Welcome to the Algorithmic Age and the need for Analytic Accuracy Assurance
1 like77 views
This document discusses the rise of algorithms and analytics in business. It notes that algorithms will increasingly manage interactions with consumers and optimize outcomes. However, historical data is not always accurate for predicting the future, especially during disruptive times. It argues that software and businesses need several critical components to identify when analytics lose accuracy and autonomously course correct, including: the ability to identify disruptions, use resilient information during disruptions, navigate disruptions autonomously or with human input, and have teams that maintain the algorithms and artificial intelligence over time. The document also discusses the increasing complexity of consumer interfaces and need for integrated platforms and processing to handle different interaction channels.
Welcome to the Algorithmic Age and the need for Analytic Accuracy Assurance
- 1. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 1 | P a g e Abstract We are entering an age where algorithms are the underlying forces that manage interactions with consumers and members of your value chain. These algorithms deliver dynamically optimized content that address the wants, needs and desires of consumers and convert the delivery of the correct content into commercial transactions or referral income opportunities. Software robots, or the autonomous software agents orchestrated and enabled with artificial intelligence, employ these algorithms to determine a path that optimizes organizational value. In most cases the employed analytics utilize historical data to determine the appropriate trajectories that optimize organizational value. There are times, however, when historical data is a poor predictor of future outcomes. These disruptive times will be commonplace during the foreseeable future. Many solutions that enlist the services of software robots available today do not have some of the critical components to identify and autonomously course correct for these disruptive times. There are some critical components are often lacking from robotic engines or common business practices and will be described in this writing. These facilities are A common framework that integrates interactions, the delivery of content, facilitation of referral income and commercial transactions into one integrated common platform‐based framework, Autonomous software capable of identifying when interactions, facilitation of referral income and commercial transactions arrive with unexpected outcomes, and can autonomously course correct, Software components devised to identify and use the information most resilient to unexpected market forces when prescribing actions to take which are devised to navigate disruption waves, Autonomous software that can robotically navigate disruption waves when possible and request swift actions from business stewards when appropriate actions to unexpected market cannot be computed, Sufficiently robust workbench capabilities that allow business stewards to review robotic actions and immerse themselves in redirecting activities when necessary and Enabling software and enabled teams tasked with the creation and maintenance of robotic software, algorithms, analytics and employed artificial intelligence at the breakneck speed of digital interactions.
- 2. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 2 | P a g e Introduction We are amidst a communication and computing transformation as disruptive as the printing press and the industrial revolution was to everything. Described as the Digital Revolution, the Algorithmic Age, the Graphene Age (because of the miracle product that will be utilized for construction, packaging, power and other uses not yet identified), the Singularity (because computational power orchestrated through autonomous software agents will displace mundane operations requiring a human presence) and other descriptions, this transformational age will use information as a catalyst to drive interactions, referral income opportunities and commercial transactions. This transformation will culminate with highly intelligent robotic software, some of which will utilize the capabilities of quantum computing. All of this robotic software will have one common attribute, that is the enlisting the assistance of algorithms, analytics and artificial intelligence to orchestrate all interactions and commercial transactions, regardless whether the playing field for these activities are the physical, digital or some other venue. The metamorphoses to this algorithmic age will take some time, partially because all of the facets necessary to operate fully autonomously are not yet available but also because there are cultural, regulatory and other constraints. Many believe this future state algorithmic age will be the status quo by the end of the century (see the Singularity University for more on this viewpoint). The road to this algorithmic age will be disruptive as components and methods necessary to realize this future state are enabled or discovered, some of which are already describable but many not. At the end of the day, consumers will collectively decide which of the innovations shaping the future state of this envisioned marketplace survive, making clear definition of this long‐ range future state very difficult if even possible. However, there are some components of this future state that are definable, many of which will be mentioned in this writing. There are six critical characteristics of the engine used for orchestrating the interactions with consumers and members of the value chain, the delivery of the most appropriate content that meets the wants, needs and desires of consumers in ways congruent with the value propositions of the organization, identification and realization of referral income opportunities and the management of commercial transactions. These are: An integrated processing environment that orchestrates across all interaction and marketing channels, whether occurring in the digital, physical (through internet of things devices) or an integrated playing field (i.e., augmented reality). The integrated processing environment will be platform‐based and will be highly customizable to meet the demands of all markets in a highly disruptive environment. Facilities that can identify when the processing activities devised for each interaction / market model has been disrupted. Disruptive forces will be a key component shaping the future state. Early warnings that allow identification and redirection of processing activities is a critical component used to redirect activities to meet the revised actions mandated by disrupted market forces. Deployment of a method to identify the most suitable and resilient information. During disruptive times, much of the information available for analysis will be tarnished because of its context buried in historical alignment, a condition not applicable during disruptive times. Autonomous robotic software capable of navigating disruptions and adjusting processing activities autonomously when time criticality is necessary and dispensing requests from business stewards when either time permits or when risks are sufficient to require business steward input.
