Succeeding at Artificial Intelligence Requires a Commitment to Agile Experimentation

I went to graduate school and got my Masters in Electrical Engineering, but that's not what I studied. The University of Arizona had a strong program in optical sciences, medical imaging, information theory, and something called "machine learning" and I opted to take classes and ultimately complete a thesis on these topics. I remember learning the math and computing of neural networks, the computer vision algorithms behind facial recognition, and the underlying mathematics of mpeg encoding.

And I remember spending countless hours in a lab testing algorithms on a Unix workstation. Would a reinforcement learning algorithm work better than a two layer neural network? Should a genetic algorithm work better, or am I programming it incorrectly? Should I apply a fuzzy controller, perform an operation in the Fourier space, focus on heuristics or prove out the underlying mathematics?



Most of what I remember is waiting for that workstation to spit out a result. There wasn't a supercomputer that I had access or a cloud environment where I could ramp up and run several experiments in parallel. In the end, artificial intelligence back then was a lot of experimentation between what data sets to test, what algorithms to apply, what parameters to configure, and how to best program them to get better performance.

Are you Ready to Experiment with AI and Machine Learning? 


I've been fortunate to have had some opportunities to develop artificial intelligence in business applications. I've developed or led teams to develop tools that enable comparing genetic and protein samples, natural language processing algorithms to extract search terms from newspaper classified advertisements, and document processing techniques for extracting building material names from construction blue prints and building specifications. What was common across all three applications is that it required significant experimentation, first to get the basic algorithm in place and then later to build up more intelligence to handle more disparate use cases with increasing quality and performance.

Is AI Today Fundamentally Easier?


The simple answer is yes, but the longer answer may be no. Today, a developer can access AI through APIs provided by IBM Watson, Google Prediction, Microsoft Cognitive Services, Amazon Machine Learning and many others MLaaS (Machine Learning as a Service) or AIaaS (Artificial Intelligence as a Service). The largest startups in AI have been funded north of $10M and the biggest companies are making multiple investments. There is a published Intelligent App Stack illustrating use cases, machine intelligence providers, and data prepping tools and other AI Marketplace overviews.



But the longer answer is maybe not. As a business person with an opportunity to apply AI or machine learning or a development team that has the priority to implement, you have many options to implement. You need developers that can normalize data sets and plug into APIs or third party services. You need data scientists that have at least some familiarity with the basic algorithms from clustering to neural networks to deep learning. You need subject matter experts that can validate the output and suggest improvements. Most importantly, you need an agile experimentation process to try approaches, configure, run, and validate results.

Lots of choices, lots of talent, lots of time to implement. But the rewards can be significant.

continue reading "Succeeding at Artificial Intelligence Requires a Commitment to Agile Experimentation"

Defining Digital Transformation, Strategy and other Digital Terminology


    In my last post, I took a stab at defining the Digital Mindset from the perspective of different roles and responsibilities in the organization. In doing so, I used several terms, some industry standard and others that are emerging so with this follow up post I thought I'd provide more color on my definition.




    Digital Terminology


    • Digital Transformation - See my older post older post, I provided my definitions for Digital Business and Digital Transformation.

    • Digital Strategy - If digital business defines an end state and transformation defines a path to get from current state to end, then Digital Strategy sets the scope, priorities and constraints of the transformation. It may state what customers segments you'll target, what digital capabilities are priorities, and what regulatory and other business factors are potential constraints. Strategy often communicates scope, priorities, and constraints in the forms of mission, goals, and values to help members of organization translate a digital vision to "what needs to be done".

    • Digital Opportunities - These are opportunities to grow revenue, new customers, and new markets through digitally driven products and services. It should be specific, so for example, "Investing in this [capability | new product | new service | product/service enhancements | new channels] we can grow revenue by [increasing prices | selling to a target customer list | growing market share in these areas]".

    • Digital Capabilities - As a technologist, I also use the word "Digital Capability" to identify technical attributes of products, services, supply chains, or business processes that strategically developed around a technical capability without assumptions on the underlying digital platform. This can include everything from reaching new customers through digital channels (a mobile digital capability, a social digital capability), the capability to process big data (big data capabilities), the ability to integrate with things (IoT), the ability to source work in a business process algorithmically (crowdsourcing, data processing services, AI as a service), or the ability to automate and integrate transactions (digital supply chain, blockchain). Digital capabilities are also what we deliver to the organization to make them more smart, efficient and collaborative and are properties of underlying systems that enable elasticity and automation.

