← Sticker Wall AGENT · DETAIL

Helping restaurants and hotels identify lost phone call revenue and making 20,000 RMB/month through reservation recovery revenue-sharing

Workflow: First, export the last 90 days of call logs for the target restaurant or hotel's extensions, calculate the missed call r

AGENT

Key Fields

FIELD STAMPS

IndustryE-commerce / Retail

RegionEurope

ScaleSME

ChannelOnline

🔧 Workflow

First, export the last 90 days of call logs for the target restaurant or hotel's extensions, calculate the missed call rate, missed call time distribution, and potential reservation loss, and turn it into a one-page diagnostic report to send to the owner or regional manager. After signing, configure the voice agent to connect to the reservation or hotel PMS system, covering high-frequency intents such as reservations, rescheduling, and business hours inquiries. Spend 1-2 hours a day inspecting abnormal dialogue recordings, manually taking over difficult calls, and optimizing the script library. At the end of the month, settle revenue sharing with the client based on the number of recovered reservations or reservation revenue. After running successfully for one store, replicate the configuration template to chain stores in the same city.

🛠 Setup Requirements

Requires understanding basic telephone system (SIP trunking, cloud PBX, call forwarding rules) and restaurant POS or reservation system (OpenTable, hotel PMS) integration logic. You need to know how to use PolyAI or its alternative platforms (Retell, Vapi) to build multi-turn dialogue flows and configure interruption, transfer-to-human, and exception fallback rules. You must be able to use reporting tools to convert call data into loss figures that the boss can understand. Spend 2-3 weeks in the early stage running the complete closed-loop of a model store, accumulate recovery data, and then replicate it to chain stores in the same city. The overall technical barrier is lower than self-developed models but higher than pure SaaS resale.

🧰 Toolchain

🔧 PolyAI or Retell voice Agent platforms
🔧 APIs for reservation systems (OpenTable, etc.) or hotel PMS
🔧 Call recording and missed call analysis reporting tools
🔧 Stripe revenue sharing settlement and reconciliation

💰 Revenue

① Revenue sharing based on recovered reservation revenue (main income): Stores pay a 10%-20% share based on recovered reservation revenue. Serving 5-8 chain stores yields a monthly income of about 20,000 to 40,000 RMB. The proportion of this path in monthly income has not been publicly disclosed (figures come from case studies and no one has independently verified them); ② Store monthly deployment and operations fees: Stores pay $300-$800/store per month. The number of signed stores has not been counted (case study data, independent verification not done), and there is no data on how much this path accounts for revenue; ③ Landing consultant annual package: Charge chain restaurant groups an annual package fee for voice Agent landing consultant services. The card mentions that similar annual packages can be priced at tens of thousands of dollars (this price is quoted from case studies and has not been independently verified). The number of closed annual packages has not been disclosed, and its proportion of revenue is also unclear; ④ Opportunity item - Performance-based wagering revenue sharing: Based on Hawksmoor (accumulating about £1.6 million in reservation revenue through the platform and 80,000 in-store diners, case source not independently verified), the fixed monthly fee is swapped for a performance-based revenue share on recovered revenue. The amount of income that can be obtained after the bet remains publicly unannounced.

💸 Cost

Voice platform invocation fees are about $0.05-0.3 per minute. When a single store has a high monthly call volume, costs fluctuate with the volume. Combined with diagnostic report tools and alert subscriptions, the initial monthly cost is about 1,000-3,000 RMB. As the customer volume increases, it scales linearly with the call volume, accounting for 20% to 30% of the revenue-sharing income.

⏱ Time Investment

In the early stage, deployment and integration for each store take about 1-2 weeks, including script sorting, system integration, and trial operation. After entering the stable period, spend 1-2 hours a day inspecting abnormal dialogues, updating the script library, and responding to customer questions, which can simultaneously maintain 8-12 stores.

🚀 Getting Started

Step 1: Pick 3 local high-ticket restaurants or boutique hotels, help them export call data for free to calculate missed call losses, and use real figures to make the first one-page diagnostic report in exchange for a pilot opportunity; Step 2: Use Retell or Vapi to cheaply build a minimum viable voice agent capable of making reservations and run it for two weeks to get before-and-after comparison data on recovered reservations; Step 3: Use the data to sign a formal revenue-sharing contract, solidify the entire process into a replicable template, and sell it to same-city chains and hotel groups.

🔑 Keys to Success

✅ Calculate the accounts before selling the product, using the missed call loss amount and recovery return to impress the boss instead of talking about large model concepts
✅ The revenue-sharing model lowers the customer's decision-making threshold, and your own income compounds with store expansion and call volume growth
✅ Focus stubbornly on a single vertical scenario (restaurant reservation or hotel booking), distilling intent libraries, script templates, and integration scripts into reusable assets
✅ Retain manual fallback and transfer mechanisms to keep the customer complaint rate to a minimum, as renewal rate is the lifeblood of this business
✅ Prioritize securing regional pilots for chain brands, using single-store data to negotiate batch deployment for the entire brand

⚠️ 风险

⚠️ Voice agents giving wrong reservation times or party sizes can trigger customer complaints; manual fallback transfers must be retained and responsibility boundaries clearly defined
⚠️ Large clients may bypass service providers after a successful run and sign directly with platforms like PolyAI; contracts must lock in the revenue-sharing period and exclusive service terms
⚠️ Increases in platform invocation fees or changes in platform policies will compress gross margins, requiring multi-platform deployment capabilities to avoid single-point dependency
⚠️ Different regions have different call recording and data compliance requirements (such as Europe's GDPR), and illegally processing call data carries legal risks

📌 Real Cases

📌 The 14 stores of the UK steakhouse chain Hawksmoor have their front-of-house teams handle phone calls concurrently, missing about 42,000 calls a year. After deploying the PolyAI voice intelligent agent, reservation revenue increased by £1.6 million annually, a case publicly and personally described by the reservations director.
📌 PolyAI, with restaurant and hotel phone reservation automation as its core scenario, secured $86 million in financing in December 2025 at a $750 million valuation. GetLatka estimates its ARR to be about $35 million, validating the payment capacity of this track.
📌 PrestoBot AI, founded by a former technical executive at Meituan Takeout, launched its smart restaurant AI phone ordering system in March 2026, running a trial in 100 restaurants in Los Angeles, USA, demonstrating that the same model is being batch-replicated by entrepreneurs in both China and the US.

Sources