- 3. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 3 | P a g e Deployment of a kill switch, which when triggered by a business steward allows them to interject their strategies and tactics to the robotically computed activities. Enlisting a team of adaptive intelligence scientists, who are tasked with the creation and maintenance of the robotic software, algorithms, analytics and artificial intelligence used to robotically orchestrate the digital activities of the organization. Many industries have an easier time of synthesizing the above 6 characteristics into their value stream. However, even the digital pure plays are finding ways to wield their influence into the physical world with devices such as Amazon Alexa, Oculus Go, Microsoft Cortana, Google Home, Apple Siri and others which are already beginning to blur the lines between the physical and digital marketplaces. Organizations will find themselves in a greater and greater need to enlist their processes with the facilities used to identify and react to unexpected market forces, autonomously course correct when possible and have the wherewithal to analyze the unexpected market conditions, bracket the conditions to enlist the services of the appropriate autonomous software and use the enlisted autonomous software to compute a trajectory through a disruption wave that benefits the organization. Integrating the Client and Value Chain Facing Components One of the attributes of the algorithmic age is the ever‐increasing means available to interface with consumers and members of an organization’s value chain. Three examples of this complexity influencing interface channels are Internet of Things devices, which allow the physical markets to integrate into the digital landscape, Intelligent voice activated services such as Apple Siri, Google Home, Amazon Alexa, Microsoft Cortana, Samsung Bixby and others allowing the control of intelligent devices through a verbal portal and mobile devices and Extended Reality (XR) devices which will augment the physical world with digital input. Further mashing of interfaces used to interact with participants of the markets can be expected, with little chance of describing what those extensions will be with any precision. It is important for organizations to manage all interaction venues through a single processing stream so that consumers can interact with the facilities brought forth into the marketplace across all venues at the whim of the consumer. The current underlying processing facilities used by many organizations will be unable to deliver the flexibility required to handle these complex interactions and convert a portion of them into either commercial transactions or referral income, thereby requiring a major overhaul to be responsive to this expanded approach to interaction interfaces. An integrated back end will allow for the integration of new and modified interface models to be employed at a speed demanded by consumers. This back end will integrate the information streams available from the interface model into a common platform so that consumer intelligence can be augmented no matter how the consumer or members of the value stream interact and so that potential referral income can be realized no matter how it arrives and which platforms are utilized to deliver the referral income opportunity.
- 4. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 4 | P a g e What the back end of that digital interaction model looks like needs to be defined, supporting a model of plug‐ins to a unified market interaction method utilizing robotic software serving interactions, transactions and adaptive intelligence. A plug and play approach to Platforms Organizations will be served best if they can easily manage the platforms that collectively deliver the organization’s interface model for the orchestration of all interactions, content delivery, transactions and adaptive intelligence used to continuously fine tune the interface model to an organization’s benefit. The future is built on platforms, platforms commonly not under the direct control of the organization. Many organizations will find this approach quite foreign as they enlist the services of more and more platforms which collectively represent nothing less than the organization’s digital identity. Many organizations do not possess a sufficiently nimble platform infrastructure that allows them to replace platforms that either do not fit the scale of the organization’s business models or themselves introduced changes to their mode of operation which are contrary to the benefit of the organization. An analytic layer which continuously analyzes the performance and participation in meeting the value propositions of the organization and prescribes changes to the platforms enlisted in the digital infrastructure of an organization. A plug and play approach to platforms will allow organizations to painlessly make the necessary changes.
