How Haymaker put three programs on autopilot: awards, media lists, and commentary tracking

Agency work has a hidden tax. The visible work is strategy, pitching, relationships, and coverage. The invisible work is everything underneath: finding the right awards before deadlines pass, keeping media lists current and deduplicated, verifying reporter emails and author pages, watching who is writing what across a dozen story categories every day. It is repetitive, detail-sensitive, and does not scale. Add a client and you add all of it again.

Three programs carry most of that tax at any earned-media agency. Haymaker put all three on Shadow. Awards and events run as a research-and-application pipeline. Media lists build and maintain themselves. Commentary tracking runs unattended every morning and lands in a team channel before the team logs on.

By hand, awards discovery means hunting for relevant programs, checking eligibility, chasing down deadlines, then writing each brief and application from scratch. On Shadow it runs as a staged pipeline. Every stage exists to protect the two things awards work turns on: eligibility and dates.

The program surfaces only eligible, real, in-window opportunities, then drafts the paperwork to go after them from the client's own materials, across the whole roster at once when needed.

Media lists sit on a three-layer model: live research across a large database of journalist profiles, the agency's own contact intelligence built across every client, and the client-facing list itself. New contacts found on any campaign flow back into the agency's institutional knowledge, so the network compounds instead of resetting per project.

The bigger shift: lists no longer decay. For the featured client, a scheduled agent runs every morning. It pulls newly observed journalists from the coverage graph, enriches each one to a full verified profile, and adds only the complete ones to the client list. It is append-only, so it never removes existing contacts, and it pushes the finished file straight to the shared drive. The list is current when the team opens it, without anyone touching it.

List-building stops being a per-campaign chore. It becomes a standing asset that maintains itself and gets richer across the whole client base.

The clearest proof is a single featured client, where all three programs run together, live, on a schedule. Early each morning the narrative graph refreshes, the living media list adds newly verified journalists, and the commentary tracker updates and posts its digest. Three coordinated agents read from a shared source layer and produce a current media list and a morning intelligence digest every day. No one is at the keyboard.

The pattern is the same in all three. Work that is repetitive, per-client, and unforgiving is now a standing program that runs itself, enforces its own quality bar, and compounds instead of resetting. Awards work loses the deadline risk, lists stay current, and tracking runs every day regardless of what else the day holds.

None of it needs an operator. The programs run on their own and hand back finished work, and the hours the tax used to claim go back to strategy, relationships, and the pitch.