    • Digital Products - These are products designed a digital customer experience first and add "brick and mortar", "paper" and other "analog" experiences later if needed. These products take advantage of artificial intelligence, location awareness, voice controls, interoperability with sensors, and access to a digital ecosystem of services to optimize the user experience and provide personalized conveniences.

    • Digital Disruptors - Are competitors that offer Digital products that can challenge or disrupt an existing business or product line. See my example of digital disruption in the newspaper industry.

    • Digital Interoperability - Implies that the data collected in one system, application or interface can easily be leveraged in other platforms. Some interoperability is achieved through a unified customer experience, for example, when an application provides access to all your data and provides similar experiences when accessing over web, mobile, and other interfaces. Sometimes interoperability is achieved when vendors agree on data and interface standards. Lastly, interoperability can occur through integration platforms such as ifttt, Zapier, and other iPaaS providers.

    • Digital Ecosystems - Digital ecosystems form when there are a wide range of services that can be easily interfaced that provide digital capabilities. Participants in the ecosystem often develop application programming interfaces (APIs) and other data interfaces to enable sanctioned consuming applications to access these services. The availability of these services implies that businesses can develop best in class applications by leveraging third party services in the ecosystem. They can also extend their customer reach and develop new businesses by making their own digital capabilities accessible. 

    • Digital Platforms - Are technologies selected by enterprises and organizations to be foundations for their digital businesses and to support digital transformation. I have a previous post on criteria to select digital platforms and proposed a process to evaluate transform enabling technologies.

    • Digital Processes - A business process that is heavily automated, robust to respond to evolving business scenarios, and is highly data driven. They can be contrasted with manual, paper driven, heavily stage gated processes that are inefficient, error prone, or have bounded scalability. McKinsey defines digital process innovation as, "A focus on the implementation of new or enhanced technology-enabled ways of working—or digital process innovation—can help companies simplify the technology landscape, reduce overall IT costs, and bring products and services to market quicker, thereby realizing greater earnings potential"

    • Digital Channels - Omnichannel is when customers have multiple ways they can interact with a product, service, business, sales person, or customer support. It also is when the sales, marketing, and support organizations have multiple ways they can interact with customers, prospects, leads, and users. A digital channel is when the collaboration, customer experience or communication occurs using digital tools, contrasted with in-person or physical channels.

    • Digital Marketing- Marketing activities to customers, prospects, and leads through digital channels. See building capabilities in digital marketing or benchmarking digital capabilities as good references. 

    • Digital Practices - Are attributed of the organization defined through roles, responsibilities and talent that implement digital strategy and transformation. MIT has a good write-up on these digital capabilities your organization needs and McKinsey's write-up discusses strategies for acquiring digital capabilities. Also see my post, What practices are needed for Digital Transformation.

    • Digital Talent - Digital talent refers to having individuals with digital skills. Common digital skills are in leveraging mobile tools, data analytics and visualization, social networking, collaboration tools, abilities to configure web tools, and even some basic coding. Beyond these basic skills, digital skills can be very role and organization dependent. If you are in sales, digital skills often includes understanding of CRM and sales automation. If you're in finance, you should have skills working in one or more ERP platforms. Here's a good study on developing digital talent.

    • Digital Workplace - From Gartner, the Digital Workplace enables new, more effective ways of working; raises employee engagement and agility; and exploits consumer-oriented styles and technologies.When I think of the digital workplace, I think of unified communication tools, collaboration environments, and enterprise mobile capabilities all aimed to help teams share information and be more productive.
    Did I miss any?


    continue reading "Defining Digital Transformation, Strategy and other Digital Terminology"

    What a Digital Mindset Means Role by Role

    We all hear and say that "Digital" requires a different mindset, a "Digital Mindset, but what exactly does that mean? If you think digital mindset is about having the latest gadget, or being the go-to person to configure your mobile device, or knowing the latest startup that is likely to disrupt an industry then you're missing the point. If you believe that digital natives have a leg up on having a digital mindset, then that might be the case when it comes to how they conduct their personal lives but not necessarily sufficient to be an active participant or role model when an organization requires a digital mindset.

    Having a digital mindset implies a new set of values and principles for the organization but applied individually based on role and priority. It starts with how customers perceive the organization's brand and their expectations to how fast and smart their product and services are relative to other options. It's services by an organization that also needs to be smarter, faster, safer, relevant internally to be able to deliver competitive offerings. This often requires individuals to understand and leverage digital technologies in new ways, and to shed old habits that are slower, inefficient, insecure, or digitally isolated from standardized practices.

    Do you have a Digital Mindset?


