120 days. One business unit.
Real freight. Measurable outcomes.
This document is the scope for a Crane Worldwide Logistics pilot of WakeTech.ai. It is intentionally narrow, intentionally measurable, and intentionally bounded by a clean exit if the outcomes are not there. Commercial terms are handled in the conversation that follows.
One business unit. One deployment. Real loads from day one.
Selected jointly. Highest-pain, highest-leverage unit.
Standup to outcome review. No drag-on, no scope creep.
No per-seat surprises. Infrastructure cost is scoped up front.
No rip and replace. No fight with your existing vendor. No multi-year migration. We pick one Crane business unit where the pain is highest, stand up a dedicated WakeTech.ai deployment in a Crane-controlled environment, and run real freight through it for 120 days. Everything we deploy is software Crane already saw in the overview: WakeTech.ai core, the AI crew, WakeEDI, WakeMail, the self-hosted routing engine, plus the customization layer built for the selected business unit.
Four phases. Each one ends with a checkpoint.
- Crane and WakeTech.ai jointly select the pilot business unit and pilot scope.
- Dedicated deployment provisioned. Crane's choice: WakeTech.ai-managed Azure subscription or landing inside Crane's own Azure tenant.
- Pilot business unit's data shape mapped: org chart, customers, carriers, lanes, tariffs, EDI partners, legacy integrations in scope.
- Success criteria signed off in writing. We do not start running freight until the scoreboard is agreed.
- Customers, carriers, locations, tariffs migrated or proxied from existing systems.
- Customization layer built for the business unit: hierarchy, customs workflows, integrations to legacy ERP and any mandatory third-party systems.
- AI crew tuned to the business unit's voice, tariffs, and customer relationships. Approval thresholds set per agent, per workflow.
- Shadow mode: agents observe live work, propose actions, but humans execute. We measure agreement.
- Live freight on the platform. Every load handled by the AI crew with a human approval on every material decision until the agent earns autonomous authority through its agreement rate.
- Weekly outcome review against the scoreboard. Anything trending wrong gets fixed inside the week.
- Customization layer iterates against real friction. The pieces that make Crane Crane get built in the open, on real work.
- Final outcome review. Side-by-side numbers: before vs. during, in Crane's own P&L language.
- Three roads. Expand to additional business units on terms agreed in Phase 1. Continue at pilot scope. Walk away clean — Crane keeps its data, the customization layer, and the documented case for what the pilot did.
Success criteria, agreed in writing before we start.
We do not ship a pilot without a scoreboard. The exact thresholds are set in Phase 1 with Crane's pilot business unit lead. The five metrics below are the standing template. Crane adjusts the targets, we agree, and we measure publicly every week.
Median time from inbound quote request to outbound quote response.
Median minutes, per channel, per customer cohort. Target set in Phase 1.
Fully-loaded cost per load to operate the pilot business unit.
Carrier spend, labor allocation, and platform cost combined. Compared against the pre-pilot baseline for the same business unit.
Loads handled per dispatcher per day.
Same dispatchers, before and during. Throughput per head, no headcount change required to demonstrate it.
Routine status checks delivered automatically; exception touches answered faster.
Reduction in inbound 'where is my load' email volume. Median exception resolution time.
Days from delivery to invoice sent, and days from invoice to cash collected.
Compared against the pre-pilot DSO baseline for the same business unit.
How often the AI crew's proposed action is the same as the human's chosen action.
Per agent, per workflow. The threshold an agent must hit before it earns autonomous authority on that workflow.
Two teams. One pilot.
- Dedicated deployment, deployed, monitored, and operated for the pilot window.
- Core platform license and updates throughout the pilot.
- Customization layer built to the pilot business unit's data shape, org chart, and workflows.
- AI crew tuning, agent approval-threshold management, skill recording for new workflows surfaced during the pilot.
- Integration work to any in-scope legacy system (ERP, customs, EDI partner, third-party tools).
- Weekly written outcome report. Founder on the line for the entirety of the pilot.
- The selected business unit and an executive sponsor empowered to make decisions inside the pilot window.
