The Company
attacca /aˈtakːa/
A Vancouver AI consulting company, built around intent.
They fail because they adopt without intent, deploy without specs, and scale without evaluation.
We help mid-market companies stay relevant through the AI revolution.
Why we exist
AI is compounding faster than most companies can adapt.
The winners of this cycle will not be the companies with the most AI tools. They will be the ones with the clearest intent — the ones who know what they are optimizing for, where humans still hold judgment, and which bets are worth making before the technology obsoletes them.
90% reliability = 6 failures a day across a 10-step workflow.
Without specifications, evaluation, and trust tiers, AI adoption becomes liability — not leverage. Teams ship agents that are confidently wrong. Leadership loses the ability to distinguish a real bet from a demo. Hyperscalers absorb the middle layer while you are still debating tools.
Our mission: give mid-market companies the decision infrastructure to navigate this — without getting replaced.
How we help
A four-step navigator, not a 40-page report.
Grounded in Dark Factory methodology, open-sourced as Attacca Forge.
- 01
Diagnose
We map where AI actually fits using the 7 difficulty axes, intent audit, and trust tier classification. Most AI projects fail at this step — before a single line of code.
- 02
Map
We score every AI bet on 5 durable verticals and 3 strategic axes. Output: a decision portfolio telling you what to build, sharpen, reposition, or kill.
- 03
Spec
Your top 3 bets become execution-ready specifications — behavioral contracts, intent contracts, eval scenarios, factorial stress tests. The same methodology we use on Tier 4 safety-critical systems.
- 04
Hand off
You leave with a 90-day roadmap and specs your engineering team (or ours) can execute. No ongoing dependency. No black boxes.
Who we are
UNIFY ORCHESTRATION AI INC.
A British Columbia corporation (BC1586506), headquartered in Vancouver. Operating company behind the Attacca product family — the Navigator consultation, the Forge open-source methodology, and the book Specs In, Software Out.
Lina Duque is the founder and general manager of UNIFY ORCHESTRATION AI INC. Jhonn Moreno is the author of Specs In, Software Out and creator of the Attacca Forge methodology.
We are a small, deliberately concentrated team. We do not scale through headcount — we scale through methodology. Our engagements are capped so every client works with the people who built the frameworks.
Our practice is rooted in production systems for regulated industries: $42M LNG operations, pharmacy call center infrastructure (Tier 4 patient safety), property management compliance under Colombian Ley 675/2001, and marketing intelligence platforms in production.
What we believe
Five pillars.
Ensemble over soloist
Single agents crack under real workflows. Coordinated systems endure.
Momentum is sacred
Every handoff preserves context. Lost intent is the real failure mode.
The conductor empowers, never replaces
AI orchestration amplifies your team — it does not eliminate judgment.
Trust is earned
Autonomy scales with evidence. Start constrained. Graduate on performance.
The score is yours
Your workflows. Your data. Your decisions. No vendor lock-in.
Proof
Production systems we have shipped.
Regasificadora del Pacífico
$42M LNG operations. Tier 4 safety-critical. 227+ sources, 5 operational manuals for Ecopetrol.
Ecomm Pharmacy Call Center
Knowledge Operating System across CSR, QC, and Loyalty teams. Prescription medication referral — Tier 4.
Edifica
Information system for Colombian propiedad horizontal (Ley 675/2001). Live with first building administrators.
Vision Labs
Marketing intelligence platform. Shipped to agency client, in production.
The methodology behind every engagement is open-sourced as Attacca Forge and documented in the book Specs In, Software Out.
Standing on
The work that shaped this.
No methodology is built in isolation. These are the people, research, and communities whose work directly influenced how Attacca thinks about reliability.
Nate Jones
AI evaluation research & failure mode taxonomyHis analysis of how AI agents fail under real-world conditions — the inverted U, reasoning-action gaps, context hijacking, and guardrail misfires — shaped the adversarial evaluation layer of this methodology.
YouTubeMount Sinai Health System
Factorial stress testing designTheir factorial design study on clinical AI agents (Nature Medicine, 2026) provided the empirical foundation for structured variation testing — the idea that you must test under controlled perturbations, not just clean scenarios.
Stripe
Engineering rigor at scaleStripe’s engineering blog — particularly their writing on building reliable systems and making infrastructure invisible — influenced how we think about methodology as developer tooling, not process overhead.
Stripe Engineering BlogThe Harness Engineer Community
Practitioner knowledge & field testingThe practitioners building and testing AI agent harnesses in production. Their real-world feedback on what breaks, what scales, and what the documentation never tells you is woven throughout this framework.
Ready to navigate this deliberately?
Start with a 10-minute readiness audit.
Free. No call required. You will receive a personalized report with your AI readiness score, your archetype, and the three highest-leverage gaps to close first.
Take the AI Readiness Audit