    Here is a summary of what having a Digital Mindset means role by role


    • As a customer, I want relevant, contextual insights that are important and useful. I want the algorithm to know my preferences and take action on my behalf where I permission.  I want freedom of digital selection anbd expect product and service interoperability. I require organizations to secure my information and want transparency on how it is used and shared.
    • As a developer, I want to be able to experiment with new digital platforms, develop prototypes, and test pilots without having to demonstrate business value and ROI up front. I want to work in an agile, low stress organization that enables me to develop solutions that empower customers.
    • As an IT engineer, I want to make sure that key operations to administrate applications are automated and that applications are secure in environments that can quickly scale up and down based on business demand. 
    • As a business analyst, I want to leverage six-sigma processes to document existing business processes so that they can be digitally reinvented using agile practices.
    • As a sales person, I want to make sure that products are competitive and continue to receive investment aligned with customer needs. I want to make sure that the CRM has the most accurate relevant information on my interactions with prospects and customers.
    • As a data scientist, I hope we safely collect data from our products, internal processes, and enterprise systems, ask questions, and leverage data in decision making.
    • As a product manager, I want to develop digital-first products that delight customers, enable their success, and grow revenue. I want to continually measure customer feedback and invest in new capabilities and partnerships that align with market needs.
    • As a marketer, I want defined KPIs around brand, leads, and customer plus the ability to experiment with digital marketing tools to optimize how we reach and what we message to customers and prospects.
    • As a customer service representative, I require appropriate access to all the information on the customer I am speaking to so that I can provide the best advice to solve their issue or present the most likely opportunities that service their needs.
    • As a financial analyst, I want to make sure that my forecasts are available to appropriate business users in the form of dashboards and data visualizations so that they can review KPIs, ask questions, and drill into the underlying data.  
    • As a change agent, I want to align the organization to future values, enable people to shed yesterday's ideals and practices, and promote behaviors and activities aligned with the organization's digital strategy. 
    • As a human resources leader, I want to make sure that we leverage digital tools to enable the workforce and ensure that we recognize and reward employees that are actively enabling the organization's digital future.
    • As a manager, I want the organization to make quick decisions. I want support of my colleagues when we have to do things fast and find ways to collaborate on solutions.
    • As a leader, I want to participate in creating a digital vision then make sure that we can review and adjust priorities based on customer need, market conditions, and our ability to execute.
    • As a shareholder, I want to make sure the organization's leadership understands and has plans targeting digital opportunities and digital disruptors while ensuring they meet digital security, privacy, regulatory and other compliance requirements. 

    But it's not enough for individuals to see their roles from a digital mindset. It takes a new collaborative, agile culture for individuals to work in teams aligned with vision an priorities.

    Good reading on the Digital Mindset



    continue reading "What a Digital Mindset Means Role by Role"

    Advice from Technology Dads on Celebrating Father's Day


    Happy Father's Day!

    I find fellow tech dad's to have an interesting perspective on raising their children. So with that in mind, here's my advice for Techie Dads on this Father's Day.





    1. Drop the Technology - I'll get this post out, then turn the tech off for the day to spend time with the family and kids. 

    2. Teach your Kids to Code - Not today, but make a commitment to find a tool, an app, or a course to get your kids at least familiar if not down right interested in coding. Have a look at Code.org, Kodable, Tynker or posts from ForbesMommyPoppins, or PCMag for inspiration.

    3. Start a blog - This is a recommendation I make to all young fathers. Make it private and get a weekly post up on what you've done with your kids and their latest milestones. It will pay dividends 5-10 years out.

    4. Consider becoming an Agile Family - I posted about becoming an agile family with a scrum board. I confess that we're not doing this in my family, but friends of mine who have tried swear by it.

    5. Learn something together - Instead of playing the latest games or watching the newest movie. Try watching a TED talk for kids together. If you're driving today, try one of these NPR Podcasts for Kids.

    6. Reconsider your privacy settings - On Facebook, Instagram and other social sites.

    7. Setup your Phone's Emergency Settings - Here's a good how to post for iPhone and Android.

    Have a great day!

    continue reading "Advice from Technology Dads on Celebrating Father's Day"

    Will Microsoft and LinkedIn Deliver on the Promise of Social CRM

    If you were an early adopter on LinkedIn, you probably got an invite from a colleague that wanted you to "join their network" without any significant justification as to what services that network provided. Early adopters like myself joined and allowed LinkedIn to access our email contacts so that they could help us grow our network.

