Request for Proposal (RFP)

WhatsApp AI Booking Assistant (Phase 1)

Issued on: 29 June 2026
Response deadline: 6 July 2026 (7 days)
Project kickoff target: 20 July 2026
Contact: mohannad@skybluegroup.com

1. Project Goal

Build a production-ready WhatsApp AI Booking Assistant that:

1. Serves as a single WhatsApp entry point for consumers seeking everyday services.

2. Qualifies leads and collects all required booking details through natural conversation.

3. Connects to multiple booking aggregator APIs (5-6 at launch) and routes booking requests intelligently.

4. Creates and confirms bookings via aggregator APIs.

5. Sends reminders, post-service follow-ups, and collects satisfaction feedback.

6. Captures source attribution data on every inbound conversation.

7. Honors a hard UX rule: the consumer must never be told "we don't have vendors in our system."

2. Scope of Work (Phase 1)

2.1 In Scope

2.2 Explicitly Out of Scope (Phase 1)

· Direct voice / phone calls (no SIP trunk, no voice notes, no call routing)

· Payment gateway integration / consumer transactions

· Multi-language support beyond English

· Native mobile apps (iOS / Android)

· Merchant-facing dashboard

· Advanced analytics / BI tooling

3. Critical Functional Requirements

3.1 "No Empty Shelves" Fallback Cascade

If the primary aggregator for a category has no slot available, the assistant must:

1. Automatically query the next aggregator in priority order (configurable per category).

2. If all aggregators in the category fail, escalate to a human handoff queue (a designated WhatsApp number / inbox we will provide).

3. The consumer must never see a message indicating that the platform lacks vendors. Graceful pivot only — examples:

o "Let me check a few more options for you, one moment…"

o "I'm connecting you with our specialist who has access to additional vendors right now."

The fallback engine must be configurable without code changes (admin can reorder aggregator priority per category via a config table or simple UI).

3.2 Aggregator Routing Engine

· Category → ordered aggregator list (e.g. Salons → [Booksy, Zenoti, Fresha link-builder, Human queue])

· Configurable failure conditions (no slot, API error, timeout, low confidence match)

· Per-aggregator timeout settings

· Audit log of every routing decision

3.3 Conversation Flow

· Natural language understanding for service type, location/area, date, time, and preferences

· Clarifying questions when input is ambiguous

· Step-by-step guided booking with confirmation before submission

· Ability to handle basic post-booking queries (status, reschedule, cancel)

· Human-handoff trigger when AI confidence is low or user requests it

3.4 Notifications

· Reminders: 24 hours before + 2 hours before appointment (configurable)

· Post-service follow-up: 2 hours after appointment end (configurable)

· Satisfaction capture: 1-5 rating + optional text comment

· WhatsApp Business template management workflow (vendor handles template creation + submission)

3.5 Source Attribution

· Every conversation must capture the source parameter from the inbound click-to-chat link (e.g. wa.me/XXX?text=SRC12) or QR code identifier.

· Source persists against the lead, conversation, and resulting booking.

· Admin console must allow filtering/grouping by source.

3.6 Admin Console (Minimum Viable)

Read-only is acceptable for Phase 1. Must show:

· Live conversations in progress

· Bookings created today / week / month

· Failed booking attempts (with reason)

· Per-aggregator success rate

· Per-source attribution counts

· System health indicators (API uptime, message throughput)

4. Non-Functional Requirements

5. Technology Preferences

We have a working hypothesis on the stack but remain open to vendor recommendations:

Please propose your preferred stack with reasoning.

6. Third-Party Costs

The following are paid directly by the client, not by the vendor:

· WhatsApp Business API provider (BSP) subscription + per-message fees

· Cloud hosting

· LLM API usage

· Aggregator API license fees (e.g. Foodics)

· SMS/OTP if required

· Domain + SSL

Vendor must estimate the monthly run-rate of these costs at a baseline pilot volume so we can budget accordingly.

7. Aggregator Integrations

We will provide access (documentation and credentials where applicable) to the following aggregators. Some have open sandboxes; others require partnership credentials that will be secured in parallel:

Vendor must design the aggregator integration layer as a plugin architecture so future aggregators can be added without refactoring core logic.

8. Deliverables

9. Intellectual Property

· All code, prompts, schemas, configurations, and documentation developed under this engagement are assigned in full to the client on final payment.

· Vendor retains no residual rights to reuse the code, prompts, or designs in other projects.

· Open-source dependencies remain under their respective licenses; vendor must provide a license inventory.

10. Timeline

11. Forward Pricing (Required in Proposal)

In addition to the Phase 1 quote, please provide an annex with the following forward pricing:

11.1 Per New Aggregator Integration

Flat rate for adding a new aggregator API after Phase 1 (covers discovery, auth setup, endpoint mapping, testing, handover).

11.2 Per Direct Merchant Integration

Flat rate for integrating an individual merchant where no aggregator API exists.

11.3 Monthly Maintenance Retainer

Quote two tiers:

· Basic — uptime monitoring, incident response, critical bug fixes

· Standard — Basic + minor enhancements (X hours/month, specify) + aggregator API change management + WhatsApp template/policy compliance

11.4 Phase 2 Enhancements

Ballpark range (not firm quote) for the following enhancements, individually itemized:

· Merchant-facing dashboard

· Multi-language (Hindi, Tamil, Arabic)

· Voice channel (phone + voice notes via TDRA-compliant SIP trunk)

· Payment gateway integration

· Advanced analytics / BI

· Scaling work as usage grows

12. Proposal Format

Your proposal must include, in this order:

1. Executive summary (1 page)

2. Company / team profile — relevant experience, team size, location, references

3. Proposed technical approach — architecture diagram, stack choice with reasoning

4. Detailed scope mapping — confirm coverage of each Phase 1 requirement (Sections 2-3)

5. Timeline + milestone breakdown

6. Phase 1 pricing — fixed-price preferred; itemized hourly acceptable

7. Forward pricing annex (per Section 11)

8. Estimated monthly third-party run-rate (per Section 6)

9. Team CVs / portfolios — especially for the lead developer

10. 3 references with contactable phone + email

11. Risks + assumptions

12. Validity period — minimum 30 days

Page limit: 20 pages excluding annexes.
Format: PDF only.

13. Evaluation Criteria

We reserve the right to:

· Request clarifications or revised proposals.

· Shortlist multiple vendors for a paid 1-week prototype "bake-off."

· Reject all proposals without explanation.

· Negotiate scope and price with any vendor.

14. Mandatory Vendor Disclosures

Please disclose in your proposal:

1. Where your team is physically located.

2. Whether you subcontract any work, and if so, to whom.

3. Any current or recent conflicts of interest.

4. Your data handling practices for client information.

5. Your standard contract terms (or willingness to sign ours).

15. Submission Instructions

Submit your proposal via email to mohannad@skybluegroup.com with the subject line:

RFP Response — WhatsApp AI Booking Assistant — [Your Company Name]

For clarifying questions (deadline: 06 July 2026), email the same address with subject:

RFP Question — [Your Company Name]

Questions and anonymized answers will be shared with all participating vendors.

16. Confidentiality

This RFP and all information contained herein is confidential. By accepting this RFP, you agree:

· Not to share its contents with third parties.

· Not to use the information for any purpose other than preparing a proposal in response to this RFP.

· To return or destroy this RFP and any derivative materials if you choose not to bid or are not selected.

Contact for all correspondence: mohannad@skybluegroup.com