Choose a task and Zeaulu will take you to the right workflow.
Search your approved SharePoint site using your Microsoft account. Results are limited to resources your account is authorized to view.
Add individual files, entire folders, or ZIP archives. Zeaulu organizes them locally so they can be found later through the Library or Knowledge Search.
Save dealer, rooftop, group, and contact information once. Select it from Dealer Context to reuse it in Prompt Builder and supported report fields.
Search by dealer, rooftop, group, contact, role, email, or phone.
Build your AI team, install quality controls, preserve approved work, storyboard complex deliverables, and build purpose-driven engines that produce repeatable results.
Give this prompt to your preferred AI program to build a connected AI operating system instead of a loose collection of prompts.
Teaches the AI how you work, what good looks like, and which preferences should shape every future result.
Creates the durable rules the AI must follow on every assignment.
Assembles the right specialists for each assignment instead of using one generic voice for everything.
Installs the committee method that turns AI into a coordinated decision team.
Routes each task to the right level of AI effort instead of running every playbook every time.
Creates reusable teams for recurring domains such as analytics, customer success, security, marketing, writing, finance, or product work.
Creates short commands that activate the right team or quality process without pasting a full prompt every time.
Combines the approved pieces into one clean instruction block for a project, workspace, or custom-instruction field.
Forces the AI to review and improve its work instead of treating the first draft as final.
Prevents confident-looking answers from outrunning the available evidence.
Uses adversarial review to find failures that ordinary proofreading misses.
Protects approved work from accidental redesign, regression, or scope creep.
Converts recommendations into an executable plan with ownership and proof of completion.
Turns corrections and failures into durable improvements instead of repeated apologies.
Turn AI from a yes-bot into a working committee. Assign roles, force useful disagreement, compare tradeoffs, and end with one recommendation.
Use this when you want AI to confirm the ask before it creates a file, report, image, presentation, or final answer.
Require AI to check every box and only deliver when it is 97-99% confident enough.
Train AI to push back when a better path, simpler process, or less risky option exists. Do not just agree.
Preserve approved presentation structure while changing only the requested content.
Use this for professional image work where identity, faces, and objects must remain intact.
Create a cleaner professional image without turning the person into a different person.
Protects faces, identity, logos, text, layout, composition, and approved source elements when editing or reusing images.
Lock the report shell, use only source-supported data, keep charts accurate, and prevent internal QA language from leaking dealer-facing.
Do not call a build ready until engineering, presentation design, CSM workflow, dealer experience, and a deliberate red-team review have independently approved it.
Use memory to make AI easier to work with over time without repeating yourself every session.
Build case recaps, CSS requests, success plans, renewal prep, and dealer follow-up with clear audience-aware language.
Explains looping in plain language and installs an iterative improve-until-ready workflow.
Separates what AI can complete now from what must wait for access, approval, information, or capability.
Converts requests into buildable requirements, scope boundaries, tests, and definition of done.
Determines whether a project is ready to launch, configure, deploy, train, or hand off.
Tests the finished workflow as the actual user would experience it.
Turns a completed solution into a workflow people actually use.
Prevents work from disappearing between teams by assigning ownership and escalation paths.
Ensures content fits the audience, brand, channel, and desired action.
Turns one approved source into multiple channel-ready assets without changing meaning.
Stops polished but unsupported marketing language from slipping into final content.
Connects campaign creation to measurement, learning, and iteration.
Prevents overpromising by identifying what the platform and tools can actually do.
Documents important technical or operating decisions so future teams understand why they were made.
Determines what must be retested before changing an approved file, app, report, workflow, or presentation.
Makes failures visible, diagnosable, and recoverable instead of silent.
Protects releases by preserving baselines, documenting changes, and preparing recovery steps.
Reviews AI work for privacy, secrets, prompt injection, metadata, and unsafe exposure.
Keeps current requirements, superseded decisions, baselines, and open questions from being mixed together.
Classifies problems by impact, urgency, risk, workaround, and escalation path.