    I followed these steps about 10 years ago and didn't find value until about a year later when my laptop crashed and I lost my Microsoft Outlook contacts. How old was my last backup? Embarrassingly, too old but lucky for me, I had just integrated all my contacts on LinkedIn. From that day forward, Microsoft Outlook, Exchange, and other tools became irrelevant to manage contact as they were now always available on LinkedIn's cloud. So long as I trusted LinkedIn with my data, I didn't have to worry about backups, moving contact lists when I changed jobs, or updating information when my contacts made career moves. 

    The Impact of Social Platforms on CRM


    At the time, I always thought that LinkedIn would evolve to disrupt Outlook and Exchange at least for my outwardly facing messaging needs. Would Exchange be disrupted? Would Social CRM mature faster with LinkedIn and other social platforms?

    The answer to date has largely been somewhat no. Many organizations enrich their contact information with data from LinkedIn and other providers. Some filter social interactions aggregated from Facebook, Twitter and other platforms into social queues in their CRM. Many stage social interactions through CRM managed workflows to capture new leads and automate marketing activities. There's a lot of social capabilities, but social platforms today are largely inputs and outputs to CRM workflows and capabilities.

    Has there been a fundamental disruption to CRM or customer engagement through these platforms? It depends on the digital maturity of the industry that you operate in, but for many businesses these tools and capabilities have been largely additive, not disruptive.

    Can Microsoft Disrupt CRM With LinkedIn's Data and Collaboration Capabilities?

    It's interesting to guess where the Microsoft and LinkedIn marriage will go - 

    • Imagine if enterprises no longer created contacts and leads from scratch or from lists. One started developing a target customer segment by querying the LinkedIn databases.
    • What if my knowledge of a Prospect was first based on the interactions that LinkedIn gathered through its capabilities (what news I read, what groups I participate in) and then supplemented by private activities performed in CRM?
    • Will CRM adoption improve if Dynamics was driven off the UX and design paradigms developed for a consumer site like LinkedIn rather than the data driven experience it and other CRM platforms have today?  
    • What if the LinkedIn profile I see is a mashup of LinkedIn's public data with my CRM data?
    • What happens when Cortana is integrated with LinkedIn and I can ask questions likes, "Are there any of my key prospects attending today's event?"
    • What will Salesforce, Oracle, and other CRM solutions do to counter Microsoft + LinkedIn?

    Interesting times...



    continue reading "Will Microsoft and LinkedIn Deliver on the Promise of Social CRM"

    5 Key Lessons for New Agile Product Owners

    At some point, organizations that want to scale to more agile development programs or take on wider scoped digital transformation efforts need to consider how best to cultivate Product Owners.

    Scaling additional development teams is a lot easier. Once an organization has a defined agile practice, a defined set of technology platforms, and a documented development standard then scaling development requires finding additional technologists, reshuffling teams so that new ones have a mix of both new and tested team members, and identifying aspects of the practice that need to scale.

    But adding product owners isn't as easy. Organizations have a choice of looking outside for people that have the skill set and experience but need to learn the business, customers, and product landscape, or, identifying existing employees that know the business but need to learn the art and skill of product ownership. The latter option is far harder to do than leaders understand especially those that equate product ownership with project management.

    What New Agile Product Owners Need To Learn


    Product ownership is a blend of several disciplines and skills, but what I focus on below are the key changes when product owners are groomed from within the organization.

    • Listening to customer and stakeholder feedback but owning decisions - Early product owners fall into a couple of traps.

      The first trap is when they do a good job of listening to customers and stakeholders, but then try to give everyone what they want. Product Owners in this camp fail to recognize that they are working with finite budget, time, and skill and their role is to digest the input then communicate a set of priorities, strategies and requirements. These product owners need to communicate the rationale behind these decisions and accept the fact that they aren't going to make every one of their stakeholders equally happy. Their job isn't to make people happy but to make decisions that drive customer adoption and growth.

      The second trap is when the product owners fall in love with their vision to the extent that they fail to listen to stakeholders including the technologists that they need to partner with on solutions. Product owners need to constantly reshape their vision and priorities based on feedback and those that don't listen may fail to adopt. Second, product owners need to share and communicate their vision repetitively and gain buy in over time, otherwise, stakeholders will communicate their objectives, often loudly and forcibly and probably late in the development process when changes or pivots are more costly. Lastly, product owners need to shape their vision based at least partially based on the organization's capabilities so they need to be presenting problems to their team and learning from their team feasible solutions.