- 5. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 5 | P a g e A word about Software Robots and unexpected business conditions Robotic processes use autonomous intelligent software robots which prescribe and synthesize interactions from both the physical and digital world and use a new class of software. The new class of software, or Robotic Processing Automation (RPA), is the orchestration layer that calls upon the services described in a pre‐defined collection of storyboards, a library of software robots capable of being called into action and a collection of platforms each serving one component of the digital representation of the organization. Collectively, these components are enlisted to orchestrate business processes without the benefit of human hands. These processes have a level of artificial intelligence to make decisions that are coded into the orchestration layer, and the success of the RPA solution is dependent on eight things: An understanding of the market which is orchestrated. The view of the market is the business model and the processes incorporated into the business model. An understanding of the platforms enlisted to interface to the participated markets and convert interactions carried out in the participated markets into commercial transactions or referral income. An understanding of how to identify disruptions and changes in phases of a disruption (aka, the disruption wave). An understanding of the consumers participating in the market. The level of influence on the needs and desires of consumers will be greater if the granularity of the information available for depicting the activity of consumers, which is used as a surrogate for the needs and desires of consumers. An understanding of how to craft the most appropriate content used to meet and influence the wants, needs and desires of consumers.
- 6. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 6 | P a g e A model used to predict the outcomes of the influence on the market achievable by each content component available to be launched into the market. A series of metrics used to measure the outcomes of content and adjust the model predicting outcomes where appropriate. A series of workbenches used to monitor the metrics and information streams available and provide the resources for adaptive intelligence scientists and business stewards to positively influence the outcomes enjoyed by the organization. These robots are codified in such a way that they enlist the services of algorithms and artificial intelligence to determine the most appropriate actions to be called next and react to small deviations from expected activities in a way that maximizes organizational value in ways described in the business models of the organization. While these RPA platforms are available today, for the most part they are capable of orchestrating business processes that represent expected outcomes, and in most available solutions, demand that business stewards take the helm when deviations from anticipated actions are experienced. The speed at which business stewards are enlisted and the time available for their actions to be thrust into the marketplace is woefully inadequate to deal with the breakneck speed of the forces that control the markets and the speed at which good and bad news circumnavigates the globe in this accelerated marketplace. Most of the implementations to date concentrate on the technical aspects of the solution and for the most part have not yet linked back to the highly fluid business models and the identification of non‐business as usual and disruptive conditions nor have the implementation of early warning systems which interrogate deviations of the assumptions used to fit business models as representative depictions of marketplaces. Expect much more activity in improving the capabilities of these platforms, plus the introduction of determining what is a potential disruption in the mix, which requires additional activities described throughout this writing. If these robotic solutions are not in your current plans in a way that impact the information and content arsenal you have at your disposal and their linkages to your strategic initiatives, then you are lagging your market leading competitors.
- 7. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 7 | P a g e There are some cultural changes that need to be accepted for consumers to embrace this new approach to everything. Consumers have shown little resistance to giving up control to the various applications imbedded in their cell phones but have shown disdain for the onslaught of robocalls hawking everything from new credit cards, vent cleaning, changing phone plans, out and out scams and a host of other activities. This new world has brought on the reduction in value of do not call lists, the need to protect credit cards from skimming and a renewed fear of identity theft. Regulatory actions taken to put privacy concerns of consumers into the forefront (i.e., GDPR) is a step in the right direction to mitigate these automation fears. Ensuring the Relevance of Software Robots, Algorithms and Artificial Intelligence The breakneck speed and highly disruptive nature of this emerging algorithm driven marketplace demands a constant creation and maintenance facility that keeps the legion of software robots tuned to the forces of the markets. Enlisting a team and arming them with the facilities to maintain the components used to compete in this automated marketplace is a critical component of the arsenal required for successfully competing. While some will argue that many organizations possess data scientists, there is a difference between what data scientists do and what adaptive intelligence scientists do and how they accomplish their tasks. Software robots and the facilities called upon by software robots are not enabled through scientist‐controlled analytics, regardless whether they are recurring or custom analytics. The tools available to data scientists are not sufficiently nimble to develop, maintain and bench test the highly critical components utilized to compete in this algorithmic age.
