Skip to main content

Exploring a small number of new product partnerships.

Build the AI your field is missing.

Amphibian partners with senior practitioners to build AI for specialty workflows broad tools get wrong.

You bring the workflow expertise, buyer trust, and distribution. We bring the product engineering, AI infrastructure, and operating model. We share the upside.

The first partnership

It started with a practitioner problem.

See the Spherient workflow

Brick Brickley, co-founder of Spherient Advisors and managing partner of Benefit Fiduciary Group, saw the same pattern across employer-advisory work: broad AI tools were impressive in demos and unreliable inside the actual workflow.

The issue was not that the models were useless. It was that the work required domain-specific documents, plan-design judgment, carrier context, source tracing, and output formats advisors already trusted.

So Brick partnered with Amphibian. He brought the workflow expertise, practitioner feedback, customer context, and distribution path. We brought the engineering team and the filesystem-agent infrastructure. Together, we are turning a specific advisory workflow into a product with Spherient.

First partner

Spherient Advisors

Workflow

Carrier materials to reviewed workbook

The deal

What the partnership actually means.

Real upside.

This is not a referral arrangement or an hourly consulting project. Partnership economics can include revenue share, licensing upside, equity, or a combination, depending on the market, product, and role. The goal is simple: if the partnership creates durable value, the partner outcome should matter.

Real involvement.

During discovery and build, expect a few hours a week: workflow sessions, review of product decisions, introductions to trusted design partners, and candid feedback on what the system gets wrong. After launch, involvement becomes lighter but remains strategic.

Real decision points.

Most conversations will not become partnerships. That is intentional. We spend several weeks learning the workflow before either side commits. If the market is too small, the workflow is too diffuse, the distribution is weak, or the trust is not there, we stop early.

Process

How it usually unfolds.

Some partnerships move faster and some slower. A typical path from first call to pilot is measured in months, not weeks.

  1. 1

    First conversation

    A focused call about the workflow, market, and why broad AI fails here.

  2. 2

    Workflow deep dive

    We review documents, edge cases, existing processes, and a small number of trusted practitioner or customer perspectives.

  3. 3

    Partnership decision

    If there is mutual conviction, we agree scope, economics, responsibilities, and any exclusivity in writing.

  4. 4

    Pilot build

    We build around one narrow workflow with a design-partner customer close enough to give real feedback.

  5. 5

    Market launch

    The product enters the partner’s network first, then expands from there if the workflow proves durable.

Partner inquiry

Tell us about the workflow.

The best inquiries are specific. Tell us the work, the documents, the judgment calls, where broad AI breaks, and who would trust you if a product existed.

We can build technology and ship product. What we cannot do from the outside is know your specialty as well as you do.

If that sounds like the right kind of partnership, send a note.

— Jack & Ali

Aim for the specific work a trusted practitioner would recognize, not the broad industry category.

We respond to specific, considered inquiries within a week.