Two engagements. One trajectory.

The Flagship Evaluation and the Yield Optimization Retainer are not alternatives — they are sequential. The 1-Week AI Potential Evaluation produces the audit, scorecard, and 90-day roadmap. The Yield Optimization Retainer is how that roadmap actually gets executed over the following 6–24 months. Most clients begin with the Evaluation and progress into the Retainer once the CFO-Ready Performance Report quantifies the savings on the table.

Package 01 · Discovery Session

1-Week AI Potential Evaluation

The forensic audit. A CFO-Ready Performance Report and the 90-day roadmap that becomes the spec for everything after.

  • 5 days on-site (or secure remote)
  • 6 CFO-ready deliverables
  • Fixed engagement fee
  • Capped at 12 companies / month
See the deliverables
Then Execute
Package 02 · Execution and Optimization

AI Yield Optimization Retainer

The operator. We execute the 90-day roadmap and run the optimised stack — compensation tied to verified P&L impact.

  • 6 to 24 months
  • Baseline retainer + success fee
  • Pay only on verified P&L impact
  • ~4 active engagements at a time
See the mechanics

From a 30-minute call to measurable yield.

Most clients enter at step two — the 1-Week AI Potential Evaluation — then escalate into the multi-quarter retainer once the CFO-Ready Report quantifies the savings on the table.

01
30 min · Free
Discovery Call
No-commitment conversation to scope where AI is or isn't generating return today.
02
1 week · On-site
AI Potential Evaluation
Forensic audit of your stack, ending with the CFO-Ready Performance Report and 6 deliverables.
Package 01
03
90 days · Sprint
Optimization Roadmap
Week-by-week execution against the prioritised use case backlog, with CFO-validated milestones.
Package 02
04
6–24 months
Execution and Optimization
Continuous operation of your AI economic engine — baseline retainer plus performance success fee.
Package 02

The 1-Week AI Potential
Evaluation.

For C-Suite and Tech Leadership requiring immediate clarity on their AI investments. We evaluate your existing systems — or mid-build projects — to pinpoint exactly where capital is leaking and where performance can be multiplied.

opportunity-map.pdf
↑ IMPACT
Strategic Bets
Quick Wins
Avoid
Fill-ins
Lead Scoring
Inbox Triage
Doc Summary
Pricing AI
Forecasting
AI Copywrite
Meeting Notes
AI Chatbot
EFFORT →
01

AI Opportunity Map

Every candidate use case in your business, plotted on a 2×2 of impact vs effort — so the leadership team can see what to ship in 30 days and what to defer.

cfo-ai-performance.xlsx
Q1 · AI Realised Value $4.2M
vs target +18%
OpEx Savings
$2.1M
Revenue Lift
$1.6M
Margin Expansion
$0.9M
Throughput Gains
$0.5M
Pipeline Velocity
$0.4M
Customer Retention
$0.3M
02

CFO-Ready AI Performance Reporting

A board-grade view of where the dollars are landing — OpEx saved, revenue uplift, throughput gains — tied to the line items in your P&L, not abstract "AI maturity" scores.

ai-scorecard.html
Strategy
8.4
Data Readiness
6.2
Talent
7.1
Governance
5.5
Adoption
7.8
03

Executive AI Scorecard

An honest read of your AI readiness across five axes — Strategy, Data, Talent, Governance, Adoption. Tells you where you actually stand and what to fix first.

90-day-roadmap.csv
W1W3W6W9W12
Workshop
Pilot build
Deploy v1
Scale & train
Board review
04

90-Day Optimization Roadmap

Week-by-week Gantt with named owners, success criteria and CFO-validated milestones. Built so PMO, product and finance can all read the same plan.

use-case-backlog.md
01 Sales lead scoring 94
02 Support deflection 88
03 RFP automation 76
04 Forecast accuracy 72
05 Onboarding agent 61
05

Prioritised Use Case Backlog

Ranked list of every shortlisted AI use case with a composite score (impact × feasibility × speed). The team leaves knowing exactly what's #1, what's #5, and why.

vendor-eval.xlsx
Cost Fit Speed Risk
OpenAI
Anthropic
Microsoft
NVIDIA
06

Vendor & Stack Evaluation

A side-by-side scorecard across the 6–10 vendors in your shortlist — cost, fit, speed-to-deploy and risk — so procurement and the CTO are aligned before the contract goes out.

How the six deliverables get produced.

The six artefacts above don't appear out of a slide deck — they're produced live, day by day, with your leadership in the room. Five days on-site (or secure remote), one senior architect embedded with your team, full CFO-ready report set delivered Friday afternoon.

Day 1
Forensic Baseline & Cost Telemetry

Live review of token spend, infrastructure utilisation, vendor contracts, and every running AI workflow — what it costs, what it returns, where it leaks. The numerical baseline for the rest of the week.

Day 2
Use Case Discovery & Impact Mapping

Leadership round-table to surface every candidate workflow, ranked live on the Impact × Effort matrix that becomes the AI Opportunity Map. Deep-dive interviews with each BU lead.

