Question 1

How does a first engagement with Arktika begin?

Accepted Answer

It starts with a 30-minute scoping call with Joshua - not a sales rep, not an account manager. We map your workflows, data, and the outcome you're chasing, then recommend the right entry point. Most engagements begin with the AI Readiness Assessment, which delivers a ranked opportunity map in two weeks.

Question 2

What does the AI Readiness Assessment actually cover?

Accepted Answer

Four layers: data, tooling, team, and workflow. We audit where your data lives and how usable it is, what's already in your stack, which teams are ready to adopt, and which workflows carry the most recoverable cost. You leave with a ranked opportunity map, scored feasibility per use case, an architecture recommendation, and a phased delivery plan. Two weeks, $12,500 flat, credited against a Growth or Scale Transformation.

Question 3

Do we need an AI strategy before we start?

Accepted Answer

No - that's usually what the Assessment produces. Most teams arrive with pressure to 'do something with AI' and a pile of half-finished pilots. We start from the workflows and the numbers, not from a model choice. Strategy comes out of the readiness work, not before it.

Question 4

What do you need from us to get started?

Accepted Answer

Less than you'd think. A 30-minute conversation covers scoping. During the Assessment we need read access to the relevant systems, a handful of stakeholder interviews, and one internal owner who can unblock access. No RFP process, no six-week procurement cycle.

Question 5

What's the difference between Growth and Scale Transformation?

Accepted Answer

Scope and parallelism. Growth ($95,000) delivers one production workstream end to end - a single high-value use case built, integrated, adopted, and measured. Scale ($150,000) runs multiple workstreams in parallel with a larger squad, deeper platform work, and an internal enablement track so your team can extend the system after handover. If you have one obvious use case, take Growth. If AI is a company-wide programme, take Scale.

Question 6

Is the Assessment fee credited if we continue?

Accepted Answer

Yes - in full. The $12,500 Assessment is credited against a Growth or Scale Transformation started within 90 days of delivery. If you don't continue, you keep the opportunity map, feasibility scoring, and architecture recommendation with no further obligation.

Question 7

What happens after a transformation ships?

Accepted Answer

Phase 05: Optimisation. Production AI systems drift - models degrade, usage patterns shift, costs creep. Our Essential ($8,500/mo) and Partner ($15,000/mo) retainers cover monitoring, evaluation, prompt and model tuning, cost optimisation, and continued build capacity. Retainers are optional; plenty of clients take handover and run it themselves.

Question 8

How long does delivery take?

Accepted Answer

Assessment is two weeks. Growth typically reaches production in 8-12 weeks, Scale in 12-20 depending on workstream count and integration depth. Phase gates are contractual: nothing moves forward without a signed outcome at the previous gate, so timelines stay honest.

Question 9

Who is actually on the squad?

Accepted Answer

A senior, timezone-aligned delivery team. A typical Growth squad is an AI Strategist, an ML Engineer, a Data Architect, a Full-Stack Engineer, and an Adoption Lead, with Joshua as strategic lead. Scale squads add engineering capacity and a platform specialist. You meet the people doing the work before the engagement starts - no bait-and-switch between pitch and delivery.

Question 10

How is this different from staff augmentation?

Accepted Answer

We're accountable for the outcome, not the hours. Staff aug gives you bodies and leaves the architecture, adoption, and measurement to you. We take a scoped outcome, run it through phase gates, and hand over a working system your team uses. If nobody adopts it, we haven't delivered - that's why Adopt is its own phase.

Question 11

How is this different from a big consultancy?

Accepted Answer

No 200-page deck, no partner-then-juniors switch, no seven-figure minimum. Nearshore delivery keeps senior squads in your working hours at roughly 40-60% of onshore consultancy cost, and every phase ends in something running - not a recommendation to buy more consulting.

Question 12

Why nearshore delivery?

Accepted Answer

Overlap hours are the whole point. Squads across the Americas work North American business hours, so architecture decisions happen live instead of over a 12-hour async lag. The cost advantage is real, but the reason clients stay is that delivery velocity looks like an internal team's.

