How we start, ship, and wrap.

Step one — first call (30 minutes, no fee). You describe what you're trying to build. We listen, ask uncomfortable questions about scope and feasibility, and tell you honestly whether we think we can help — or whether the right answer is a different team, an off-the-shelf tool, or no project at all. About a third of our first calls end in "you don't need us."

Step two — NDA before any detail leaves the room. Mutual, straightforward, our template or yours.

Step three — problem framing (1-2 weeks, fixed fee). We dig into the actual problem with your team. By the end, you have a written framing: the problem, the constraints, the architectural choices, the metrics that mean success, and a credible scope and timeline. Roughly half of framings reveal that the project should be smaller or different from what was originally asked.

Step four — SOW. A written, signed Statement of Work with deliverables, milestones, fees, payment schedule, IP terms, and exit conditions. Fixed price where the scope is defined; time-and-materials where it can't honestly be fixed. You always know before signing whether you're getting one or the other and why.

• Senior engineers on every project. We don't put juniors on client work. The people who scope your project are the people who deliver it.
• Weekly written updates. Same day every week. Decisions made, blockers, what's next, hours burned versus budgeted. No surprises at the next milestone.
• Working demos at every milestone — not slides. If we can't show it running, the milestone isn't done.
• Code in your repository, written under your conventions, with the tests and CI that we'd want for our own work.
• Direct access to the people doing the work. There is no account-management layer. You get the engineers' Signal numbers if it matters.
• An exit plan from day one. If we vanish, your team can run the system. We document the assumptions, the failure modes, and the runbooks before we hand off.

• A single decision-maker. Not a steering committee. One person who can say yes or no within a business day.
• Access to the systems, data, and people we need to do the work. Where that's blocked by your own internal processes, we'll work with you to unblock — but the clock keeps running.
• Responsiveness to blocking questions, ideally within 24 hours. We assume you're busy; the SOW reflects that. Beyond a certain delay, we may have to pause and bill accordingly.
• Real feedback, not polite sign-off. If something is wrong, tell us early. We can fix things in week two; we cannot fix things in week ten.
• Payment on the schedule in the SOW. Romanian invoices, typically 14-day terms, EUR or RON.

• White-label dev shop work. Our name is on what we ship.
• Engagements where the right answer is an off-the-shelf product. If we think you should buy rather than build, we'll tell you on the first call and recommend specific products.
• Promise specific Claude model outputs. Models evolve; evaluations and guardrails are how we manage that. We commit to evaluation metrics and behaviour bounds in writing, not to specific completions.
• Retainer arrangements without a defined workstream. We don't bill for being on call. We bill for work shipped.
• Greenfield platform builds where Claude is incidental. We're an AI consultancy; we partner with full-stack teams for the rest.

Your data stays yours. We never train models on your data, never share it, and never use it for any purpose outside the engagement.

Code we write for you is yours under work-for-hire terms specified in each SOW. We retain rights to general tools, libraries, and methodologies we develop independently of the engagement and which we make available to other clients on the same basis.

Confidentiality is mutual and default. We never reference clients, project details, or content publicly without explicit written permission. The "anonymised case studies" on our homepage are published only with sign-off from the clients involved.

Our preference is fixed price when the scope is defined enough to commit. We carry the risk of execution; you carry the risk of scope change.

When the work involves genuine discovery (early-stage architecture, capability evaluation, novel agentic patterns), we charge time and materials against a not-to-exceed budget, with weekly burndown reporting.

Invoices are issued from RAN RESULTS SRL with full VAT details. Payment terms are stated in each SOW; default is 14 days from invoice date, in EUR or RON, by bank transfer. Late payments past 30 days may pause the engagement.

If you're unhappy, tell us immediately. We acknowledge within 24 hours and propose a path within 3 business days. We've never had a client formally invoke a dispute clause, and we'd rather keep that record.

If we miss a specification that's clearly defined in the SOW, we re-do or refund the affected portion of work, at your option. The mechanics are in each SOW.

Our total liability under any engagement is capped at the fees paid to us under that engagement. This is industry-standard; the specific clause is in each SOW.

Governing law for each engagement is Romanian law, with venue in the courts of Bucharest, unless the SOW says otherwise.

This is a description of how we operate, not a contract. Each engagement is governed by a separate, signed SOW that takes precedence over anything on this page. Where this page and an SOW disagree, the SOW wins.

If you're considering an engagement and any of the above is a deal-breaker, flag it on the first call. We'd rather know in week zero.

To start a conversation, send a brief through the contact form or write to contact@ranresults.com. We read every message ourselves and reply within a few business days.

RAN RESULTS SRL
CUI: RO32881995
Prel. Ghencea 65 B, Bl. H, Sc. 3, Et. 1, Ap. 4
Cod 061702, București, România