Three ways to put us in the room.
We work with a small number of teams at a time. Most engagements start with The First Conversation; the shape that fits usually becomes obvious once we know what is actually wrong. Tell us if none of these fit and we will tell you who is.
The work is not the tool. The work is the change.
Most organizations are not failing at AI because the tools are bad. They are failing because they bought tools before they understood the problem, ran pilots that never became operations, and ended up with a collection of subscriptions that changed nothing about how the business actually runs.
AI transformation is an operational change. It means identifying where AI removes a real constraint, building against that constraint with architecture that holds, and making sure the people who live with the system can use it and maintain it without an engineer standing next to them. That is the work. The tool is just the last step.
The engagements on this page are built for that arc. The First Conversation finds what is actually wrong before anyone talks about building. Architecture makes sure the design holds before you commit to it. Build-with ships the piece that fixes it. You do not need all of them. You need the right one at the right moment.
Not sure what you actually need yet?
Most people arrive sure they need a tool, or sure they need “AI,” and most of the time they are solving the wrong problem. The First Conversation runs the other way around. You tell us what is slowing the business down. We find what is actually wrong before anyone talks about building anything.
It goes in two steps. First a short fit call, free, to make sure there is something here worth your time and ours. Then The First Conversation itself: a focused, paid working session where we diagnose the real issue and hand you a prescription. What to fix, in what order, and whether AI belongs anywhere near it.
Sometimes the answer is a workflow, not a model. Sometimes it is an integration you already pay for and never wired up. Sometimes it is do not spend money on this yet. You leave with a direction you can act on, whether or not you ever work with us again.
Best fit for operators who know something is off but cannot name it, and want a straight answer before paying for a build. This is the doorway to everything else on this page.
Get the design right before you build it.
Most AI projects do not fail in the build. They fail in the decisions made before it: the wrong retrieval approach, a data model that cannot grow, a vendor lock-in nobody priced, an agent design that demos well and falls over in production. By the time that surfaces in code, it is expensive to undo.
Architecture is a bounded engagement that settles those calls up front. We map the problem, design the system, choose the stack and the vendors with reasons you can defend, and hand you a written architecture your own engineers can build against. A fixed scope, a real deadline, and a deliverable you keep whether or not we build it with you.
We design against the same three Hexaxia frameworks we run internally. AGF defines what the systems are allowed to do. APF defines how they behave. ASIP defines how they change, keeping every update under human review so what you put in place does not drift out from under you. See the frameworks.
Best fit for teams about to commit real budget to an AI build who want the design pressure-tested first. Worst fit for teams who already know the design and just need hands to ship it. That is Build-with.
We sit with your engineers and ship.
Build-with is not a consulting report. It is a sprint your team runs with ours embedded in it. We open PRs, write the architecture doc, sit in the standup, and leave you with a system that works instead of a deck that describes one.
Every engagement runs against a defined outcome. A working retrieval pipeline. An agent that does the job. A vendor migration you actually finished. We set the scope with you, the deadline is real, and the deliverable is code in your repo with your name on the commit history.
We move fast because we run our own infrastructure. RhizomeRAG for memory. Hextant for the strategic context. The same primitives we ship as products are the ones we bring to your stack. You inherit our tooling without paying for the rebuild.
Best fit for teams with a real roadmap and a missing piece of senior AI experience. Worst fit for greenfield discovery where the problem is still being defined.
Second opinion when the stakes are real.
Some decisions deserve a focused outside read. An architecture you are about to commit to. A vendor contract you are about to sign. A team build you want to validate before you start hiring against it. The kind of question where a generalist answer is worse than no answer.
Advisory is point-of-decision work. We come in for a specific question, do the read, deliver a written judgment with the reasoning, and leave. No retainer. No ongoing relationship implied. A clean second opinion from operators who have shipped the thing you are about to ship.
Best fit for engineering leaders who need to make a defensible call and want an operator opinion to back it. Worst fit for teams that need someone embedded over time; we work in scoped engagements, not standing seats.
Same team, three shapes.
If you are not sure which one fits, start with a conversation. The shape often becomes obvious in the first thirty minutes.
Architecture
Build-with
Advisory
Time commitment
1 to 3 weeks, scoped
2 to 6 weeks, embedded
Single engagement
Output
Architecture and decisions
Working code in your repo
Written judgment
Continuity
Project-scoped
Project-scoped
Point-in-time
Who from us
Aaron, named
Aaron plus 1 to 2 builders
Aaron, or Jay where his read fits
When it fits
You need the design right first
You need senior AI delivery
You need a defensible call
Tell us what you are trying to do.
If we are the right fit, we will say so. If we are not, we will point you at someone who is. We work with a small number of teams on purpose.