Checks whether outputs are readable, navigable, understandable, and usable by different users.
Makes confidence meaningful by tying it to evidence, assumptions, and risk if wrong.
Defines when AI should pause instead of forcing a risky or unsupported deliverable.
Creates a purpose-built AI operating engine around the result the user wants to achieve.
Chooses the right engine for the requested outcome before work begins.
Coordinates teams, playbooks, storyboard, gates, access, and delivery into one unified result.
Evaluates whether an engine is producing the intended results and should be improved, split, merged, or retired.
Plans sequence, narrative, user journey, and visual progression before final production.
Defines the core story that can be reused across decks, reports, videos, campaigns, and training.
Produces verified, balanced, source-supported intelligence rather than a search-result summary.
Evaluates alternatives and produces an accountable recommendation with tradeoffs.
Creates audience-specific communication that produces a defined response or action.
Builds visually professional, logically sequenced, brand-compliant presentations.
Transforms source data into an accurate, understandable, decision-ready report.
Converts requirements into a tested release without breaking approved functionality.
Moves a solution from readiness through launch, adoption, and sustained use.
Creates or modifies visual assets while protecting identity, logos, composition, and approved elements.
Defines the full sequence for turning AI into a repeatable performance engine.
Manage your Zeaulu experience, local Library, approved connections, privacy controls, platform guidance, and report defaults.
Your display name and selected working role appear in report attribution. Role selection is not account authentication.
On uses the Light theme. Off uses the Dark theme.
Zeaulu automatically detects Windows or macOS and applies the correct keyboard guidance.
Each test package is loaded fresh automatically without clearing browser cache or cookies.
When enabled, generated content favors dealer-facing language, hides internal coaching labels, and suppresses internal notes from report output.
Reports use the black, white, and gray Zeaulu theme by default. Author metadata is preserved in generated files, but hidden from the CSM-facing UI.
Connections are managed here inside Zeaulu. Nothing is connected or uploaded unless you intentionally configure and use an approved service.
Used only when you select Push to Copilot.
Use only an approved internal HTTPS endpoint.
Use only an approved internal HTTPS endpoint.
Reserved for an approved CDK-owned API and security implementation.
Zeaulu does not automatically upload Library files to an external AI service.
0 files stored locally
Library files remain on this device. Back up the Library before moving Zeaulu to another computer or browser profile.
Files are stored in the local Zeaulu Library for search and retrieval.
CRM KPI Report: Zeaulu supports 9 CRM KPI source options: Desking Deal Detail Report, Desking Deal Summary, Desking Sold Details Report, Internet / Phone Combined Performance, Internet Response Time, Dealership Lead Source Stats, User Activity Completion, User Activity Performance, and User Appointment Confirmation. Use the native CSV exports for CRM reports and the XLS exports for the three eDESK Desking reports. Load any combination for a partial report or all 9 for full CRM KPI coverage.
Performance Report: Upload the existing Performance Report source/export package supported by the Performance Report builder. Group totals are shown when group-level source data is present. Store-level values are shown only when store-level source data exists.
Utilization Report: Upload the existing Utilization Report source files or import template values used by the Utilization Report builder. The dealer-facing report should generate only after supported dashboard/status values are populated.
Intelligence Suite: Upload the Intelligence Suite source files, ZIPs, or reports used by the existing Intelligence Suite builder. The builder should parse by filename, ZIP internals, headers, and source columns when available.
Enter the report context before uploading source data.
Add source files for the utilization review. Files remain on this device.
Review file recognition, duplicates, and report readiness before generating.
Preview is generated from parsed ZIP and CSV data using the approved CDK report shell.
Export the parsed, data-driven report as a standalone offline HTML report.
Convert files, optimize attachment size, and package working files without leaving Zeaulu.
Browser-safe conversions are handled locally. Complex Office conversions that require a server engine will fail closed instead of pretending to convert.
Images are compressed locally. PDFs and Office files are measured and flagged when they need external compression; Zeaulu will not fake compression results.
