms: An Idea Whose Time Has Come | Blog

In today’s increasingly competitive environment, enterprises need to package their offerings with superior and memorable experiences to remain relevant. They need to streamline their efforts to deliver a unified and seamless digital experience to stakeholders. While they’ve attempted to achieve this with point solutions such as CRM platforms, campaign management tools, and other experience management solutions, their disjointed and incompatible portfolios have often created more problems than solutions.

Enter the Digital Experience Platform (DXP)

In response to an obvious need, vendors including Adobe, IBM, Oracle, and Salesforce have created a digital experience platform or DXP. We define a DXP as a comprehensive suite of solutions enabling enterprises to deliver a content-rich, stakeholder-driven digital experience (DX), encompassing all digital touchpoints.

Its main function is to digitally enable the three pillars or modules of DX – content management, brand engagement, and digital e-commerce – so enterprises can create business value through a well-structured and unified experience.

The Digital Experience Platform (DXP)

1

  • Content management: A DXP offers various services across the content management lifecycle, such as dynamic templates for designers, a library of frequently used content, and widgets and tools for reviewing and publishing content to multiple platforms, which help enterprises effectively and centrally manage the content they publish.
  • Brand engagement: A DXP unlocks numerous aspects of brand engagement across functions including marketing, advertising, sales, and experience management. With capabilities like end-to-end campaign automation and drag-and-drop tools to design customer journey maps, a DXP enables experience-as-a-service for enterprises.
  • Digital e-commerce: A DXP activates different facets of digital e-commerce with solutions like AI-enabled merchandising, visual merchandising, automated management and maintenance of product data, and central dashboards to manage all websites.

In addition, a DXP has tools to help deliver a data-driven experience across the customer experience value chain by enabling functions such as sales, marketing, merchandising, and content publishing via different modules.

Beyond the basics

Most of the DXPs in today’s market provide the same basic services. But the leading DXP providers also provide ancillary, value-add services on top. Some of the most popular are omnichannel services, API-integration, and tools for improved developer experience.

Per our recently released research report, BigTech Battle: Digital Experience Platforms (DXP) Assessment – Rise of the Digital Experience Platform, the leading players are adding more functionality to the DXP to enhance its features and functionality. For instance, they are helping make the development process less technical with the help of services such as What You See Is What You Get (WYSIWIG) interfaces, drag and drop functionality, and templates to create new experiences. This significantly reduces the creative team’s dependency on the technical team and improves the overall efficiency of the experience delivered. The top providers also have tools for end-to-end omnichannel customer journey mapping and enable the use of “win scores” to prioritize sales opportunities and probability metrics to measure the experience delivered.

2

These players are also using technology to enhance the functionality of the different solutions they offer, such as AI for content creation, event-based automation (cart abandonment), and advanced analytics solutions.

Simply put, a DXP is a more efficient way for an enterprise to manage its DX. In today’s increasingly competitive market, enterprises need to leverage a platform-based approach to deliver a compelling and sticky experience.

For more insight on the DXP market and a detailed analysis of current vendors, please read our report: BigTech Battle: Digital Experience Platforms (DXP) Assessment – Rise of the Digital Experience Platform.

Please share your experiences with the digital experience platform and the overall experience ecosystem with us at [email protected] and [email protected].

Software Is Eating the World, but Services Are Eating Software | Blog

“Software is eating the world,” wrote Marc Andreessen, co-founder and general partner of venture capital firm Andreessen Horowitz, in an essay published in The Wall Street Journal in 2011. But today, it’s clear that services are eating software. The implications of this trend are very significant for companies. The advantages are clear, but it’s also clear that there are challenges. Most companies today are not set up to deal with a services world. I believe they need a new set of management and operating models that allow companies to get clarity on what they are doing with services and allow them to stay in control.

Read my blog in Forbes

In AI We Trust, Thanks to AI Checking Software | Blog

The increasing popularity and uptake of Artificial Intelligence (AI) is giving rise to concerns about its risks, explainability, and fairness in the decisions that it makes. One big area of concern is bias in the algorithms that are used in AI for decision making. Another risk is the probabilistic approach to handling decisions and the potential for unpredictable outcomes based on AI self-learning. These concerns are justified, given the implicit ethical and business risks, for example, impact on people’s lives and livelihood, or bad business decisions based on AI recommendations that were founded on partial data.

The good news is that the software industry is starting to address these concerns. For example, last year, vendors including Google, IBM, and Microsoft announced tools (either released or in development) for detecting bias in AI, and recently, there were more announcements.

IBM

Last year IBM brought out:

  • Adversarial Robustness 360 Toolbox (ART), a Python library available on GitHub, to make machine learning models more robust against adversarial threats such as inputs that are manipulated to derive desired outputs
  • AI Fairness 360, an open-source toolkit with metrics that identify bias in datasets and machine learning models, and algorithms to mitigate them

Last month, IBM further augmented its offerings with the release of AI Explainability 360, an open source toolkit of algorithms to support the understanding and explainability of machine learning models. It is a companion to the other toolkits.

Cognitive Scale

Cognitive Scale recently unveiled the beta of Cortex Certifai, software that automatically detects and scores vulnerabilities in black box AI models without having access to the internals of the model. Certifai is a Kubernetes application and runs as a native cloud service on Amazon, Azure, Google, and Redhat clouds. Cognitive Scale also unveiled the AI Trust Index. Developed in collaboration with AI Global, it will provide composite risk scores for automated black-box decision making models. This is an interesting development that could grow to become a badge of honour for AI software, and a differentiator for those with the most trusted rating.

The Reality of Bias

While these announcements and those made last year are good news, there are aspects of AI training that will be difficult to address because bias is all around us in real life. For example, public data would show AI that there are many more male CEOs and board members than female ones, leading it to possibly conclude that male candidates are more suitable for shortlisting for a non-executive director vacancy than women. Or public data could lead AI to increase mortgage or auto loan risk factors for individuals living in a particular zip code or postcode to unreasonably high levels.

It is the encoding and application of these kinds of biases automatically at scale that is worrying. Regulations in some countries address some of the issues, but not all countries do. Besides, the potential for new threats and risks is high.

There is still a lot more for us to understand when it comes to making AI fair and explainable. This is a complex and growing field. As demand for AI grows, we will see more demand for solutions to check AI as well.

Request a briefing with our experts to discuss the 2022 key issues presented in our 12 days of insights.

Request a briefing with our experts to discuss our 2022 key issues

How can we engage?

Please let us know how we can help you on your journey.

Contact Us

  • Please review our Privacy Notice and check the box below to consent to the use of Personal Data that you provide.