Selected Work

A large supply chain logistics firm that handles returns for large electronics brands in Europe needed to improve their efficiency and tracking to meet new contractual requirements. I led the development of a strategy using machine vision to track the receipt of boxed items, automated sorting of content, automated storage and retrieval of equipment, and tracking of returns through re-boxing and re-shipment. The implementation used a variety of automated guided vehicles to move inventory throughout the facility, an automated warehousing system for storage, safety cameras to alert fork truck drivers to human presence, and with the arrival of COVID, infrared detection at the entrance to identify staff that did not meet policy guidelines for on-site work. The resulting solution reduced the average processing time for a package from 2hrs to 22min, increased facility capacity by 30%, improved handling accuracy from 92% to 99%, reduced re-shipment costs by 6%, and reduced fork truck incidents to zero.

Supporting material

The Age of Customer Centric Manufacturing Presentation · August 2019

Wall Street loves well-run companies, but delivering outsized returns is tough. For a business services provider, the traditional route leveraged acquisitions in a highly fragmented market, but as the pickings became slim, line extensions emerged as an equally strong path to revenue. The challenge lay in servicing their tens of thousands of customers weekly. At a local facility the team loads out each truck based on statistical consumption norms. Friction enters the process when the driver arrives and the forecasted need doesn't match the actual need. Each trip back to the truck takes time and risks impacting the next customer. Out of stock items wreak even more havoc, requiring either a follow-up visit or a special dispatch.

Having reached out to the President and COO with the bold claim my idea could improve route efficiency by 6%, I worked with the VP Product Strategy to develop an AI solution using low-cost cameras to provide actual inventory counts. The solution improved loadout accuracy and route efficiency by 12%. Further, paired with additional analytics, the cameras opened new paths to revenue in theft detection, consumption analysis, and process optimization.

Auto insurance is marketed solely with humor. The industry believes the products are so well understood, universal, and price similarly, that brand value is better driven by mayhem, Flo, an Emu, a gecko, and Professor Burke than substance. As a result, churn is common and expensive for the insurers, running 16%-29% annually. I had an idea. If an auto-insurer developed a widely applicable product that protected customers regardless of the vehicle in which they travelled (ride share, taxi, school bus, etc.) instead of insuring their vehicles, they would create a new revenue stream while gaining visibility among those not already customers, while obtaining valuable data for improving their own claim adjudication process.

I created the attached video and sent a personalized version to an executive at a handful of auto insurance and service providers my research identified as forward thinking. A product built on the same concept later launched as a $15-a-month subscription. It uses AI on a phone to identify when a vehicle accident has occurred, connects the customer immediately with a crisis center able to provide safety guidance, and contacts 911 with precise location and identification information.

Supporting material

Redefining Insurance Concept film · Vimeo

Networks are viewed as boring infrastructure to most C-suite executives. To take advantage of my creativity and executive presence, I was assigned to a pilot Business Outcomes program tasked with leveraging the desired business outcomes of customers to engage them in large, innovation driven transformations. To limit risk, only customers who spent less than $50k per year could be targeted, but each program had to be worth at least $25M in revenue to Verizon. Clients needed to sign-off on an eight-stage business case that represented their agreement to work with Verizon to develop and implement the solution. Once completed, the business case was turned over to the sales team for pursuit. All of the business cases were tied to revenue generation for the customer, whether new digital services or enhancements to existing services.

To engage with sales teams, I spoke to their desired outcome of building executive relationships. I assembled a matrixed team and used a design thinking approach to build 2-3 truly innovative ideas that would solve real business challenges, each one rooted in a Verizon capability. I then solicited meetings by reaching out to C-suite executives and suggesting I had an idea that would help them achieve a specific desired outcome by anywhere from 2% to 5%, and if they were interested, I’d appreciate being connected to the right person. Using this approach, I was able to create relationships at the C-suite and senior executive levels with five companies who now represent over $50M/yr in spend with Verizon.

Supporting material

Redefining Banking Concept film · Vimeo

When the CTO of a Tier 1 bank puts 30 minutes on his calendar and invites you to a meeting without any explanation, you go. Upon arrival I was presented with the following problem. The Federal Reserve had tasked his bank with figuring out how to ensure 100% uptime, the continuation of banking operations even if both the primary and secondary data centers were unavailable. He explained that the bank was under close regulatory scrutiny. What I didn’t know at the time was that he was testing me. Would I, as the newly established Verizon Customer Success person on the team, throw products and services at him or admit it was an unsolvable problem. What he didn’t expect was the solution I presented.

Banking data for each customer I theorized couldn’t be very large, maybe a single megabyte at most. He agreed. The Fed only cared about people with savings and checking accounts, correct? Yes. And they wanted to make sure that during a disaster, people would have access to at least some of their funds, if not all? Correct again. What if a copy of all the most recent transactions was stored on the customer’s phone, in the secure sandbox where the app resides. It would be encrypted and only available to the app. In the case of a data center outage, the phone would act as a proxy for the bank’s data center, able to initiate transactions at branches and ATMs and store the results. When the data centers were restored, at the next interaction the updated data would be uploaded and processed and life could continue as normal.