Question 13

What's the difference between the Arktika AI Engine and Applied AI?

Accepted Answer

The AI Engine is our internal platform; Applied AI is what we build for you. The Engine is the evaluation, orchestration, and delivery tooling our squads run on - accelerators for retrieval pipelines, agent scaffolding, evaluation harnesses, and monitoring that let a squad reach production faster. Applied AI is the client-facing practice: readiness, roadmap, build, adoption, and optimisation delivered against your workflows. You engage Applied AI; the Engine is how we deliver it faster.

Question 14

What tech stack do you build with?

Accepted Answer

Whatever fits your stack, not a single opinionated framework. Most engagements combine Anthropic Claude or OpenAI for model infrastructure, LangGraph or LangChain for agent orchestration, Pinecone or pgvector for retrieval, dbt and warehouse-native tooling for data, and AWS, GCP, or Azure for cloud. Where an open-weight or self-hosted model fits the compliance profile better, we build that instead. Recommendations come out of the Assessment, based on what you already run.

Question 15

Who owns the IP of what you build?

Accepted Answer

You do. Fully. Every line of code, every model artefact, every pipeline and automation built during an Arktika engagement transfers to your company at handover under a full IP transfer agreement signed before work begins. We keep nothing proprietary in your production path.

Question 16

How do you handle NDAs and data security?

Accepted Answer

An NDA is standard and signed before the Assessment begins. Your data is never used to train models, never shared across clients, and never leaves the environments you approve. We work inside your cloud accounts wherever possible, follow least-privilege access, and can align to SOC 2, HIPAA, or regional data-residency requirements when your compliance profile requires it.

Question 17

What if our data isn't ready for AI?

Accepted Answer

That's a finding, not a failure - and it's common. Roughly half of assessments surface data readiness as the first constraint. When that happens the roadmap sequences data foundation work ahead of model work, so you build on something stable instead of shipping a demo that breaks in production.

Question 18

How do you measure whether it worked?

Accepted Answer

Against baselines set in Phase 01. Every use case gets a measurable target before we build - cycle time, cost per transaction, deflection rate, throughput - and we instrument the system to report against it. Clients typically see 40-68% cost reduction on the workflows we target, measured against their own pre-engagement numbers, not an industry benchmark.

Question 19

What if adoption stalls internally?

Accepted Answer

Adoption is a delivery phase, not an afterthought. Phase 04 covers enablement sessions, documentation, champions inside each affected team, and usage instrumentation so we can see where uptake stalls and fix it. A system nobody uses is a failed engagement regardless of how good the model is.

Question 20

What if the Assessment says AI isn't the answer?

Accepted Answer

Then we say so. Some workflows need process redesign or plain automation, not a model - and telling you that in week two is cheaper than discovering it in month six. The opportunity map ranks by expected value and feasibility, which sometimes means the honest recommendation is to fix the data or the process first.

Question 21

What does "founder-led" actually mean in practice?

Accepted Answer

Joshua O'Grady is the strategic lead on your engagement. He runs the scoping call, leads the readiness work, chairs the phase gate reviews, and is reachable directly - not through a ticketing system. When you work with Arktika, you work with the person accountable for the outcome.

Question 22

Can we reach Joshua directly during our engagement?

Accepted Answer

Yes. Every client has a direct line - WhatsApp, email, LinkedIn. No support queue, no "your account manager will be in touch." If something comes up mid-phase, you reach the person responsible for it.

Question 23

How many engagements do you run at once?

Accepted Answer

Deliberately few. Arktika takes on a small number of transformation engagements per quarter so squads stay senior and the founder-led model stays real. If you're planning a timeline, it's worth a quick call to check current availability.

Questions, answered.

Getting Started

Programs & Scope

Squads & Delivery Model

Technology, IP & Security

Outcomes & Measurement

Working with Joshua

Talk to a founder, not a sales rep.