Many agencies I know depend on a growth engine they cannot steer: referrals. A frontier agency needs a second engine, one based on public thinking, useful participation in communities, and a clear view of where its industry is going next.
This was written with AI based on my voice memos.
TLDR: Invest in meditation and communities!
Referrals are valuable. They are also irregular, difficult to forecast, and difficult to scale predictably. You can do excellent work and still discover that a satisfied client has nobody to introduce you to.
If you are fortunate, referrals support growth for years. Until the well dries up without warning.
In our decade of existence, I will not name how many agencies have gone bankrupt.
This is why people turn to content so they become visible. But publishing is no longer a differentiator by itself. Companies publish, founders maintain feeds, and I see generative AI making competent content easier to produce. Producing more of the same is not a strategy.
The frontier agency should become a content thinker, not merely a content creator.
A content thinker is a practinioner, not a theorist. I will explain why and how…
Referrals are proof, not a growth system
A referral proves that somebody trusted your work enough to attach their reputation to it. The mistake is to lean too much on it.
Waiting for referrals is reactive. Engineering a system to get them is proactive.
A system gives you actions you can repeat, signals you can observe, and choices you can improve. A referral pipeline alone gives you none of those reliably. You cannot decide when somebody will remember you, who they will meet, or whether that person has a problem you can solve.
The answer is not to abandon referrals. It is to stop asking them to carry the entire business. Keep earning them through good work, but build a parallel system that helps the right people encounter your judgment before they need your services.
Content creators follow attention. Content thinkers create direction.
Remember the stoics? They were not fancy jugglers, circus attention-grabbers, or theorists.
A content creator usually begins with the feed: What is trending? Which format is performing? What will the algorithm reward this week? Those questions can improve distribution, but they rarely move a business or a category forward.
A content thinker begins somewhere else:
- What is wrong in our industry today?
- Which belief is true now but unlikely to remain true in five years?
- Where is value moving as technology changes the cost of execution?
- What have we learned that could help our partners make a better decision?
The resulting content is not commentary attached to a business. It is part of how the business learns. It clarifies a position, attracts disagreement, finds future partners, and exposes weak assumptions while there is still time to change them.
I think of this as research and development in public. The goal is not to sound certain about the future. The goal is to develop a more useful point of view by testing it where other people can challenge it.
Communities can become a compounding growth engine
One strong alternative to referral dependency is not a larger broadcasting machine.
When people ask me “how to become a content thinker” they believe the effort should focus on volume.
Actually, I think it should be more like a laser than a flashlight first. Broadcasting should go deeper into participation in communities.
That does not mean entering a group, dropping a link, and waiting for leads.
It means joining as a participant and volunteering useful work without calculating the immediate return. Answer a difficult question. Share a framework. Introduce two people who should know each other. Help an organizer improve an event. Give a strategic partner a stronger way to explain what they do.
Some of those contributions may eventually become paid projects. Many will not. Contributing without making a transaction the condition can help build trust because the work is useful before a commercial relationship exists.
The most interesting communities may sit outside your immediate commercial bubble. A technology agency can contribute to groups organized around entrepreneurship, education, cities, sport, travel, or independent work. Entering with a real capability and enough humility to listen creates connections that a narrow industry feed cannot.
This is consistent with how we think about shared leadership at Camplight. Value does not have to flow from a central authority. It can emerge from people contributing different strengths to a common problem.
Why I expect the agency model to move from hours to outcomes
Public thinking only matters if it helps the agency anticipate structural change. One hypothesis I am testing is that software agencies will see a declining connection between hours worked and software produced.
Previously only premium agencies did outcome-based partnerships. I think this will become the status quo in 2027+.
The traditional outsourcing model sells access to human capacity. More scope usually requires more people or more time. An AI software factory can change that relationship by connecting product context, coding agents, verification, release, and feedback into a managed delivery system.
I do not expect outsourcing to disappear. I do expect selling person-hours to become a less dominant form of collaboration. I expect more clients to buy an outcome, a capability, or a continuously improving system. Agencies that make this shift will be responsible for designing and operating the factory that produces that result.
That is the current frontier, but a content thinker cannot stop there.
If software factories become normal over the next five years, what becomes wrong with the software-factory category itself? Asking that question now is more valuable than spending five years defending today’s positioning.
What becomes scarce after execution becomes abundant?
I see four candidates.
1. Strategic alignment
A factory can produce quickly and still produce the wrong thing. The more execution accelerates, the more important it becomes to align every system with a coherent product and business strategy. Speed magnifies direction, including bad direction.
2. Taste and judgment
Most meaningful problems have several technically valid answers. The hard question is which answer belongs to this company, for these customers, at this moment. When production becomes easier, a recognizable signature becomes more valuable. Someone still has to decide what is clear, appropriate, humane, and worth making.
3. Collective attention
Our attention is already fragmented. If autonomous systems continue to accelerate execution, I expect more options, updates, and decisions to compete for it. Teams will need the ability to return to a shared problem and hold it long enough to think coherently.
This is why I consider meditation a practical operational training, not a wellness accessory.
Meditation trains us to return our attention after distraction.
Organizations need a collective version of that skill: the ability to notice drift, restore context, and continue from the same strategic center.
4. Coordination between autonomous systems
I expect software factories to work increasingly with other software factories. Shared agents may connect engineering, marketing, sales, operations, and support.
Agents with wallets and agent-to-agent protocols will unlock a bigger, faster and more continuous economy than what we had before.
The difficult work will include defining boundaries, authority, evidence, and escalation across those systems. The right level of autonomy will depend on the task and its consequences, not on a universal promise of full automation.
My five-year operating system for a frontier agency
Nobody knows exactly how this transition will develop. My response is not to predict harder. It is to build a practice that keeps learning.
- Think publicly. Publish questions, models, experiments, and changes of mind. A useful argument creates more trust than a polished claim with nothing at risk. Create more than you consume!
- Run one content experiment every week. Test a new format, question, community, or frame. Make each experiment spiral upward by carrying its evidence into the next one.
- Prefer timeless ideas to disposable reactions. Trends can start a conversation, but durable principles keep helping people after the feed has moved on.
- Contribute before you ask. Give communities work they can use. Do not disguise prospecting as generosity.
- Choose proximity over vanity. Live conversations, small events, working sessions, and direct exchanges create understanding that impression counts cannot measure.
- Build concentric circles of trust. Kevin Kelly’s 1,000 True Fans is a better mental model than chasing universal reach. For an agency, I would focus on 1,000 people who value our thinking, 100 who actively support it, and 10 who would build something with us.
- Question the category before the market does. Do not merely manage the agency in its current state. Ask what it needs to become, move it there, and only then optimize the new model.
The frontier is a posture, not a destination
Innovation has no fixed boundary, and nobody has a reliable map of the next five years. That uncertainty is not a reason to wait. It is a reason to build an agency that can observe, think, contribute, and reposition itself continuously.
In my view, a frontier agency should not try to publish more noise than everybody else. It should earn attention by making its thinking useful. It should not enter communities only when it needs a pipeline. It should become part of their capacity. And as technology changes the unit of value, it should learn to take responsibility for outcomes instead of defending the billable hour.
I expect referrals and content to keep mattering. But I believe the durable advantage will come from the quality of the thought behind them and the relationships built around that thought.
What is accepted as true in your industry today that will no longer be adequate five years from now? That is where I would start thinking in public.