- A pilot business unit lead who owns the scoreboard and unblocks operational decisions.
- Access to the existing data needed to seed the deployment (customers, carriers, locations, tariffs, recent freight history).
- Access to in-scope legacy systems for the integration work.
- Operator hours from the pilot business unit during onboarding and shadow mode.
- A clear yes-or-no decision at Day 120.
Human approval on every decision until the agent earns its way out.
Every AI agent in the WakeTech.ai crew operates under a four-layer safety model: skill scope, tool-level allowlist, approval threshold per workflow, and an audit trail every action writes to. During the pilot, every material agent decision requires a human click until that agent reaches the agreement rate Crane sets in Phase 1. Autonomy is earned, not granted.
Each agent has a defined catalog of skills. New skills require explicit assignment. Agents cannot self-promote into new workflows.
Per agent, per workflow. Set by Crane's pilot lead. Until the agent reaches the threshold, every action waits for a human.
Every action, every approval, every decision logged with actor, timestamp, and reasoning. Reviewable in Crane's deployment in real time.
Crane keeps everything Crane brought in. And what we built for Crane.
Pick the shape that fits your P&L.
We will quote against your actual numbers, in any of the three pricing structures below, or a hybrid built from them. The recurring number depends on your volume and headcount profile, which is what the diligence form on the next page captures. The setup cost we size and commit in writing the day you submit.
One monthly number. Predictable. Easy to budget.
Setup fee at signing + go-live. Flat monthly platform fee. 36-month minimum term. Annual escalator built in. No per-seat or per-load surprises.
Operations with steady volume that want a single line item on the P&L. Procurement loves it. Finance loves it. Easy expense allocation.
You pay the same in a slow quarter as a busy one. If volume grows materially, we revisit the number on renewal, not mid-term.
Pay only for the work the platform does.
Smaller setup fee at signing + go-live. A per-load price billed monthly. Monthly minimum so the platform is viable in a downturn. 36-month minimum term.
Operations with variable volume or who want compensation aligned to throughput. If your loads compress in a soft market, your bill compresses with you.
You take some upside risk too. If volume grows fast, your bill grows with it. Per-load math is more legible to operators, less legible to procurement at first read.
Base + variable. The shape most enterprise deals end up in.
Setup fee at signing + go-live. Base monthly fee that covers infrastructure and team. A small per-load fee above a stated volume threshold. 36-month minimum term.
The most common shape for mid-market and enterprise deals. The base covers our cost to run your deployment, the per-load piece scales with the value we are creating.
Two line items on the invoice instead of one. Easier to model upside for both sides. Harder to compress to a single number for a back-of-napkin comparison.
We commit the setup fee number in writing on day one. It covers dedicated deployment standup, data migration, customization layer built to your business unit, AI crew tuning to your voice and customer base, and integration work to in-scope systems. Half at contract signing, half at go-live. No professional services overage. No surprise "additional implementation hours" line items. If we underestimated something, we eat it.
We do not charge per dispatcher, per user, or per agent. The whole AI crew is included. Adding humans does not add cost.
The platform's API surface is yours to use without a per-call meter. EDI volume, document parsing, agent activity — all included.
Whatever path you pick, the formula is in the contract. The number on next month's invoice is predictable from your actual operating data.
Get a scoped proposal in one business day.
We model the deal against your actual numbers, not a template. Setup cost we commit on day one. Recurring price comes back in whichever of the three structures fits your P&L best.
Open the diligence form →Scoped, transparent, defensible.
Pilot pricing is set against the pilot business unit Crane selects, the scoreboard Crane signs off on, and the diligence numbers Crane shares with us. We hold three commercial principles regardless of unit selection or path chosen:
- One scoped fee for the 120-day pilot. No per-seat creep, no metered surprise.
- Expansion terms agreed before we start. Crane knows the next step's commercial shape before signing the pilot.
- Walk-away clean if the scoreboard is not met. Crane keeps the data and the customization layer. WakeTech.ai walks.
Questions about cost, scope, or any of the above go directly to the person who wrote this page. No procurement layer between Crane and the founder.