    • Participating in Team Commitment is Key to Agile Success - New agile product owners often fail to fully understand the importance of getting a team's commitment and learning to manage to their velocity. Even when teams estimate and forecast a schedule, getting the team to formally commit insures that they have a shared understanding of the prioritized stories and have a plan to address them in the allotted time. Without commitment, new product owners may find their teams falling behind schedule, completing stories with quality issues, introducing unnecessary technical debt, or unhappy - all issues if the organization is going to maintain a culture and practice of excellence. 

    • Adopting the Team's Agile Tools and Practices - Because many organization equated product ownership with project management, they tend to promote individuals with management skills into this role. Many will make the mistake of trying to manage the team through tools that have worked for them in the past and try to force agile project schedules into other tools like MS Project, spreadsheets, PowerPoints, etc. It's a huge waste of effort duplicating this information, it leads to inaccuracies, and it will likely frustrate the team when their product owner isn't collaborating with them with tools and practices selected. Product owners need to start their role by learning and using the selected practices and tools. If anything, they should drive the team to adopt my recommended best practices in configuring agile tools.

    • Driving Data Driven Decisions - The best way agile product owners can get both stakeholders and agile team members on board with their priorities is to demonstrate that there is data backing their decisions. "What does the data tell you?" is the first question I ask product owners when reviewing their vision or priorities. If the answer is that we have limited data, then my follow up is, "Where in your backlog have you prioritized data collection priorities?" 

    • Leveraging Estimations To Shape Vision, Priorities, and Solutions - Understanding that estimation is not a contract. I have written several posts on agile estimation as a key tool to form a dialogue around solutions, to drive a discussion on MVP, and to help teams plan releases. Agile product owners have to use estimation as feedback to make decision, but some fall into the trap of using estimates to manage their teams to fixed timeline, fixed scope releases. Sometimes it works, but when estimates are done early in the process or performed by less mature teams they will have degrees of inaccuracies. Agile product owners that "box in" their teams to fixed schedules and deliverables will lose their ability to adjust priorities or worse, they may degrade the culture that drives both execution and innovation.

    Lastly, what steps are they taking on to learn their craft? I encourage agile product owners to adopt some self reflection. Are they adopting any of the 20 bad behaviors of agile product owners? Are they listening to their team's and prioritizing some work to address technical debt?

    Better yet, are they learning some of the practices of strong agile product owners?

    continue reading "5 Key Lessons for New Agile Product Owners"

    Agile, Citizen Development, Innovation, IoT are Keys to Digital Transformation - MIT CIO

    A couple of weeks ago I attended MIT's CIO symposium that focused on Digital. Let me share with you some of my key learnings

    • Agile is foundational to transformational efforts. Where CIO have had success is bringing customer experiences to the developer to help shape priorities, designs, and solutions. Agile was mentioned by most of the CIO on various panels throughout the day.My view - You need to work on getting the C-Suite truly on board with agile transformation.

    • The workplace of the future was by far my favorite panel of the day bringing several different perspectives on empowering and extending the workforce with digital capabilities. "The workplace of the future is driven by self-selection - someone having the skills stepping to solution," was Allison Mnookin's, CEO of citizen development platform Quickbase, explained. Mike Morris of crowd sourcing platform TopCoder view is that "A workplace is a perk" and that "Organizations need to find the best person in the world that can help solution." So either bring new digital tools to your workforce and let them step up to solution, or bring the best skills from the outside to solution for you. This panel asked and answered a difficult question, "Are you paying for effort or rewarding for results?" More on Citizen Development

    • Corner off innovation budgets is a key step CIO are taking and acknowledging as a best practice to ensure that innovation driven experimentation can occur outside of slower financial and governance practices.They agree that portfolio and ideation tools are useful to promote the best ideas, assign to the most capable teams, and provide them sponsorship to enable experimentation and learning from accepted failures. My view on calculating ROI too early in Digital Transformation.

    • Is IoT a platform? Not yet, said a panel of experts including George Collins (CTO, Deloitte Digital), and MIT Professor Sanjay Sarma. "IoT is a series of capabilities, for it to become a platform it needs focused standards that develop an ecosystem. Tipping point for IoT becoming a platform given is when there is semantics and syntax for coordinating between things." Basically, that means that the protocols that have been developed are proprietary and primitive - my analogy the TCP/IP of IoT - and that the industry will need easier to use, interoperable, and semantic protocols - for it to scale to wider use, My view - AI will be key to scaling IoT.

    Other posts about MIT CIO 2016



    continue reading "Agile, Citizen Development, Innovation, IoT are Keys to Digital Transformation - MIT CIO"
    Share