Colleagues coordinate beneath towering stacks of paperwork, where abundant information makes coordination the real work

When Intelligence Gets Cheap, Coordination Becomes the Work

For a long time, producing useful information was expensive.

If an advocacy campaign wanted to understand a complicated regulatory proposal, somebody had to read it. If a coalition needed a briefing document, somebody had to research and write it. If organizers wanted different messages for different constituencies, a communications team had to develop them. Research, synthesis, writing, analysis, design and even basic administrative work consumed time and money, and that constraint shaped organizations. A lot of capacity was built around producing things.

We are beginning to operate under a different set of conditions. Artificial intelligence can already summarize a lengthy document, compare competing arguments, draft a memo, generate a dozen versions of a message and analyze a large body of text in the time it once took somebody to get started. The quality varies, and plenty of the work still requires human review, but the direction is fairly clear. Producing plausible analysis and content is getting much cheaper.

The interesting question for advocacy is what happens after that. If every organization can produce more research, more messages, more strategies, more summaries and more content, the supply problem starts to recede. We do not necessarily become better at making change. We become capable of producing far more material related to making change, which means some other constraint has to absorb all of that new capacity.

In practice, we can already see where many of those constraints are. Someone still has to decide which information matters. People have to determine what they believe, which strategy deserves attention, whose interpretation they trust and which opportunities are worth pursuing. Groups with overlapping interests have to figure out where they can cooperate. Someone has to know who else should be involved. Leaders need access to people who can make decisions. Communities need reasons to participate. People who disagree still have to find enough common ground to work together.

Those things have always mattered. As information and intelligence become easier to produce, they matter more because scarcity has a way of moving.

We wrote recently about one part of this shift in Review Is the Product, where we looked at how cheap content makes validation, legitimacy and situated judgment more valuable. That still feels right, but review is only one place where the bottleneck is moving. We are starting to see the same shift play out across the larger work of building and coordinating networks.

The work moves downstream

Think about a fairly ordinary advocacy situation. A new policy proposal appears that could affect communities across several states. Ten years ago, simply understanding the proposal might have consumed a meaningful share of a small organization’s capacity. Staff would divide up the research, call a few experts, review earlier policy, develop a summary and eventually circulate an analysis.

Now imagine dozens of organizations involved in the issue, each with access to tools capable of producing a credible initial analysis within minutes. That represents a real gain in capacity, but it also changes the nature of the work. There might now be fifty analyses instead of five. Each organization can develop its own strategy, its own language, its own list of implications and its own recommended response. Every participant can produce more material for every other participant to consider.

The difficult work starts shifting from production toward interpretation. Which analysis reflects what people are actually seeing on the ground? Where are the important disagreements? Is an apparent conflict substantive or simply the product of people using different language? Who has experience with a similar fight? Which risk is urgent and which can wait? Which organization is positioned to do something useful? Who is already talking to the relevant decision maker? Where would another intervention add value, and where would it simply duplicate work already underway?

These are difficult questions because there usually is not one correct answer waiting to be discovered. They involve context, judgment, relationships, timing and knowledge distributed across many people.

AI can contribute to that work. It can probably help us see patterns that would otherwise remain buried. But increasing the analytical capacity available to every participant does not resolve the problem of deciding what a group of people should do together. In some circumstances it may make that problem harder. If ten people each arrive with one possible course of action, a group has ten ideas to work through. If those same ten people can each arrive with twenty credible options, they have created a much larger coordination problem.

More intelligence at the nodes does not automatically create a more intelligent network. We have spent much of the past two decades at Netcentric Campaigns working inside that distinction.

Networks have never been information systems alone

It is tempting to look at a network and see communications technology. There are email lists, databases, websites, group chats, directories, social platforms and whatever tool comes next.

The technology matters, but when we assess whether a network is actually useful, the questions quickly become human. Do people know who else is involved? Can they find the person with the expertise or relationship they need? Are participants comfortable asking each other for help? Is there enough trust to exchange useful information rather than carefully polished talking points? Is there common language that allows people from different organizations to understand one another? Can people see what others are learning? Do new participants understand where they fit? Can someone with an idea find other people willing to work on it?