He sat across from me for two minutes with a blank stare and deafening silence. He turned to the sales director who’d facilitated the meeting and told him “I like this guy.” From that point forward, we had carte blanche access to the CIO and his team, who helped me get in front of the retail banking SVP to discuss improving the customer experience.

Supporting material

Where Are We Going Now? Executive briefing · Verizon, May 2019

While advising a top 3 bank on cloud computing, I developed an AI strategy predicated on the belief that the majority of “unbanked” people, a primary topic in the news at the time, were poor at managing their own money due to a lack of knowledge and impulse control. I developed a concept titled Data as Currency that espoused using AI to help customers make better financial decisions in their everyday life by injecting insights and guidance into purchase transactions via their mobile device. The goal was to instill discipline by demonstrating the opportunity cost of unnecessary purchases, while exposing them to upstream banking services. I shared this idea with the CTO who loved the concept and sponsored a meeting with the Chief Data Officer. From that meeting we co-developed the bank’s AI strategy which started with his organization beginning a massive overhaul of their data quality.

Supporting material

Data as Currency Conference handout · Business Intelligence Innovation Summit, October 2013

AT&T had been largely excluded from deals at a major retailer, only used as a second provider where the primary wasn’t available. Upon assignment to the team in a Customer Success role, I immediately sought to understand before being understood. I learned their network strategy was a mess, control having been ceded to vendors with no traceability to the company’s strategy. Leading with generosity, I asked for a copy and returned with a blueprint for a five year, vendor agnostic strategy that aligned with the public statements of the CEO on vision and strategy, combined with industry insights I gleaned from the Internet. The Director of Networking was astonished, appreciated my effort, and after a couple of iterations my strategy and architecture roadmap was submitted. I soon learned it was the only one submitted to the VP of Enterprise Architecture who was leading an effort to professionalize IT. He asked to meet and walk him through the modifications I made to his template, which he incorporated. I shared my thoughts on how to leverage my work to get responses from his peers, treating it as a campaign and suggesting the CIO approve budgets in order of submission. I was soon talking with the CIO about overall IT strategy and the adoption of emerging technologies. Among several topics, I shared my recently developed concept for in-store 3D printed hard-to-find parts. He introduced me to the VP of Retail Strategy who showed me how similar my concept was to theirs, and he added me to his team as an “advisor”. When the Voice Transformation initiative was announced, I was asked to help write the RFP, and then asked to create the evaluation framework, both adopted as best practices by the customer. I implemented a firewall between my role and sales to prevent any impropriety. AT&T won the $35M contract, growing account revenue by 10x in less than nine months. Prior to my assignment, I’d never worked in the retail industry.

Supporting material

Revolutionary Retail Executive summary · AT&T, October 2014

While at AT&T in 2011, I was asked to participate in an Oxford style debate at the InterOp Enterprise Cloud Summit in Las Vegas where I was presenting on Cloud Economics. The resolution: private cloud will always be superior for enterprises to public cloud. I was asked to argue the pro side despite being a strong critic of private clouds, having referred to them as a weak crutch in both my Network World blog and at prior conferences. My opponents: Peter Coffee of Salesforce and Peter Magnusson of Google Cloud fame. I was informed 30 minutes before the debate that my partner had flight issues and wouldn’t arrive in time.

The pre-debate survey was narrowly tilted to the con side at 46 to 44. Truly I had an unfair advantage because I talked to Fortune 500 CIOs and CFOs all the time, my opponents did not. I was already in the thick of the fight, trying to change minds while they were figuring out how to get access. For the debate my platform was simple: security, privacy and cost predictability matter more than the vague promise of lower capital expenditures. I successfully blunted every argument they made. Lower cost. Maybe, but enterprises could get close enough on their own without sacrificing privacy and security. Greater flexibility. Not really, because most enterprises don’t see wild swings in consumption, and those areas that do can be migrated to public cloud when it makes sense. Over the course of 45 minutes our battle raged. When the post-debate votes were tallied, private cloud, the position I didn’t agree with, won: con 30, pro 54, with a few attendees unable to pick a side.

The low profile guy from AT&T bested the experts from Google and Salesforce simply by listening to his customers.

Invited to sit in on a meeting at the request of an audit partner, a large utility's SVP of Smart Grid shared their vision: give customers more granular knowledge of how consumption impacts cost. He walked me through their Smart Grid architecture, single board computers installed inside of meters and connected via a mobile network to the Internet and their data centers. Throughout the day, each smart meter would stream usage data, which in turn would provide up-to-the-minute costs to customers via a mobile application and customer portal. I asked how billing was done today. A mainframe intensive task that took several days to complete. Following up I asked why bring all the data back to overloaded mainframes? To address customer billing disputes. I asked how many customers disputed bills monthly? Less than 5% each month, and they were usually resolved with a credit due to the lack of data. The problem was the cost of moving all the data.

