AI speed on top of stable, regulated infra.

The two sides of this partnership already do the thing the other side doesn't want to do. Aaron framed it on the April 29 call: "a firm that does what you do doesn't want to do what we do, and we can't do what you do because we're doing all of this." That's why this is clean, not why it's complicated. Trabian isn't positioned to build apps for clients — and shouldn't be. Black Flag isn't positioned to run a bank's core connections — and shouldn't be. Together, the line is bright.

Mesh is, in Aaron's own framing, "the platform that allows for AI speed against compliance and regulated systems" — all the logging, all the connections to cores, payments, and partners, baked in. Jack Henry, Q2, Fiserv, Alkami. Licensing deals to get fintechs into Q2 and Alkami, not just talk to them. That ecosystem is on lock. The stability is what gives an application layer permission to move fast.

Black Flag's job is everything that sits above that line: facilitation, AI use-case shaping, application development, agentic workflow design with human-in-the-loop guardrails, demos that win procurement, and the slow grind of getting an FI's people to actually adopt the thing. We take partners from vague idea → facilitated clarity → prototype → alpha → beta workflow → production AI application, using Mesh as the substrate underneath. Keith's framing on the call: "we can go from nothing to alpha better than anybody else."

The Connecticut bank-analysis startup that surfaced on the call is exactly the shape of client this is built for. Real product to build (an analysis platform a state controller wants every Connecticut bank reporting against), real banks already on the hook, regulators paying attention — and a team that doesn't have the application + AI muscle to ship the front half. Mesh gets them connected to cores and CRMs. Black Flag builds the product.

Trabian owns the stable financial-services infrastructure. Black Flag owns the AI application and adoption layer on top. The two halves combine into one complete delivery motion for any bank, credit union, or fintech client — without either side having to step into a business model that isn't theirs.

flowchart LR
    Client["Bank · CU · Fintech
(has an AI use case)"]
    BFD["BLACK FLAG
Application Studio
Facilitation · App build · Adoption"]
    Mesh["TRABIAN / MESH
Banking Substrate
Connectivity · Logging · Compliance"]
    Cores["Cores · Payment Rails
Q2 · Jack Henry · Fiserv · CSI · Alkami"]

    Client ==>|asks for AI app| BFD
    BFD ==>|ships app, owns adoption| Client
    BFD -.->|rides the substrate| Mesh
    Mesh -.->|governs connectivity| Cores

    classDef bfd fill:#1E5BB8,stroke:#4985d8,stroke-width:2px,color:#FAFAFA;
    classDef mesh fill:#1A1A1A,stroke:#4985d8,stroke-width:2px,color:#FAFAFA;
    classDef ext fill:#0A0A0A,stroke:#3A3A3A,stroke-width:1px,color:#A1A1AA;
    class BFD bfd;
    class Mesh mesh;
    class Client,Cores ext;
      

Four moves, in twelve months, walked Trabian from "custom dev shop" to "platform operator with a workflow-management roadmap." Each one widened the surface area where an application studio like Black Flag can plug in.

Mesh is already in production, already serving customers, already connecting cores. The remaining gap — the one a banker or a fintech CEO actually feels — is the application layer on top: the screens, the agents, the workflows, the human-in-the-loop copilots that make Mesh tangibly valuable to a buyer. Trabian doesn't want to staff a 30-person app-build org to fill that gap; it wants a partner that can. That's where Black Flag sits.

Frame the partnership as a shared services lane Trabian can sell, staff, and showcase. Black Flag operates as the embedded AI-application studio for Mesh, taking client ideas through a repeatable motion that respects Trabian's governance posture at every step.

Each successful build becomes a candidate repeatable module — sold across the 200+ relationships Trabian already has, demoed to prospects evaluating Mesh, and used as evidence the platform actually ships application-layer outcomes.

Not a buffet of seven. Three deeply-told stories, each pegged to a real customer already in the Mesh / Trabian / Q2 / Jack Henry orbit. Each is narrow, sandbox-safe, human-in-the-loop on day one, and answers a question those banks are already asking. Every one of them deliberately leans into Black Flag's two unfair advantages: we go from nothing to alpha faster than anybody else, and we manage agents that write the software — which is how a five-person team ships work other firms need thirty to ship.

The shape across all three: Trabian sells the platform, refers the application work, and sees revenue from both sides — without taking on the headcount or business model of an apps shop. Black Flag ships the apps, manages the agents that wrote them, keeps the FI's compliance posture intact, and turns each engagement into a reusable Mesh-backed module the next ten clients can buy off a shelf.

Trust gets built one stage at a time, and the timeline below in section 08 traces what that looks like for a single 90-day engagement. Across all three engagements in section 05, the same pattern applies: crawl in synthetic data, walk on a real alpha with a design partner, run as a productized Mesh-backed module. Each step earns the next.

Every custom AI application ships with an agent policy card: business purpose, user role, data sources, permitted/prohibited behaviors, allowed actions, actions requiring human approval, logging, failure modes, escalation path, retention model, success metrics. Early work lives at Levels 0–3.

The chart below is the real shape of one of these engagements — applied to any of the three stories in section 05. Defined, low-risk, fixed-fee. Synthetic data through day 60. Clear go/no-go on day 90 — design partner identified and beta scope agreed, or a clean exit. Crawl, walk, run mapped to the timeline rather than treated as an abstract ladder.

The work isn't just "the model" — it's the full delivery: web app + mobile companion, backend, the workflows and data contracts between Mesh and the client, the stakeholder management that gets the FI's people on the same page, and the agent governance that keeps compliance teams calm. The differentiator across all of it: BFD's team manages agents that write the software. That's how a five-person shop runs at the velocity of a thirty-person one — without sacrificing the compliance posture banks need.

gantt
    title AI APPLICATION DELIVERY · TRABIAN + BLACK FLAG (90-DAY ALPHA)
    dateFormat X
    axisFormat Day %s

    section Crawl · BFD-led
    NDA + Mesh orientation                :crawl1, 0, 10
    Use-case workshop + workflow map      :crawl2, 5, 18
    Agent policy card + risk register     :crawl3, 10, 24
    Clickable prototype + demo script     :crawl4, 14, 30

    section Walk · Joint
    Backend + data contracts (Mesh ↔ client) :walk1, 22, 50
    Web app + mobile companion build         :walk2, 28, 60
    Synthetic-data alpha                     :walk3, 35, 60
    Internal user feedback + iteration       :walk4, 45, 60

    section Run · Design partner
    Design partner identified (from 200+ FIs) :run1, 55, 68
    Integration + compliance gating           :run2, 60, 82
    Production data, governed scope           :run3, 70, 90
    Adoption playbook + reusable module spec  :run4, 75, 90
      

What the chart doesn't draw, but matters: every cell that has BFD building, has BFD agents building under BFD oversight. That's the speed multiplier that makes 90 days realistic on a custom AI app where most shops would quote nine months.

Everyone who's reading this — Aaron, Matt, Trey, Eli, Keith — already knows the thesis. The remaining question is whether there's a small real thing the five of us could build together in the next 30 days that would tell all five whether this works. The Connecticut bank-analysis referral is one obvious shape; an internal Mesh demo app is another. We're flexible on the wedge; we're firm on doing one.