rootwise — AI implementation studio

01 · Workflows

What we actually build.

Real workflows, not “AI transformation” theatre.

We build custom AI systems around the workflows that already slow your company down: documents, emails, internal knowledge, approvals, quotes, invoices, support and operations.

Delivery / logistics

Pain: Dispatchers manually triage emails, proof-of-delivery, invoices and exceptions across TMS / ERP.

Solution: AI triage parses documents and emails, creates tasks, updates shipment records and escalates exceptions for human approval.

Accounting / bookkeeping

Pain: Teams manually reconcile invoices, purchase orders, bank references and email approvals.

Solution: AP copilot extracts data, matches it to PO / contract, flags mismatches and drafts entries for Xero / Sage / QuickBooks / ERP.

Retail / e-commerce

Pain: Support drowns in questions about orders, returns, refunds and delivery status.

Solution: Support copilot reads order history, policy and customer thread, then drafts the response or action. Risky refunds go to human approval.

Property management

Pain: Tenant emails, photos, lease clauses and contractor updates pile up into chaos.

Solution: Maintenance triage turns emails and photos into work orders, checks lease / rules, prioritizes urgent cases and updates the tenant thread.

Wholesale / distribution

Pain: Sales teams turn PDF / email RFQs into quotes manually, checking SKU, stock, margin and delivery terms.

Solution: Quote assistant extracts SKUs and quantities, maps them to ERP items, flags missing stock or margin risk and drafts a quote for approval.

02 · Best fit

Who this is for.

We are a good fit when there is a real workflow, real data and someone inside the company who owns the problem.

Best fit

20–250 employees
Manual document, email or ops workflows
No internal AI team
Clear workflow owner inside the company
Budget for production, not just a demo
Real data we can work with

Not a fit

Cheap chatbot or generic AI integration
Vague “AI transformation” without a concrete workflow
Staff augmentation / hourly dev
High-risk AI: hiring, credit, medical, biometric, law enforcement
No internal owner of the workflow
No access to data or SMEs

03 · Where to start

Start with the smallest proof that can kill or justify the build.

Most AI projects shouldn’t begin as a €80k build. We map the workflow, test the risky part, and only then commit to production.

01 — First touch Free

We understand the task, flag obvious risks and give a rough path.

02 — AI Process Audit €2.5–5k · 3–5 days

We map the workflow, data, tools, blockers and first implementation path.

03 — AI Feasibility Sprint €7.5k founding · €12–15k list · 10 working days

We test the risky part on real data before you fund the build.

04 — Production AI System v1 From €25k · 4–10 weeks

We turn the validated workflow into production software your team can use.

05 — AI Care Plan €1.5–12k/mo · monthly

Monitoring, fixes, model migrations and improvement sprints.

04 · First signal

First signals — what could break before we build.

First touch free

You walk us through the task. We flag the tricky parts, what to watch out for, sketch a rough breakdown and give a ballpark on timeline and cost.

Aligning expectations
Ballpark cost and timeline
A proper hello

AI Process Audit €2.5–5k · 3–5 days

We map your workflow, data, tools and operational constraints. Then we decide where AI actually makes sense, which workflows go first and which risks need testing.

Risk line: Model quality, cost per user, latency, scale — anything that could break the project goes on the table early.

workflow map
risk map
data readiness
first implementation path

05 · Flagship

AI Feasibility Sprint.

Test the risky part on real data before you fund the build.

10 working days · €7.5k founding · €12–15k list

You walk away with

prototype
evals
model / cost comparison
architecture
risk map
go / no-go recommendation
fixed quote for production

Guarantees

Sprint fee credited toward production if we move to v1 within 60 days.
No-build honesty guarantee: if the workflow should not be built, we say so.
60-day critical bug warranty for founding clients.
1 month Essential Care included after production launch.

Founding offer

We are new to this market, not new to production software. For our first 3–5 EU/US clients, we offer a limited founding rate in exchange for the chance to earn our first public proof in EU/US.

Trust exchange: We ask for one of: a public case study, a named reference or a private reference call after delivery.

06 · Production

Once the risky part is validated, we build.

Production AI System v1 · from €25k · 4–10 weeks

Setting the course

We write the spec and user scenarios. In parallel, we work out architecture, infrastructure and design.

Writing documentation
Planning the build and checkpoints
Covering infra and services

Build phase

We go heads-down on the build per the agreed plan. You get status reports as tasks close.

// dev mode
team.focus("product")
project.online === true

Checkpoint demos

At agreed points in the plan, you see a demo of the work in progress. You validate we’re heading where you wanted to land.

Demo at agreed checkpoints
Validation at every stage
Approving course corrections
Status reports as tasks close

07 · Handover

Product ships. The keys are yours.

Everything we built along the way is yours. From here it’s your call — stay with us on a care plan, or run it on your own.

Full ownership

Your code.
Your accounts.
Your infrastructure.
Your documentation.
Your decision.

What we hand over

repositories
accounts and access
infra and deployment
logging and monitoring
runbooks and onboarding
known risks
handover checklist

AI Care Plan €1.5–12k/mo · monthly

Monitoring, fixes, model migrations and improvement sprints. Essential / Growth / Scale tiers.

Or run it on your own

Open codebase, no hidden dependencies, runbooks, onboarding docs and handover sessions for your next team.

∞ · Referral

Bring us a client.

Know a company drowning in documents, emails, internal knowledge or manual operations? Introduce us to the decision-maker. If they become a client, we pay a transparent referral fee after the first invoice is paid.

5% of the first paid project invoice · capped at €5,000

What counts: A warm introduction to a decision-maker at a company that is not already in our pipeline.
What doesn’t: Cold lists, spam outreach, intros to people with no buying role or referrals where the fee is not disclosed.
Important: We do not pay employees, officers, procurement staff or decision-makers of a referred company unless their employer has approved it in writing.
Terms: Valid for 6 months from intro · Payout after invoice is paid · EU-compliant invoicing required.

Refer a client →

AI that survivesfirst contactwith real work.

What we actually build.

Who this is for.

Best fit

Not a fit

Start with the smallest proof that can kill or justify the build.

First signals — what could break before we build.

AI Feasibility Sprint.

Once the risky part is validated, we build.

Product ships. The keys are yours.

Bring us a client.

No black box. No lock-in. The product is yours.

Every account in your name.

Data, properly shelved.

Repositories.

Why this model.

Servers and recovery.

Logging

On-premise & Fine-tuning

Incidents.

Ship updates without fear

Zero .env files.

GDPR bundle.

Mobile apps.

An easy way in.

Modern UI standards. Intuitive by default.

Domain and certificates.

Payments, auth, and third-party services.

Thank you for your trust.

Next steps

Sota report

Support.

Growth.

On your own.