Day 3
Vendor & Stack Audit

Side-by-side evaluation of current vendor contracts vs optimised localised alternatives. Surfaces the bloated line items and the open-source substitutions worth piloting in the next 90 days.

Day 4
CFO Alignment & Backlog Prioritisation

Joint sessions with finance to translate AI workloads into P&L lines — OpEx vs CapEx, capitalisable engineering, revenue attribution. Every shortlisted use case scored on impact × feasibility × speed; the 5-axis AI Scorecard goes on the wall.

Day 5
90-Day Roadmap & CFO-Ready Report Delivery

Week-by-week execution plan with named owners and CFO-validated milestones. Full report set — Opportunity Map, Performance Report, Scorecard, Vendor Eval, Backlog and Roadmap — delivered Friday afternoon. Leadership team aligned and ready to start Monday.

From a bloated stack to an economic engine.

A side-by-side of what we typically replace versus what we typically install. Numbers are taken from real Carion engagements — see the projects page for full case studies.

Before · Bloated

What we typically rip out

  • Metered proprietary LLM APIs on high-volume workloads $129K / mo
  • Over-provisioned cloud GPUs running at < 20% utilisation Idle spend
  • Black-box agentic vendors with no token-level telemetry No P&L attribution
  • Generic enrichment pipelines uncoupled from sales motion 0.8% conv.
  • Compliance/data-residency liabilities on outsourced inference Vendor lock-in
After · Optimised

What we typically install

  • Localised open-source inference for high-volume text/agentic flows < $15K / mo
  • Right-sized compute with dynamic scaling + LLM cost routing > 70% util.
  • CFO-attributable cost lines tied to every workflow Daily P&L
  • Agentic enrichment rebuilt around your ICP and sales playbook ~2.0% conv.
  • Sovereign architecture under your control — data, models, ops Vendor freedom

How the success fee actually works.

The retainer covers our dedicated senior team. The success fee only triggers on financial outcomes verified by your CFO. If the optimization doesn't move your numbers, the success portion isn't paid — full stop.

Step 01 · Always Paid

Baseline Operational Retainer

Secures the dedicated senior team — model routing, infrastructure operations, agentic tuning, monthly executive review. Sized to your stack, paid quarterly.

Step 02 · Quarterly

Joint CFO P&L Verification

Every quarter we sit down with your CFO and reconcile realised OpEx reduction and revenue lift against the baseline locked at engagement start. Only verified, attributable impact counts.

Step 03 · Only If Results

Performance Success Fee

A pre-agreed share of the verified P&L impact — never on theoretical projections, only on dollars that hit your books. No results = no success fee. We share the risk because we're confident in the yield.

What real engagements have returned.

Realistic ranges drawn from recent Carion engagements across financial services, energy, and B2B SaaS. The yield curve below traces the typical trajectory from negative ROI at engagement start to stabilised yield by Month 6. Full case studies on the Projects page.

−40 to 80%
Monthly TCO reduction
High-volume AI workloads (lead-gen, support, back-office)
1.5× to 2.5×
Conversion-rate uplift
B2B agentic enrichment and qualification pipelines
15–25%
Stabilised Operational ROI
From a negative starting point · plotted on the curve below
3–6 months
Median time to break-even
From Evaluation delivery to positive P&L contribution
+30% +15% 0% break-even −20% −40% M 0 M 1 M 2 M 3 M 4 M 5 Stabilised Negative ROI $129K / mo OpEx ~Month 3 Break-even Stabilised yield +15 to 25% Operational ROI
Engagement start · negative ROI Break-even crossover Stabilised yield

Which package fits your stage?

The Immersion is built for organisations that need clarity and a plan; the Retainer is built for organisations ready to operationalise and scale.

1-Week AI Potential Evaluation AI Yield Optimization Track
Best for Leadership teams needing a forensic audit, CFO-ready figures, and an executable 90-day plan. Enterprises ready to execute that plan and scale the optimised architecture under continuous senior operation.
Duration Five days on-site (or secure remote). Full report set delivered Friday afternoon; 90-day roadmap by end of the following week. 6 to 24 months. Tightly scoped quarterly milestones with named CFO-validated KPIs.
Who's in the room Senior Carion architect plus your C-suite + BU lead. Dedicated senior pod (architect + ops engineer + economist) embedded with your team.
Headline deliverables AI Opportunity Map · CFO-Ready Report · Executive Scorecard · Vendor & Stack Eval · Use Case Backlog · 90-Day Roadmap. Optimised AI infrastructure in production · CFO dashboard with monthly P&L attribution · Continuous model-routing & agentic tuning.
Fee structure Fixed engagement fee. Baseline operational retainer plus a Performance Success Fee tied to the OpEx savings and revenue lift you actually realise.
Capacity 12 companies per month. Limited senior-pod availability — typically 4 active engagements at any time.
Outcome focus Clarity, prioritisation, and a CFO-defensible roadmap. Operationalised yield — measured monthly in OpEx reduction and revenue lift.

Not sure which package fits?

Book a 30-minute consultation. We'll listen to where your AI investment stands today and recommend the right engagement — or honestly tell you it's not time yet.