I asked if they had considered a distributed architecture, in which the data and processing sat locally on the smart meters, only communicating the final consumption numbers upstream monthly, with AI to adjudicate disputes.

Over dinner that night, I walked him through a new architecture using AI to deliver real-time cost estimates, manage traffic to minimize cost, and use lower cost off-hours periods for larger network tasks. Costs were minimized, speed of the overall solution maximized, customer needs met, and data available when needed. Over the next several weeks we finalized the approach, which the customer adopted and created enough momentum to shift the entire industry to match.

Supporting material

Controlling the Smart Grid Client deck · PwC, October 2008

Upon arrival at PwC, I was removed from my new hire training in Atlanta halfway through my first day and told to get on a plane to Miami. I joined a team architecting a solution to transfer money more efficiently to South America for a Latin American financial services company. Although hired as a project manager, I had a very technical background and quickly took over as the lead architect. This change in role earned me an invitation three weeks later to a partnership meeting between PwC, Sun Microsystems and iPlanet, a developer of high-end internet backend systems. During the meeting I was assigned as the PwC representative to join a Sun and iPlanet team the following day to discuss building the first Spanish language portal on the Internet for a Chilean phone company, Americatel. My job was to elicit requirements. Perfect.

I arrived at the meeting to find 30 IT professionals including two managers, the only ones who spoke English. I speak no Spanish. Nobody else arrived. Panicked, I called my boss who was dumbfounded. He told me to do my best, that’s all he could ask. I huddled with the managers: what was their expectation. Your job is to walk us through the architecture and design of an Internet Service Provider (ISP). The closest experience I had was using one. Not perfect.

I did about 30 minutes of quick research and phoned a friend from the financial services project to join me the following day to discuss user access and authentication. I introduced myself and was immediately asked which model Cisco router would we be using. Undaunted, I broke the solution space down and we walked through all the components of an ISP and how they worked together over the next six hours. I got four hours of sleep that night as I pored over every document I could find on ISP architecture and design. I arrived the following day, joined by my counterpart from that project, and we completed the workshop. I delivered a design and quote three days later, defended the design at an orals presentation three days after that, and we were awarded a $6M contract to implement 123.com, followed by a $2M extension to build a customized email service on top of the portal, news aggregation, personal web space and calendar service in the first release.

Supporting material

Americatel Project Kickoff Kickoff deck · PwC

Affina was in a difficult position. The company generated $50M a year in revenue and was losing $25M a year to churn as clients decided latest generation technology was more valuable than white glove service. Customers representing about 80% of our revenue had signaled an intent to put their business out to RFP. Having deployed a sales force automation solution that included teaching salesmen who’d never used a laptop, I was tasked by the CEO to retain as many of these customers as I could. He gave me a list of names and access to the company plane.

I interviewed the account manager for each customer, all of them high brand value companies, including our largest customer, who represented nearly half of our revenue. I scheduled meetings with the buyer at each customer and started what a VP called “Brian’s apology tour”.

At each stop I listened to every complaint and concern my customer could voice. When done, I told them we had received the message and were in the process of moving to a new platform with the functionality they demanded. While we would never lower our quality, we would certainly face obstacles and setbacks, and we might even need their help. On that tour I gained the support of every leader, 100% retention of their business, with two even offering to loan me their own IT resources for weeks at a time to help out as needed.

General Motors spun EDS out and in doing so turned over several software licenses, including my two products, WinVLT and CalTools, used during the development of GM cars worldwide. GM had already decided to sunset both as they moved to a new provider of in-vehicle controllers, Siemens, who would also provide a new software suite. Despite several offers to share our intellectual property in the interest of our mutual customer, who strangely still paid my salary despite me being allocated to EDS, Siemens went at development on their own. I sensed impending disaster.

I decided to find a way to keep both products alive, just in case GM changed its mind. I had only support level funding, five months at most, and needed to find new revenue sources. I reached out to newly displaced Delco Electronics and signed a distribution deal bundling our tools with their controllers. I pushed GM international companies to buy licenses and support, explaining the tool had been sunset and EDS now owned the rights. Through this effort I discovered engineers at two foreign auto manufacturers were illegally using our software, and reached out to offer licensing agreements in the form of cease and desist letters.

While negotiating one of those deals, I was pulled into a meeting with the lead engineer for GM vehicle calibration who informed me the new supplier had missed every deadline and they were now past the point of no return. He lamented that WinVLT and CalTools were dead. Well, not really. Within 60 days we had a new contract that took us from subsistence to full funding and completed the overhaul and migration to 32-bit windows in 16 weeks. As a team, we generated more revenue per head than any other group at EDS, by a wide margin.