TThough CIOs’ in uence and prestige have grown markedly over the last decade, the primary source of their credibility continues to lie in maintaining e cient, reliable IT operations. This is, by any mea- sure, a full-time job. Yet along with that responsi- bility, they are expected to harness emerging tech- nology forces. They stay ahead of the technology curve by absorbing the changes that leading-edge tools introduce to operational, organizational, and talent models. Finally, an ever-growing cadre of en- terprise leaders with “C” in their titles—think chief digital o cer, chief data o cer, or chief algorithm o cer—demand that CIOs and their teams pro- vide: 1) new products and services to drive revenue growth, 2) new ways to acquire and develop talent, and 3) a means to vet and prototype what the com- pany wants to be in the future.

As growing numbers of overextended CIOs are realizing, the traditional operating model that IT has used to execute its mission is no longer up to the job. Technological advances are creating en- tirely new ways of getting work done that are, in some cases, upending how we think about people and machines complementing one another. More- over, the idea that within an organization there are special types of people who understand technology and others who understand business is no longer valid. Technology now lies at the core of the busi- ness, which is driving enterprise talent from all op- erational areas to develop tech  uency.5

The time has come to build a new operating model. As you explore opportunities to reengineer your IT shop from the top down, consider the fol- lowing areas of opportunity:

• Reorganizing teams and breaking down silos: In many IT organizations, workers are organized in siloes by function or skillset. For example, network engineering is distinct from QA, which is di erent from system administra- tion. In this all-too-familiar construct, each skill group contributes its own expertise to di erent project phases. This can result in projects be-

coming rigidly sequential and trapped in one speed (slow). It also encourages “over the wall” engineering, a situation in which team members work locally on immediate tasks without know- ing about downstream tasks, teams, or the ulti- mate objectives of the initiative.

Transforming this model begins by breaking down skillset silos and reorganizing IT workers into multi-skill, results-oriented teams. These teams focus not on a speci c development step— say, early-stage design or requirements—but more holistically on delivering desired outcomes. A next step might focus on erasing the boundar- ies between macro IT domains such as applica- tions and infrastructure. Ask yourself: Are there opportunities to share resources and talent? For new capabilities, can you create green eld teams that allow talent to rotate in or out as needed? Can some teams have budgets that are commit- ted rather than  exible? The same goes for the siloes within infrastructure: storage, networks, system administration, and security. What skill- sets and processes can be shared across these teams?6

A  nal note on delivery models: Much of the hype surrounding Agile and DevOps is merited. Reorganizing teams will likely be wasted e ort if they aren’t allowed to develop and deliver products in a more e ective way. If you are cur- rently testing the Agile-DevOps waters, it’s time to wade in. Be like the explorer who burned his boat so that he couldn’t return to his familiar life.

• Budgeting for the big picture: As functional silos disappear, the demarcation line between applications and infrastructure fades, and pro- cesses replace tasks, IT shops may have a prime opportunity to liberate their budgets. Many older IT shops have a time-honored budget planning process that goes something like this: Business leaders make a list of “wants” and cat- egorize them by priority and cost. These proj- ects typically absorb most of IT’s discretionary budget, with care and maintenance claiming the rest. This basic budget blueprint will be good for a year, until the planning process begins again