- 8. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 8 | P a g e Identifying when Analytics will yield Less than Accurate Results and fixing them During disruptions, the trends available from historical data will not be representative nor useful in forecasting future outcomes nor useful in prescribing impending interactions. Conditions that represent a departure from expected outcomes must be identified and called upon as an early warning so that actions can be taken to adjust the information called upon to yield accurate outcomes from analytics. These early warning beacons and the facilities they call upon, rarely available today, are tasked with identifying and bracketing unexpected outcomes so that they can instruct software robots to not utilize the pre‐defined collection of activities as the roadmap cast to represent activities that yield expected outcomes. These facilities must determine the characteristics and magnitude of the departure from expected outcomes. Specialized software robots should be employed to identify these conditions and send alerts and bracketed conditions to the robotic software orchestration engine to accommodate the market conditions accordingly.
- 9. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 9 | P a g e Ensuring the availability of the Information Catalyst and Data Privacy There has been much stated about data privacy recently. Much of this conversation has focused on the regulatory characteristics of data privacy. Consumer intelligence is necessary to serve the wants, needs and desires of consumers. Without the readily availability of the catalyst for commercial transactions and referral income, namely consumer information feeding consumer intelligence, the ability to convert interactions, commonly delivered optimized dynamic content, into commercial transactions and referral income will be severely marginalized. Organizations must grasp the principle that assuring consumers that they have their backs when it comes to data privacy is required to ensure the necessary free flowing information feeding consumer intelligence and a cost of doing business in the algorithmic age. Readying for the Algorithmic Age Four organizations, commonly called FANG (Facebook, Amazon, Netflix and Google) are used as the today’s benchmark for organizations that successfully thrive by utilizing the information they collect from consumers to deliver services that serve the wants, needs and desires of consumers while understanding how to monetize this collected information. A checklist is provided as a vehicle to benchmark whether your organization can minimally exploit the deluge of information available to your organization to serve the wants, needs and desires of consumers. Many organizations will not possess the characteristics portrayed in the checklist today, but many will possess the underpinnings necessary to prioritize the initiatives that will improve the ability to thrive in the algorithmic age. The checklist is provided as a vehicle to communicate the continued improvement in capabilities required to ready the organization for the algorithmic age.
- 11. Welcome to the Algorithmic Age and Analytic Accuracy Assurance When & Why Analytics Lose Accuracy and What to Do About It 11 | P a g e About the Author Mark Albala mark@infosightpartners.com Mark Albala has been one of the early practitioners in wielding information as a competitive weapon and has served as an advisor to vendors and analysts serving the information insight industries. Through much of Mark’s career, he has served as an innovator and trusted advisor to the global 2000. He has led the geographically disbursed teams who helped address the information strategic, governance and architectural issues plaguing the global 2000. Of specific interest has been and continues to be assisting organizations optimize the value they can extract from information. This includes aligning information with business processes, protecting personal information entrusted by consumers, required to understand, predict and satisfy their wants, needs and desires and eradicating the reasons for not using information at times when its use would be most beneficial to the organization. Mark has held several senior management positions most recently leading globally distributed organizations dealing with the most difficult information centric issues organizations tackle. Mark currently serves as President of InfoSight Partners, L.L.C., a firm which delivers both technical and financial advisory services to the global 2000. Mark’s role at InfoSight Partners is focused on helping companies come to grips with their information challenges and help them devise the means to manage their information as assets or the organization. Mark has been a champion of helping organizations monetize their information assets, which demands they devise mechanisms and practices to measure its value. Most recently, Mark has been focused on writing about practices devised to help organizations improve the management of their information assets. One key component of managing information assets is determining a value for information assets. The first of his major publications, Thriving with Information in the Digital Economy, is available on Amazon, and the second book, Thriving with Digitally Adaptive Intelligence, is expected for release during 2018. The third book, currently being planned, will cover what it takes to prepare for the autonomic age (the end state of the digital economy), specifically focusing on actions to allow software robots to manage the full spectrum of interactions, whether carried out in the digital or physical environment.