Skip to main content
Docklyx
PricingBlogContact
esen
Log inStart free trial

On this page

All modules

  • 01Portal + Bookings
  • 02Dock + Yard
  • 03Multi-site
  • 04QCD Scoring
  • 05Alerts + KPIs
  • 06Cerebro AI
Home/Features/Cerebro AI
AI
Module 06

An autopilot that doesn't just suggest. It acts — and lets you roll back.

Cerebro AI is an autopilot for dock and yard ops: real-time slot suggestions in the carrier portal, an optimizer that drafts moves on your board, a WhatsApp bot that can auto-schedule, predictive insights, plain-text queries — and a decision queue where any exception waits for a human. Every automatic action runs through simulation mode first, is logged, and rolls back with one click.

Predictive insights 4h / 24h / 7dPlain-text queries → tablesSlot optimizerWhatsApp botIncident classificationAssisted carrier scoringDry-run + rollbackManual / auto / off modesPortal slot suggestionsSimulation mode
Start free trial →See the demo
At a glance
Six capabilities
Insights · queries · slot intelligence · WhatsApp · classification · scoring
Modes
Manual (suggests) · Auto (acts) · Off — configurable per action, per DC
Safety
Simulation mode · dry-run before execute · rollback with double protection
Cooldowns
Insights every 7 days (Pro) · 30 days (Starter)
Horizons
Predictions at 4h, 24h and 7 days
Plans
Starter (basic) · Pro (full suite) · Enterprise
Decision queue
Anything outside policy waits in an exception inbox for a human
Demo · plain-text queries

Ask your ops. No SQL required.

Pick a question and watch it translate, validate and execute. The self-correction loop rewrites the query on errors.

app.docklyx.com / reports / askCEREBRO
Ask it
Questions land in the suggestions inbox. You can save them as favorites.
→
Converting your question
Checking it's safe to run
Running the query
Module anatomy

Three modes, one cycle with a brake.

Every automatic action passes through four gates: suggest, preview, execute and — if something went wrong — roll back. By default, Cerebro only suggests. You raise the autonomy level action by action.

manual
Manual · suggests

Default. The AI proposes. You accept or dismiss.

auto
Auto · acts

The AI executes without waiting. Always with dry-run + rollback.

off
Off · disabled

The capability doesn't run for that action.

AGENTIC ACTION CYCLE
1Suggest

Insight with evidence · links to source · expires in 7 days

→
2Preview

Dry-run: 'this would move 3 slots' without touching anything

→
3Execute

Action applies. Logged with who, what, when

→
4Roll back

One click reverses the action. Protected against double-rollback

Evidence — every insight links back to the row or event that triggered it.
Cooldown — Pro regenerates every 7 days; Starter every 30.
Audit log — who changed the mode and when is recorded.
Six capabilities · two on the way

Every feature earns its place in the yard.

01Predictive insights

Reports that see 4h, 24h and 7 days ahead.

Reads your bookings, dwell times, incidents and no-shows to spot peak hours, bottlenecks and trends. Each prediction names its time window. Refreshes every 7 days on Pro, 30 on Starter.

02Slot intelligence

Picks the best slot for the carrier. Suggests moves on your dock board.

Combines available capacity, predicted congestion and the carrier's 90-day on-time history. Every move runs in simulation first, and every suggested slot is checked against your real schedule before the carrier sees it.

03Plain-text questions

Ask in plain English. Get a table back.

Type the question the way you'd ask a coordinator. The system answers safely — read-only, no personal data, retries automatically on errors. Save your favorites for next time.

04WhatsApp bot · auto-schedule

Carriers book, cancel — and even say 'whatever you've got'.

The bot understands messages like 'move my Friday slot to 3pm' and resolves them. With auto-schedule on, it accepts open-ended replies like 'whatever you've got this week' and picks the best fit. Anything unclear goes to a human.

05Incident classification

From 'he got aggressive' to type + severity.

When the guard describes an incident in their own words, the AI sorts it into a category (aggressive, no PPE, false docs, bribery, etc.) and suggests a severity. The manager confirms in one click.

06Assisted scoring

The AI explains why this carrier is a risk.

On top of the carrier's A–F grade, the AI adds plain-English context like 'no-shows tripled last month'. Shows up as a note on the carrier page. The grade itself doesn't change.

soon
07Document scanning

Scan bills of lading and packing lists.

Coming soon. Pulls reference numbers, recipient and SKU automatically from document photos. Lands integrated into the post check-out flow.

soon
08Plate reading

Automatic recognition at the gate.

Coming soon. Reads the truck's plate as it pulls up and matches it against the one on the booking. On a mismatch, the guard gets an alert before granting access.

Action matrix

Every action with its own safety gate.

The default is Manual. Admin moves to Auto only what the team already trusts. Everything is logged and reversible.

Action
Default mode
Guardrail
Reassign slots
Manual
Dry-run · shows how many bookings would move
Offer slot to waitlist
Auto
Rollback restores slot to the original owner
Classify incident
Manual
Manager confirms type + severity before saving
Reply on WhatsApp
Auto
On ambiguity, the bot escalates to a human
Save favorite query
Manual
Never writes · read-only
Generate insight
Auto
7d cooldown · atomic reservation avoids dupes
Auto-schedule from WhatsApp
Manual
Free-text intent only · falls back to slot suggestions · transporter confirms
Adjust schedule (move appointment)
Manual
Rollback restores original time · routed through automation policy engine
Contact carrier
Manual
Skipped on auto trigger — messages can't be unsent
A day with Cerebro

From policy to rollback.

