Every business leader knows the invisible tax that eats away at productivity, innovation, and trust. That tax is corporate politics, and it is more corrosive than most operational problems because it hides in plain sight.
The pattern is familiar across organisations of every size. Vital context sits trapped in private email chains. Gatekeepers hold on to information to protect their turf. Decisions get shaped by internal lobbying rather than by clear data. Over time, healthy teams quietly erode. High performers grow tired of playing games, information silos multiply, and strategic momentum stalls, often without anyone being able to point to the exact moment it happened.
Leaders often reach for cultural fixes when this shows up: town halls, workshops, offsite retreats. Culture matters, and none of these efforts are wasted. But they rarely touch the root cause. Politics thrives wherever data is fragmented, subjective, or held back. When access to truth depends on who you know, power follows proximity rather than merit.
Removing the gatekeepers
This is where the conversation on enterprise AI needs to move on from drafting copy and automating routine tasks. Autonomous AI agents and unified data platforms are changing how information moves through an organisation. They act as a neutral, accessible layer of operational truth that does not answer to any single department.
Information hoarding becomes harder to sustain once customer interactions, pipelines, service histories, and operational metrics live in one connected system. Knowledge stops working as a bargaining chip. Anyone can query the system directly, without waiting on a gatekeeper’s permission.
Evaluation built on impressions and personal narrative starts to lose ground too. When performance reviews, resource allocation, and handoffs run on real time data rather than on who tells the better story, political posturing has far less to work with.
Execution becomes more standard as a result. Autonomous agents run cross functional workflows according to business logic and customer need. Department rivalry and hierarchy have little influence over how the work actually gets done.
A surprising case in point
I recently came across an AI use case that caught me off guard. A company looking to improve customer outcomes removed human call centre managers from on the job coaching entirely and handed that role to AI.
The decision itself was not what stood out to me. What stood out was why the agents responded so well to it. To them, AI coaching felt objective. It did not colour feedback with personal bias or favouritism, and it did not base an assessment on one call or one bad day. There was no manager mood to read, no politics to navigate, and no fear that a single misstep would follow them around for months.
The result was a genuine gain in psychological safety. Agents stopped worrying that the AI would judge them personally or carry that judgment outside the coaching framework. Once the human tendency to play favourites or hold grudges was taken out of the equation, the feedback became something agents could simply trust and act on.
It is a small example on its own, but it points to something larger. When the system giving feedback has no turf of its own to protect, people stop performing for the evaluator and start focusing on the work in front of them.
Technology alone is not enough
An AI system cannot fix a toxic culture by itself, and it would be a mistake to expect it to. If leaders keep rewarding political behaviour, or override objective insight to favour allies, no amount of new technology will save them from the consequences of that choice.
What AI does do well is make the cost of politics visible. Once the data sits in front of everyone, ambiguity stops being a hiding place. Bypassing a data backed recommendation, or quietly sitting on a resource someone else needs, becomes a visible choice. It demands an explanation in a way it never did before.
The goal: Leadership built on trust
AI in the workplace is not really about replacing human judgment. It is about making room for better leadership to emerge. As routine coordination, data aggregation, and status reporting move over to objective systems, leaders find themselves spending less time managing friction between people.
That freed up time tends to go toward the parts of the job that were always the point. Mentoring people. Working through hard customer problems. Building the kind of psychological safety that real trust depends on.
Strip out the gamesmanship, even partially, and people are left with the space to do the best work of their careers. That is a modest promise on its face, but for most organisations, it would be a significant shift.
- Linda Saunders, Country Manager & Senior Director, Solution Engineering Africa at Salesforce
