Meet Wally
Your built-in AI,trained on your org.
Wally is Turtini's AI assistant. 650+ tools across every module — CRM, accounting, planning, email, hospitality, the in-browser IDE — wired to your real org data. Voice or chat, walkthrough then action, and every preview tells you whether you can undo it.
Powered by Claude. Private to your org. Included with every plan.
What Wally does
Eight buckets, one chat box.
Wally is read-and-write. It works across every module you have on, with the same tools you'd reach for manually — and a few that don't have a button anywhere else.
CRM, deals, and follow-up
Search any account, contact, opportunity, or quote with a sentence. Add notes, log activity, move stages, schedule follow-ups, draft emails — all from chat. Pipeline weight, slipping deals, and cold contacts surface without a BI tool.
Accounting v2 — close, runway, audit
Draft month-end close, surface anomalies (Anomaly Radar), forecast cash runway, generate the board pack and MD&A narrative, categorize bank txns, verify the hash-chained audit trail, post closing entries — and ask the books in plain English.
Planning — Wally Insights
Critical-path analysis, EVM rollup, risk-weighted Monte Carlo, capacity heatmaps. Wally surfaces the schedule risks before they slip, drafts mitigation plans, and writes them into the project as new tasks once you confirm.
Hospitality — restaurants, lodging, retail
Auto-86 menu items when ingredients deplete (Smart Cascade). Recommend room rates from occupancy + comp set. Draft pre-arrival guest emails. Post folio room-charges. Run an occupancy report. The same Wally for every storefront.
Email, socials, articles, decks
Draft a campaign, suggest an audience segment, enroll contacts in a journey, explain why a send underperformed. Publish and tag help articles. Spin up a slide deck from your CRM data and refresh it from the source on demand.
Builder + IDE pair-programming
Edit Builder pages by description ("add a discount block to the homepage"). In the in-browser IDE, Wally is the AI pair-programmer — reads files, proposes diffs, commits on confirm. Restore from any safety snapshot in one ask.
Environments + module ops
Launch and tear down OpenShift labs, forecast environment cost, extend a soon-to-expire lab. Enable or disable platform modules. List what your org is paying for, in plain language.
Daily briefing + walkthrough
"What happened yesterday?" "What's on my plate today?" Wally rolls up activity across every module. Ask "how do I…" and Wally walks you through the matching help article — then offers a one-click action card to do it for you.
How it works
Walkthrough first, action second, no surprises.
Wally explains before it does. Every write previews, and the preview tells you whether you can take it back. Every action is recorded in a ledger you can verify.
01
Type, click, or just say "Hey Wally"
Open Wally from the chat icon in the top nav, the Ask Wally button on every record page, or the wake word on devices with a mic. Wake-word listening is local + free; the actual transcription only kicks in once you say "Hey Wally."
02
Walkthrough first, then an action card
For "how do I…" questions Wally narrates the workflow from the in-app help library — so you learn — then offers a confirmable action card if a tool covers the task. You can decline the card and follow the walkthrough manually, or click "Do it" to have Wally execute.
03
Every write previews before it commits
Wally never silently mutates your data. When it's about to write — a new note, a stage change, a journal entry, a published article — it shows the field, the record, and the new value. One click to approve, one click to send back with feedback, one click to drop.
04
Undo where undo exists — and it tells you first
Every Wally write is recorded in your org's hash-chained audit log. Where the action can be reversed, the card says so before you confirm and an Undo link stays live for 24 hours; the change reverses cleanly, including downstream effects (a refund un-posts the related JE too). Where it cannot be reversed — an email already sent, money already moved — the card says that instead, while you can still decide.
Twelve of the 133
A taste of what's actually wired up.
Real tool names from the Wally tool registry. The full set spans every active module on your org and grows as the platform grows.
accounting_draftMonthEndCloseDrafts every closing entry with a confirm-each-line preview
accounting_anomalyRadarSurfaces unusual transactions before they hit your P&L
accounting_cashRunwayLive runway forecast against your actual burn
project_evmEarned-value rollup across every active project
project_criticalPathCritical-path with risk-weighted Monte Carlo schedule
lodging_rateRecommendationSuggests nightly rates from occupancy + comp set
restaurant_smartCascade86Auto-86s menu items when their ingredients deplete
email_suggestSegmentBuilds an audience from a plain-English description
email_explainResultsDiagnoses why a campaign over- or under-performed
deck_createSpins up a slide deck from CRM, Planning, or Restaurant data
environment_costForecastForecasts lab spend before you launch a workshop
ide_restoreSnapshotRestores any IDE safety save with one ask
Built on Claude
Private to your org. Auditable to the line.
Wally runs on Anthropic's Claude. Your conversations and your records stay scoped to your org — never shared across tenants, never used to train a foundation model.
Tenant-scoped
Wally only sees data for the org you're signed into. Switch orgs, switch context — no leakage.
Auditable
Every action lands in the Wally activity log with the prompt, the tool call, the resulting record diff, and an Undo link.
Always verifiable
Wally cites the records it's reasoning about with clickable links — jump straight to the source to verify before you act.
Hash-chained
Accounting writes (JEs, closes, audit chain) are appended to a tamper-evident hash chain so any retroactive edit is detectable.
Voice runs local
The "Hey Wally" wake word uses the browser's on-device Web Speech API — no audio leaves your machine until you trigger it.
Org-priced + transparent
Wally usage (Claude tokens, STT seconds, TTS chars) bills back to your org as platform usage at a published +15% markup, posted to your books as an expense.
Sovereignty-ready
Admins can point Wally at an Anthropic- or OpenAI-compatible endpoint they operate — AWS Bedrock GovCloud (FedRAMP High / IL5), Google Vertex AI in any region, vLLM or Red Hat AI Inference Server in your VPC, Azure OpenAI, or a custom gateway — so prompts and responses never traverse Turtini's infrastructure on the inference path. Covers every customer-facing AI surface: chat, Builder Docs, AI Builder Review, Builder Sites, Builder Design.
Included
Wally is on for every org.
No add-on. No separate Wally licence. No usage tier on your plan. Wally consumes Claude tokens and STT/TTS minutes; that's billed back to your org at our published +15% markup and posts to your books as an expense automatically.
No credit card. No trial period. Wally is on from day one.
Ask Wally anything.
Sign up free, open the chat icon in the top nav, and put Wally to work — or just say "Hey Wally" on any device with a mic.