Those questions became the basis for much of our network practice because we kept seeing the same pattern across different issues. The value of a network is partly what is available through it, but as a network matures, the value increasingly becomes who is available through it.

We often describe this as a progression. People may initially come to a network for a service, but over time they stay for the people. Eventually, the people themselves become part of the promise.

That shift is easy to underestimate. A useful database can be copied. A briefing service can face competitors. A new technology can make an existing service cheaper almost overnight. A trusted relationship between two people working on the same problem cannot be replicated in the same way.

A network participant who knows which researcher has spent fifteen years studying an issue, which local leader understands a particular community, which organizer can get five groups into the same room and which funder is actually willing to consider an unconventional idea has access to something much richer than information. They have access to a functioning social system.

The strongest networks create more of those connections over time. People contribute knowledge, experience, introductions, feedback, resources and credibility. Other participants make those contributions more useful by connecting them to needs elsewhere in the network. That kind of value becomes more important in an environment where basic information is plentiful.

Cheap content makes judgment expensive

There was already too much information before generative AI arrived.

People working in advocacy have never suffered from a shortage of reports, newsletters, webinars, opinion pieces, policy analyses, conference notes, strategy documents and requests to read one more important thing. The challenge was often finding the useful signal inside all of it. Now we are adding machines capable of producing dramatically more.

One reasonable response is to build better filters. Some of those filters will themselves be powered by AI. We will have tools summarizing other tools, agents watching information streams and increasingly sophisticated ways of searching large bodies of material. Filtering helps, but it does not eliminate judgment. It changes where judgment gets exercised.

A model can rank documents according to criteria we give it. Someone still has to decide whether those criteria make sense. It can identify a pattern across hundreds of reports. Someone needs enough context to know whether that pattern should change strategy. It can recommend five possible messages. Someone still has to understand which language will land with a community whose trust was earned long before the model entered the picture.

Much of good advocacy judgment is accumulated socially. People know who has been right before. They know who understands a place or institution. They know which person is cautious for good reason and which person always sees catastrophe coming. They know who can disagree without blowing up the collaboration. That information rarely appears cleanly in a database.

The same is true of attention. We can generate unlimited things that deserve attention, but a city council member still has the same number of hours in a day. A journalist can only pursue so many stories. A foundation still has to choose among competing investments. An organizer can only sustain so many relationships. A community cannot attend every meeting or respond to every campaign.

As production expands faster than attention, the ability to direct attention becomes a form of power. Trust is one of the ways people manage that scarcity. We use trusted relationships to decide what to open, which invitation to accept, what warning to take seriously and when to act quickly without independently verifying every piece of information.

Healthy networks make those trust pathways stronger and easier to use.

Coordination is not command

There is another wrinkle here that matters for advocacy. When people hear “coordination,” they often imagine a central organization making decisions for everybody else. That model can make coordination faster in the short term, but it can also concentrate power, suppress useful experimentation and create a fragile system where too much depends on a small number of leaders.

Most of the networks we build and nurture do not function that way.

A strong advocacy network has many leaders. Power is pushed to the edges. People act from different positions, with different resources and different forms of authority. One person may understand the policy. Another has community relationships. Someone else can communicate publicly. Another participant knows how to get resources moving. People shift roles depending on what the situation requires.

The job of network infrastructure is to make distributed leadership usable without turning the network into an organization with a single chain of command.

That requires some shared conditions. People need ways to communicate across organizational boundaries and enough common language to understand what others mean. They need feedback from different parts of the network so strategy can respond to what is happening. Shared resources have to be discoverable. Social ties need enough strength to survive disagreement. Participants need clarity about the larger purpose even when their individual approaches differ.

Those are among the Seven Elements of an Effective Network that we have used for years: leaders, common language, communications grids, feedback mechanisms, shared resources, social ties and clear vision. They were developed long before the current AI moment, and their relevance is now stronger than ever.

A communications grid without trust can become a torrent of material nobody can evaluate. Shared resources without navigation become clutter. Distributed leadership without common language can produce confusion. Feedback without a clear vision can pull a network in a hundred directions. Strong relationships without mechanisms for assembling resources may create a wonderful community that has difficulty exercising power.

