Skip to main content

Applied Politics

The relational and political navigation that determines whether an applied AI deployment actually lands. The third role most lists miss.

The frame surfaced for AAS via Andros Wong, citing a Steve Bartlett Diary of a CEO clip on the irreplaceable roles inside an AI-era organization. The taxonomy is three:

  1. The domain expert / manager who knows the work the agents are being asked to do.
  2. The AI agent whiz who is fluent with the harness, the model, and the architecture.
  3. The relational and political glue who can read the room, build the coalition, and shepherd the decision through the organization.

Most public taxonomies of "AI roles" name the first two. They omit the third. The omission is structural: people who can do the third are usually too busy doing the third to write a LinkedIn post about it.

This doc names the third role and treats it as a legitimate applied AI competency that sits alongside technical skill and domain expertise.

What Applied Politics Is​

Applied politics is the discipline of moving an applied AI deployment through the actual social, organizational, and political reality of the room it has to land in. It includes:

  • Reading which stakeholders have informal power and which have formal title.
  • Identifying which department wants the change, which is threatened by it, and which is indifferent.
  • Building a coalition of supporters before any kickoff meeting.
  • Sequencing rollout so the wins land in front of the people who will fund the next phase.
  • Naming the politics in private rooms so the politics do not blow up the project in public ones.

This is a load-bearing skill, often dismissed as "soft skill," that determines whether technically-correct AI work creates value or sits on the shelf because the organization rejected it.

Why Most Taxonomies Miss It​

Three reasons:

  1. It is invisible from the outside. The technical work and the domain knowledge are legible. The political work is the work that happens between meetings, in side conversations, in the careful sequencing of who hears what when. The output is "the project landed." The mechanism is hidden.
  2. It does not map to a job title. "AI Politician" is a role nobody puts on a business card. The function shows up under a dozen different titles: Chief of Staff, BD Lead, Strategic Advisor, Head of Implementation, Account Executive, Community Leader.
  3. The crypto era trained a generation to disrespect it. "GM, we are all in this together" worked when the industry was small and idealistic. It stops working the moment money and incumbents are involved. Applied politics is the recognition that grown-up institutional reality requires institutional skill.

The omission is expensive. Applied AI work that ships without political navigation has a high failure rate, and the failures are misread as "the technology was not ready" when the actual issue was that nobody got the deal through committee.

The Three Shapes​

Applied politics shows up differently in different roles. Three patterns are worth naming.

In the Applied AI Consultant​

A consultant entering a corporate deployment has to navigate the existing power structure of the client. Who is the actual decision-maker? Who has veto power they will not announce? Which department is paying for the project, and what does that imply about who gets to set scope? Which middle manager is going to be threatened by the rollout, and how do you neutralize the resistance without humiliating them?

The technical work is real, and it is the smaller part of the engagement. The larger part is shepherding the deployment through the political terrain so that on the day the system goes live, the relevant humans have all already agreed it is going to work. See Applied AI Consultant for the role overview.

In the Community Leader​

A chapter leader running an Applied AI Society chapter in a new city has a different version of the same skill. Who are the existing players in the local AI scene? Which ones are you allying with and which ones are you complementing? How do you build the chapter without alienating the meetup organizer who has been running things for three years? Which local business owners need to hear about you from a peer, and which need to hear about you from you?

This is applied politics at the community scale. It is what separates a chapter that becomes the connective tissue of a city's applied AI economy from one that fizzles inside six months because nobody warned the founder which fights to skip. See Community Leader for the role.

In the Chief AI Officer​

A CAIO embedded inside a company has the hardest version. They have to build a coalition across executive peers, navigate the threat the role poses to incumbent CIOs and CTOs, manage the ambition of an enthusiastic CEO without overpromising, and shepherd the workforce through real change without losing the people whose tacit knowledge is the most valuable input to the AI deployment. The political surface is everywhere at once.

The CAIOs who succeed do it because they treat applied politics as a central skill, on the same level as the technical decisions. The technical decisions matter. The political decisions decide which technical decisions get made. See Chief AI Officer for the role.

What This Means For Practitioners​

If you are early in an applied AI career, the highest-leverage skill to develop alongside harness fluency is the ability to navigate institutional rooms. That includes:

  • Spending real time learning how decisions actually get made in your target environment. Not how org charts say they get made.
  • Building a personal network of people across functions who will tell you the political truth about a deployment before you walk in.
  • Practicing the side-conversation move. Most applied AI deployments are decided in the conversation right after the official meeting ended.
  • Developing the discipline to refrain from undermining an executive in public, even when you are right and they are wrong. The cost of the public correction is more than the deployment is worth.

If you are evaluating whether an applied AI engagement is set up to succeed, ask who the political owner is. If the answer is "we have not assigned one," the engagement will struggle.

Applied Politics Is Not Empty Politics​

Applied politics is the operating discipline of moving real work through real institutions. The empty version of politics optimizes for personal advancement at the expense of the work. The applied version uses institutional skill to make sure the work actually happens.

The distinction matters because some technical practitioners reject "politics" as a category. What they are rejecting is the empty version. The applied version is something they need.

Further Reading​