What DAOs Need to Do Next

Most DAO governance conversations in 2026 are still about what went wrong: voter apathy, proposals that died in committee, treasuries that funded nothing anyone remembers. The post-mortem is a well-worn format by now. DAOs have spent the better part of a decade proving that decentralized governance can work. The harder question, the one worth asking in a quieter market, is what it takes to make it work consistently, at scale, over time. As a founder I watched systems stall, and almost never in the design. They stalled in the operation. Here is what actually moves the needle.

The vote isn’t the hard part

The most common failure point in DAO governance is not the vote. It is the silence after it. A proposal passes, the community celebrates, and then nothing moves for six weeks because no one actually owns the execution.

This is cheap to fix and rarely fixed. Every decision record should carry the rationale, the owner and the deadline before the vote even closes. At The Sandbox DAO the author and the milestones were defined before anything went to a vote, so the day a proposal passed, execution already had a name attached to it. That single habit removes most of the post-vote drift people misread as apathy.

Fix intake before you blame voter fatigue

Voter fatigue is real, but it is usually misdiagnosed. The problem is rarely the frequency of votes. Token holders are being asked to judge proposals that were never ready to be judged, with no clear scope, no success criteria and a vague budget, which leaves reviewers to either chase clarifications or vote half-blind.

We solved this at Sandbox with the unglamorous work at the front of the pipeline. My admin team, two project managers, tore every proposal apart before it reached the community, pushing on scope, budget and feasibility until it held up. A separate special council, nominated at the DAO’s creation, gave each proposal a lighter read and a simple signal: positive, negative or neutral. By the time something reached a vote, the hard scrutiny had already happened, and the vote was a decision rather than an investigation.

Delegate with real authority, not just more votes

Sending every decision to a token holder vote sounds democratic and breaks down in practice. When everything needs a vote, nothing moves quickly, and the contributors who care most disengage first. The answer is delegation that actually delegates, not a retreat from decentralization.

Our grant program is the clearest example. The Sandbox DAO Grant Program put four elected grant managers in charge of approving proposals in their own fields, up to fifteen thousand dollars each, with no community vote on any individual grant. The community voted on two things only: the program itself, and who the four managers would be. Everything inside those guardrails moved at the speed of one accountable person rather than a token poll, while my admin team ran the day to day so the system kept its rhythm between votes. The word that makes this work is accountable. Each manager had a defined mandate and had to answer for what they approved. Delegation without that is just diffused responsibility with extra steps.

Use AI in the operational layer, not the decision layer

AI is getting a lot of airtime in these conversations, and some of it is hype. Where it genuinely earns its place is around the decisions, not inside them. It will not exercise judgment on a hard call, but it will strip friction from the work surrounding one: flagging a proposal that arrives missing key information, turning a three hour community call into a briefing the council can act on, keeping context from evaporating when contributors are scattered across time zones. The teams getting value from it are using it to make existing processes faster and more consistent, not to automate the choices themselves.

Measure outcomes, and make the case in the open

Most DAOs measure activity: proposals submitted, votes cast, turnout. Those are easy to collect and tell you almost nothing. The numbers that matter are the time from submission to decision, and the share of approved work that actually gets delivered. At Sandbox, effectively every approved proposal reached execution. The only shortfalls were a couple where the author walked away from their own project before the final milestone, which is a different problem from a system that cannot ship.

There is one more thing worth watching: whether decisions are being made with enough context to make them well. The clearest miss I can point to was a proposal to diversify the treasury into ApeCoin and build a partnership around it. It went nowhere, not because the idea lacked merit, but because the case was never made clearly in the community Spaces and the forums.

A decision is only ever as good as the context the people deciding actually have.

You cannot always plan the ending

The honest version of the wind-down story is that I did not plan The Sandbox DAO’s ending from the start. Almost nobody does, because it feels like planning for failure. What made the eventual wind-down orderly rather than messy was not foresight. It was that the system had been built cleanly and documented thoroughly the whole way through, so when the mandate ended, the records, the processes and the accountability trail were already there to wind down against. The lesson is not to script your ending on day one. It is that disciplined operations hand you a clean exit even when you did not see it coming.

What it really takes

None of this is technically hard. The hard part is doing it consistently, across people, over time, which is exactly what most DAOs are not built to sustain alone. The ones that manage it have someone owning the operational half of the system: not the governance design, which gets drawn once, but the daily running of it, the intake, the accountability, the tracking, the communication. That operational half is the part of GovOps most DAOs skip, and it is the difference between a framework that looks right and one that delivers.

If your governance reads well on paper but stalls the moment a decision has to become action, the gap is operational, and it is fixable. That is the work we do at arasak.io.