Engagements

(Three ways to start)

Clear scope,agreed up front.

A fixed-fee sprint, a fixed-scope build or an ongoing partnership. Each is led by the same senior team, and the price is agreed before work starts.

Engagements

Three ways to start.

Fees are agreed before work starts and exclude VAT. Every engagement begins with a complimentary 45-minute planning session.

Strategy Sprint

2–3 weeks

Fixed

fee, agreed on scoping

For leadership teams that need to know where AI is worth it, and where it is not.

  • Stakeholder interviews and process review
  • Data and systems readiness check
  • Ranked opportunities with value and risk
  • Governance and policy starting point
  • Board-ready roadmap and business case
Discuss a sprint

Build & Launch

8–12 weeks

Fixed

scope and fee

One high-value AI system, designed, built and put into use with your team.

  • Use-case definition and success measures
  • Experience design and architecture
  • Build and integration with your systems
  • Evaluation, guardrails and human review
  • Launch, training and results report
Plan a build

AI Partner

Ongoing

Monthly

retainer

Senior AI leadership on call while you scale across teams.

  • Fractional head of AI
  • Portfolio and vendor oversight
  • New use cases delivered each quarter
  • Monitoring, evaluation and reporting
  • Executive briefings and team training
Discuss a partnership

Always included

Whichever way you start.

  • 01A senior consultant from first call to handover
  • 02Weekly working sessions, not big reveals
  • 03You own every file, prompt and line of code
  • 04A complimentary 45-minute planning session before any proposal
  • 05Price agreed before we start. No open-ended invoices

Questions

Before we talk.

The questions we hear most in a first call. If yours is not here, ask us.

advisory@newbranch.co.za

Yes. Some of the clearest returns come from smaller organisations automating a few well-understood processes. We scope the work to your size, and the first session will tell you quickly whether it is worth doing.

Whichever fit the job. We have no reseller arrangements, so we compare hosted models, open models you run yourself and the AI features already inside your software, then recommend on cost, quality and risk.

Every system ships with a test set built from your own cases, clear limits on what it may do, a person in the loop where it matters, and a record of what it did and why. You can audit it, and you can switch it off.

Under a mutual NDA, with least-privilege access, and in line with POPIA, GDPR and your own security policies. Your data is not used to train third-party models without your explicit approval.

You do. Code, prompts, evaluation sets, configuration and documentation are handed over, with enough training that your team can run and change it without us.

Yes, and it usually works better that way. We can lead the whole engagement or join your team for strategy, design or build.