How MakerDAO scaled community ops and technical docs across global timezones.

MakerDAO is a DeFi protocol with one of the longest-running and most distributed communities in crypto. The user base spans North America, Europe, East Asia, Southeast Asia, and Latin America. The Discord server runs at full activity across every timezone. The Telegram presence covers regional pockets where Discord is not the dominant chat channel. The technical documentation is the canonical reference for builders, governance participants, and integrators across that entire footprint. The community ops and docs functions together carry the weight of how the protocol is understood, adopted, and built against.

The structural problem at the community ops layer was the timezone gap. The core contributor base sits primarily in US and European business hours. Discord questions arriving from Tokyo, Singapore, Manila, or Seoul at 2am Eastern time would sit in the queue for hours, often answered the next morning by a contributor who had to context-switch back into the original conversation. The same pattern hit the Telegram regional pockets where non-English questions sat unanswered for half a day at a time. The community experience was structurally worse for users outside the core contributor timezones, which is a problem when a protocol is trying to build adoption in those exact markets.

The docs problem was different in shape and similar in cause. The protocol ships changes frequently. New collateral types, governance updates, parameter changes, integration partnerships. Every change has documentation implications. Builders and integrators rely on the docs being current. The labor required to keep the docs current with the protocol cadence was structurally larger than the docs team could carry. Articles aged. New developer onboarding material lagged the actual state of the protocol by weeks at a time. Builders asking the same recurring questions in Discord were a symptom of docs that had not caught up to where the protocol had moved.

MakerDAO came to EOI looking for a fractional engagement that treated community ops and docs as one system instead of two. The structural insight from the engagement was that most of the Discord questions arriving at 2am Eastern were tier-1 protocol questions with answers that already lived in the docs. The labor that was being spent answering those questions in Discord could be redirected into keeping the docs current, if there was an AI layer answering the Discord questions against the canonical docs in real time. The two problems were two sides of the same content function.

The shape of the engagement that worked was the AI Content Department applied to the community-plus-docs surface. Agents trained on the protocol documentation, the governance forum archive, the historical Discord question patterns, and the technical reference. Multilingual coverage so the Telegram regional pockets got native-language responses instead of translated approximations. A docs cadence that kept the canonical reference current with protocol changes, with the agents drafting explainers against governance forum threads and the human team approving and publishing. The broader case for this kind of content consolidation sits in What is a Fractional AI Department and the crypto-vertical version of the engagement sits at AI for Crypto + Web3.

The other reason the fit was clean was the protocol-specific knowledge requirement. A generic moderation bot would have shipped wrong answers on collateral mechanics, governance procedures, and risk parameter logic, because the protocol surface is too specialized for off-the-shelf models to handle accurately. The agents got trained on the canonical references, the governance archives, and the technical specifications. The accuracy threshold for protocol questions held because the corpus was protocol-specific, not general crypto knowledge.

The most immediate result was the timezone gap closing. Discord questions arriving from Tokyo, Singapore, Manila, or Seoul at 2am Eastern got answered in seconds instead of hours. The regional community experience that had been structurally worse than the core US and European experience caught up. CSAT-equivalent measures across the Discord moderation flow improved measurably in the first month. Users in the regional pockets stopped feeling like second-tier audiences and the adoption pattern in the markets the protocol was trying to build into reflected that.

The docs side compounded more slowly and matters more for the long-term protocol health. The lag between protocol changes shipping and the docs catching up collapsed from weeks to under a week. Builders relying on the docs got the right answer when they looked. New developer onboarding material stayed current with the actual state of the integrations. The recurring questions in Discord that were symptoms of stale docs started dropping as the docs caught up. The feedback loop between community and docs became a real loop instead of a theoretical one.

The structural result is the one that matters most for a DAO operating model. The core contributor labor that had been glued to tier-1 community moderation got freed for governance work, risk parameter analysis, and the strategic protocol work that needed contributor judgment. The contributor capacity that was being consumed by Discord question-answering moved into the work that advanced the protocol. The community ops layer stopped being a labor sink and became a feedback signal feeding governance and docs both.

The deeper lesson is that DeFi community ops and technical docs are one system at the protocol level. Most protocols staff them as two functions because the labor markets organize them as two functions. Running them as one fractional engagement collapses the latency on both sides, closes the timezone gap, and produces a feedback loop where community questions feed docs updates that deflect future community questions. The same pattern works across every protocol and on-chain product we work with, which we have detailed at AI for Crypto + Web3 and the matching AI Content Department page.