Kristy Stengl
Founder, True North NexTech
Twenty years at the intersection of data, cloud, and AI — most of it spent translating between the people who set strategy and the people who build the systems.
Why this practice exists
I kept seeing the same failure. An organization would run a promising AI pilot, get a good result, and then stall — for eighteen months, sometimes indefinitely. The post-mortems usually blamed the model, or the vendor, or the pace of the technology.
It was almost never any of those. It was that nobody could trace where the data came from, nobody had measured its quality, and nobody had decided who was accountable when the system was wrong. The pilot succeeded precisely because someone had hand-picked the inputs — which is exactly what production does not do for you.
True North NexTech exists to work on that layer: the data foundations and governance underneath AI, and the engineering practices that keep both honest as a system grows.
How I work
I am a practitioner, not a slide deck. I can sit in a leadership meeting in the morning and review a schema migration in the afternoon, and I think that combination is the whole point — governance written by people who have never shipped a model tends to produce frameworks nobody in engineering can act on.
Engagements are scoped so you can buy a decision without buying a relationship. If the honest answer is that you should not proceed, I will say so and the engagement ends there.
Background
Two decades building and governing data platforms, most recently in healthcare provider data at national scale. That has meant enriching a 7.4-million-provider directory against federal and state registries with per-field provenance and confidence scoring; building the governance tooling that let business analysts own a data model directly instead of through spreadsheets; and standing up documentation that regenerates itself daily from live sources so it cannot drift.
Alongside that, a parallel body of work in higher education — rebuilding technical curricula as engineering systems, and building a production platform that gives instructors rubric-aligned draft grading without ever taking the academic judgment out of their hands.
The common thread across all of it is making trust legible. Provenance that travels with the value. Documentation that cannot go stale. Confidence scores that surface disagreement rather than hiding it. Systems that show their evidence.
Read the case studies