Nothing happens without you being able to see it first — and undo it later.

Day 0
Admin sets policy

Per action: manual / auto / off. Every change is logged with who and when.

01
00:05
Nightly cron runs

Insights generate with atomic reservation to avoid dupes. Cooldown 7d/30d by plan.

02
07:12
Manager opens the dashboard

Sees 4 fresh insights. Each with linked evidence and a dry-run button.

03
07:13
Optimizer dry-run

'This would move 3 slots and offer 2 gaps to ACME'. Nothing touched.

04
07:14
Accept and execute

Action applied. Affected carriers get notified. All logged.

05
09:04
Plain-text query

'Avg dwell per dock yesterday'. Translate → validate → execute. 312ms.

06
11:28
Something went wrong

A carrier complained about the reassignment. Rollback in one click, protected against double-reverse.

07
14:00
Weekly review

Admin reviews the log: what ran on auto, what was rolled back, why.

08
Day 7
Promote a rule out of simulation

After a week running in simulation, contact-carrier on dwell breach is moved from Manual to Auto with confidence — every dry-run was reviewed.

09
07:50
Exception lands in the decision queue

The bot sees an intent it can't safely resolve (out-of-window request). Instead of acting, it queues it for the manager — full context, one click to approve or reject.

10
How it connects

What it shares. What it receives. What depends on it.

Module
Shares →
Receives ←
01Portal + Bookings
Bookings · no-shows · punctuality
Reassignment and cancellation suggestions
→
02Dock + Yard
Real dwell · dock utilization
Slot-optimizer proposals
→
03Multi-site
Per-site policy · regional context
Insights scoped to the active DC
→
04QCD Scoring
A–F grade as feature for the optimizer
Assisted-scoring annotations on the carrier page
→
05Alerts + KPIs
Events: rollback, insight accepted, query ran
Critical insights go to the alert inbox
→
06Audit log
Every automatic action with who, what, when
Policy changes visible in history
→
What it moves

Four levers when Cerebro runs.

14 min
Time per question

What a BI report used to take — now it's a plain question and a table.

-22%
Unused dock gaps

The optimizer catches gaps that used to stay empty because nobody had time to review them.

72%
Insight acceptance

Of the insights Cerebro generates, that share ends in an applied action — not noise.

0
Un-reversible actions

All automatic actions are reversible. There's never an 'oops' you can't undo.

Measured across DCs with Cerebro AI on Pro for ≥ 3 months · Q1 2026.
Module questions

Things people ask in the demo.

If it's not here, sales replies same-day — not a form into the void.

Every automatic action is reversible with one click. Dry-run lets you see the effect before executing; the audit log records who ran what; rollback is protected against double-reverse.
Google Gemini by default, via a centralized gateway. On Enterprise you can point to a dedicated or self-hosted model.
No. The system only runs read queries — nothing can write, modify, or delete your data. Queries also can't return personal information or data outside your role's access.
Pro: every 7 days per DC. Starter: every 30. Atomic reservation prevents duplicates. If you access before cooldown, you see the current insights.
Only when an admin enabled Auto mode for a specific action. Queries are always read-only. The default for everything is Manual.
It only responds to the carrier's verified number. And only books within windows the DC allows — if anything falls outside policy, it escalates to a human.
← Previous moduleAlerts + KPIs
PRICING

A plan for every yard size.

Monthly or annual billing (save 17%).

Starter
$299 USD
per month
Start
  • -1 distribution center · Up to 8 docks · 10 users
  • -Booking portal + QR check-in
  • -Yard management + dock planner
  • -Real-time dwell alerts
POPULAR
Growth
$539 USD
per month
Start
  • -Up to 2 distribution centers · 15 docks · 20 users
  • -Everything in Starter +
  • -Automated WhatsApp to carriers
  • -Dwell cost tracking
Professional
$879 USD
per month
Start
  • -Up to 3 distribution centers · 25 docks · 30 users
  • -Full QCD scoring + cost tracking
  • -Cerebro AI (NL queries, slot optimizer)
  • -KPI alerts + daily digest
  • -WhatsApp bot + notifications
Enterprise
Custom
multi-site · SSO
Talk to sales
  • -Unlimited distribution centers / docks / users
  • -Outbound webhooks + ERP/TMS
  • -99.95% SLA · SOC 2
  • -Dedicated CSM
  • -Optional on-prem

21 days free · No credit card required · 1-day setup

AI
Next step

Let your ops ask you, not the other way around.

We stand up an environment with your real data. We set Manual on everything. You move things to Auto when you trust them.

Start free trial →Book a demo

Here when your yard needs it.

Docklyx

Dock and yard intelligence. Optimize your logistics operations with AI.

Product
  • Features ↗
  • Pricing ↗
Company
  • About ↗
  • Contact ↗
Resources
  • Blog ↗
  • Dwell Cost Calculator ↗
  • Yard Operations Audit ↗
  • Dock Capacity Calculator ↗
Legal
  • Privacy ↗
  • Terms ↗
© 2026 Docklyx. All rights reserved.
Mexico · EN ⌄