Methodology
The 90-Day Payback Methodology
How we deliver production-grade AI automation with 90-day payback periods. The framework we built after watching too many AI projects fail.
MIT's 2025 GenAI Divide report found that 95% of enterprise AI pilots deliver zero P&L impact. S&P Global separately reported 42% of companies scrapped most AI initiatives in 2025. The failure isn't the technology — current AI models are genuinely capable. The failure is the implementation model.
Most AI consulting engagements proceed in roughly this order: 1) build a pilot to demonstrate the technology, 2) try to scale the pilot, 3) discover the integration and production work is harder than the pilot suggested, 4) abandon. The pattern is so common it has a name — “pilot purgatory”.
Our methodology is designed to skip pilot purgatory entirely. Production-grade build from day one. Operational ROI as the project north star, not technology demonstration. Human-in-the-loop on consequential decisions. Compound value through ongoing optimisation rather than launch-then-decay. Six principles, three phases, one outcome: AI that works on day one and compounds value every month.
Six Principles
The six principles that drive every engagement
1. Diagnose before deploying
Every engagement starts with operational mapping — not technology evaluation. We sit with the people doing the actual work, watch how it flows, identify where the friction is, and project the specific ROI of the recommended automations before we build anything. The Diagnose phase is where most failed AI projects went wrong: they started by picking the technology and worked backwards. We do the opposite.
2. Production from day one
We don't build pilots. Pilots get abandoned. Every workflow we deploy is built to production quality from day one — proper error handling, retry logic, audit logging, monitoring, integration with your actual systems. The 'pilot then maybe productionise later' pattern is responsible for most of the 95% of AI projects that fail. We skip the pilot phase and ship production-grade work from week one.
3. Conservative ROI projections
We project ROI during Diagnose based on operational metrics — hours saved, cost recovered, revenue lift — using your specific cost structure, not industry averages. We project conservatively because over-promising is what destroys consultancy reputations. If our projection is wrong, it's wrong on the upside, not the downside. Our standing target: 90-day payback on first production automation against our cost.
4. Human-in-the-loop on consequential decisions
AI drafts, humans decide. Every workflow we build has explicit human approval gates on consequential decisions — customer communication, record modifications, payment triggers, anything financially material or relationship-sensitive. Fully autonomous AI on consequential workflows is a risk profile most businesses don't actually want once they understand it. Conservative routing protects the business and the client relationship.
5. No vendor lock-in by design
Every workflow we deploy is built so you own it. The automation logic, integrations, credentials, documentation — all yours, all transferable. If you ever want to stop working with us, your systems keep running and any competent technical contractor can take them on. We don't build platforms we control; we build systems you own. No proprietary licences to us. No revocation if you leave.
6. Drive the compound, not the launch
The biggest mistake we see in AI deployments is treating launch as the finish line. AI systems that work well at launch and aren't tuned afterwards degrade. AI capabilities improve every quarter; your business changes every quarter. Drive (ongoing optimisation) is where the compound value lives — typically 2–3x the original automation's value over 12 months. We design every engagement assuming Drive continues, even though there's no lock-in.
Three Phases
Diagnose. Deploy. Drive.
Diagnose
1–2 weeks • Operational mapping, ROI-projected implementation plan.
We sit with the principals and senior staff, map the operational workflows end-to-end, audit the existing systems stack, and identify the one or two automations with the strongest ROI case. The output is a written plan with projected hours saved per workflow, projected payback period, and a clear go / no-go recommendation. You walk away with enough to decide on next steps even if you don't proceed with us.
Deliverables
- Operational workflow mapping documented
- Systems and integration audit
- Ranked automation roadmap with ROI projections
- Compliance review for your industry
- Written go / no-go recommendation per recommended automation
Deploy
3–16 weeks • Production build, integration, training, go-live.
We build the recommended automation(s) to production quality, integrate with your existing systems, train your team, and deploy. Most engagements pick one workflow first (Accelerator tier) or 2–3 integrated workflows (Growth tier). We don't push three automations on day one — pick the workflow with the worst friction, ship it, then expand once the team has lived with the first one for 30 days.
Deliverables
- Production-quality automation deployed and running
- Full integration with your existing systems
- Team training and documentation
- Audit logging and monitoring
- 30-day post-deployment support and tuning
Drive
Ongoing monthly (optional) • Tuning, edge-case handling, new automation builds.
Post-implementation, most clients continue with a monthly retainer for ongoing optimisation. Drive is where the compound value lives — typically 2–3x the original automation's value over 12 months because AI capabilities improve, your business evolves, and new automation opportunities surface continuously. No lock-in. You own everything we built; if you stop Drive, your systems keep running.
Deliverables
- Monthly performance review and tuning
- Quarterly new automation builds
- Edge-case handling and rapid issue resolution
- Strategic input on emerging AI capabilities
- Compliance updates as regulatory framework evolves
The 90-Day Math
Why 90 days is the target
Most engagements pay back inside 90 days against the cost of the project. The math: a typical Accelerator engagement (AU$25–40k) that compresses 10–15 hours per week of senior or specialist time creates measurable cost recovery starting in week 4 (when the automation goes live) and compounds for the next 8–12 weeks until the cumulative recovered cost exceeds the engagement fee.
The 90-day target shapes what we propose. We don't recommend automations that don't pencil. We won't propose a project where the math requires 18 months of compound value to justify the cost. If the projected payback period exceeds 4–5 months for the first automation, we'll usually suggest a smaller scope or a different workflow.
The 90-day payback is the floor, not the ceiling. Multi-workflow engagements (Growth tier) compound across functions, often delivering 2–3x the projected first-year value once optimisation cycles add up. Transformation engagements pay back over 12–18 months but unlock organisation-wide capability that single-workflow wins can't. The methodology scales; 90-day payback is just the proof that the math works.
Want to see what 90-day payback looks like for your business?
Free 30-minute Diagnose call. We'll look at your operations, project a specific ROI case for the one or two automations with the strongest fit, and tell you upfront whether the math works.
Book a Diagnose call