Network capacity comes from the interaction among these things.

The abundance problem is a civic problem

The same pressures that show up inside an individual advocacy network do not stop at the network boundary. When more people and organizations can produce more information, analysis and activity, the challenge of connecting that capacity starts to become a civic infrastructure problem.

That is part of what we are thinking about through our Networked Nation initiative. Our civic capacity has not kept pace with the pressures being placed on it. Trust has weakened, information environments have become harder to navigate, institutions and organizations are stretched thin, and the social connections that help people work through disagreement have frayed. Networked Nation is an effort to think seriously about the infrastructure needed to rebuild that capacity and help people work together across those conditions.

The encouraging part is that the underlying civic energy has not disappeared. There is no shortage of people who care about what happens in their communities. There is no shortage of organizations working on democracy, climate, health, justice or other public problems. There is certainly no shortage of communication. What is harder is connecting all of that activity so people can find one another, learn from one another and assemble enough capacity to act when it matters.

AI raises the stakes because it increases what individual people and organizations can produce. A small community group may soon have analytical and communications capabilities that once required a much larger institution. Local leaders will be able to process complicated information faster. Volunteers will be able to create materials and conduct research that previously required specialists. People who have been excluded from certain forms of technical work may gain useful new capabilities.

There is real promise in that distribution of capacity, along with a risk of extraordinary fragmentation if the connective infrastructure does not improve alongside it.

Imagine thousands of civic actors who are all individually better equipped to analyze problems, formulate strategies and communicate their views, but who remain poorly connected to one another. They can produce more without necessarily sharing more. They can act faster without knowing what others are doing. They can build competing interpretations of the same situation. Established organizations can become overwhelmed trying to absorb the output of newly capable participants.

The design problem for Networked Nation is therefore larger than providing people with better technology. We need an environment where distributed capability can become collective capacity.

That means helping people locate expertise and relationships across boundaries. It means making the network navigable as it grows and creating ways for local observations to travel without requiring everything to pass through a central institution. It means maintaining an invitational culture so new people can enter rather than finding a civic system that feels finished and controlled by insiders.

It also means getting better at assembly. A useful network does more than tell us that 5,000 people possess useful resources. It helps the right combination of those people find one another when a problem demands what they collectively have.

That assembly function may become one of the defining pieces of civic infrastructure in an age of abundant intelligence.

What remains stubbornly human

There is a tendency in conversations about AI to sort work into things machines will do and things humans will continue doing. That framing may be too clean for what is actually happening. A more useful question is how the value of different kinds of human capacity changes when some forms of cognitive production become cheap.

Years of network work give us some places to watch. Judgment becomes more valuable because somebody has to interpret an expanding volume of analysis in context. Relationships matter because information moves differently through people who know and trust one another. Reputation helps people decide whose judgment deserves weight. Attention remains bounded even as the supply of material competing for it expands. Access matters because many institutions and decision makers remain finite no matter how much information is produced about them. Coordination matters because distributed capability only becomes collective power when people can align enough of their activity to matter.

Trust runs through nearly all of those things, and we should think of them as infrastructure rather than soft additions around the edges of serious strategy.

We have seen networks struggle when they stop creating value for new people, become too difficult to navigate or lose the ability to aggregate resources into useful action. We have also seen what happens when people can find one another, develop shared language, exchange useful resources and build relationships strong enough to support action. The network itself begins to generate a kind of capacity that no participant owns alone.

That is worth keeping in view as we become capable of producing more intelligence than any advocacy effort has ever had available. The next generation of civic work will certainly use better machines, but the harder challenge is building better ways for people to think, decide and act together.

What Are You Seeing in Your Network?

If you are seeing this shift in your own network, whether through information overload, coordination challenges, difficulty finding the right people and resources, or simply more activity than your existing infrastructure can absorb, we would be interested in comparing notes. Netcentric Campaigns works with advocacy leaders to strengthen the relationships, systems and network capacity that make collective action possible. Get in touch and tell us what you are seeing.