Marvin hlongwane

**VoicePulse Analytics** is a **cross‑channel voice analytics platform** for restaurant groups (10–200 locations) that already run voice agents or heavy phone volume but lack insight into what guests actually say. Today, operators see only call counts and average handle time while **churn reasons, menu confusion, and promo performance are buried in audio**. VoicePulse ingests recordings and transcripts from drive‑thru agents, Phone2POS‑style IVRs, and call centers to surface **actionable patterns**. - **Menu intelligence**: detects items frequently misunderstood, modifier friction, and common substitution requests by region. - **Lost‑order tracking**: identifies calls where the guest abandoned before completion and tags likely reasons (price shock, wait time, UX). - **Voice‑of‑customer themes**: clusters complaints and compliments into ops dashboards (wait times, packaging, delivery issues). - **Agent benchmarking**: compares AI vs human performance, consent‑safe, to tune where automation actually helps. - **Compliance & privacy‑first**: regional data residency, auto‑redaction of PII, and configurable retention to satisfy regulators. The moat comes from **cross‑brand benchmarks and increasingly accurate restaurant‑specific taxonomies** that general purpose call‑analytics tools lack.

**ServerWhisper** is a **whisper‑mode voice coach for casual dining servers**, delivering real‑time upsell prompts through discreet earpieces during table‑side ordering. Mid‑market casual chains complain about **plateaued check averages and inconsistent suggestive selling**, while current training is static and quickly forgotten. ServerWhisper listens to the order (via table‑side tablet mic or beltpack) and quietly suggests **high‑margin add‑ons or bundles** tailored to that guest’s context. - **Contextual suggestions**: uses time of day, current items, party size, and promo rules to propose 1–2 relevant upsells only. - **Micro‑training**: post‑shift, servers see which upsell lines worked and practice with short, TikTok‑style playbooks. - **Location experiments**: HQ can push A/B scripts (e.g. dessert vs appetizer push) and measure impact on check averages. - **Compliance guardrails**: ensures suggestions align with alcohol policies and allergy disclosures. - **Hardware‑light**: works with cheap Bluetooth earpieces and existing POS/tablet mics; no big hardware rollout. This builds a moat via **unique labeled data on what verbal scripts convert** across demographics and concepts.

**GhostVoice Hub** is a **voice ordering router for ghost kitchens and virtual brands**, built for operators running 5–30 brands out of a single facility. These kitchens juggle Doordash/UberEats/white‑label web orders but **have no coherent way to handle voice‑driven reorders from loyal customers or call‑in catering** without chaos. GhostVoice gives each brand a unique phone/voice channel and uses a **central brain** to route orders to the right make‑line and prep station. - **Brand‑aware agents**: separate voices, greetings, and menus per virtual brand, all administered from one backend. - **Cross‑brand upsell**: if Brand A is slammed, intelligently offers similar items from Brand B with accurate ETAs. - **Kitchen load balancing**: factors ticket queue and cook times before confirming pickup/delivery times with customers. - **Catering mode**: long‑form conversational flows tuned for corporate and event orders, auto‑generating quotes and deposits. - **Marketplace integration**: syncs inventory and 86’d items across third‑party platforms and voice in real time. Defensibility comes from **deep operations integration** and **multi‑brand order routing**, not just generic call bots.

**Phone2POS** converts **inbound phone calls** to restaurants into structured orders pushed directly to the POS, targeting busy independents and small chains still drowning in phone volume. Owners complain across Reddit and G2 that online ordering hasn’t killed phone calls and that third‑party call centers are expensive and error‑prone. Phone2POS sits on a Twilio‑like number or existing line forward and runs a **conversational IVR** that understands accents, modifiers, and pickup vs delivery, then posts straight to the POS with ticket printing. - **Drop‑in deployment**: keep your number, just forward to Phone2POS; 30‑minute onboarding. - **POS‑native tickets**: creates checks with correct revenue centers, taxes, and kitchen routing; no re‑typing by staff. - **Smart clarifications**: dynamically asks for missing info (e.g. sauce choice, well/medium) based on restaurant’s forced modifiers. - **Callback & SMS**: automatically texts confirmations, ETAs, and links for payment or tip capture. - **Fraud and prank filters**: phone reputation scoring and pattern analysis to reduce fake orders, a big pain for pizzerias and wings shops. Pricing is **per‑call + SaaS seat**, with a clear, immediate ROI in reclaimed staff time and fewer errors.

**KDS Talkback** adds **voice‑driven controls and alerts to kitchen display systems**, targeting high‑volume kitchens using KDS but still relying on shouting and tickets to coordinate. Line cooks and expo staff complain that **hands‑busy workflows make KDS under‑utilized**, and vendors like Toast/KDS don’t invest deeply into voice UX. KDS Talkback runs on inexpensive mics/speakers above the line and ties into existing KDS APIs to let staff **update, query, and be alerted by voice**. - **Hands‑free actions**: “Mark table 42 fries done”, “86 mussels”, “Fire mains for 23” mapped directly to KDS events. - **Priority alerts**: when a dish is about to miss promised time, the system speaks a short, localized alert to the right station. - **Expo queries**: expo can ask “What’s still missing on 14?” and get a concise spoken rundown. - **Shift learning**: tracks bottlenecks and suggests station‑level prep or menu tweaks based on recurring slow items. - **Vendor‑agnostic**: lightweight adapters for major KDS vendors; sold as an add‑on, not a replacement. Moat is built through **granular operational data and station‑specific voice models** trained on noisy, hot kitchens.

**ShiftVoice OS** is an end‑to‑end **drive‑thru voice agent** for QSR franchise groups (20–300 locations) that plugs directly into existing POS and headsets. Franchise operators struggle with **chronic understaffing, slow lanes, and missed upsell scripts**, while current AI pilots from giants (McD, Wendy’s) are closed ecosystems and inaccessible to mid‑market brands. ShiftVoice routes every order through a **noise‑robust, menu‑aware agent** that handles 80–90% of interactions and gracefully escalates to staff on edge cases. - **Franchise‑grade integrations**: certified connectors for Revel, Toast, Brink, Xenial, NCR used by QSRs, not just indie POS. - **Menu‑tuned NLU** trained per brand (combos, LTOs, regional items) with live sync from POS pricing. - **Lane optimization**: real‑time queue estimates, dual‑lane balancing, and “fast order” prompts based on historical patterns. - **Upsell engine**: A/B‑tested scripts per daypart; tracks incremental revenue per lane, per store. - **Franchise analytics**: compare AI vs human order time, accuracy, and upsell by location and shift in one console. This is defensible via **brand‑specific training data, integrations, and lane performance benchmarks** that compound over time.