Confidential · CZ-2026-0619
Solution Scope & Investment
AI for Travel.
An operational AI service for 360 Edge that runs the booking-lifecycle admin for travel consultants, verifies every output against the source, and ships as a fully managed service.
01
The Challenge
Consultants spend their day on admin, not on clients.
Every booking carries the same manual load. It is slow, easy to get wrong, and it scales badly as volume grows.
01
The email firehose
100+ emails a day across clients, suppliers and marketing, waded through by hand.
02
Double entry
A quote is built in Excel, then re-typed into Word. The same data, keyed twice.
03
Fee hunting
Cancellation, baggage and seat fees live scattered across supplier emails and memory.
04
Open loops
Reconfirmation windows and late documents rely on someone remembering to chase.
02
The Solution
A back-office assistant the whole agency can trust.
It drafts and proposes. A consultant always approves. Five principles keep it cheap, accurate and effortless to run.
01
Cheap by default
A small fast model does the routine work. A larger model is called only for hard reasoning.
02
Human in control
Nothing reaches a client or supplier without a person signing off.
03
Verified first
A second pass checks every output against source before a consultant sees it.
04
One source of truth
Everything the agency knows lives in one searchable place.
05
Fully managed
We host, secure, monitor and update it. The agency simply uses it.
03
How it works
One flow, every task.
Consultant asks in plain language"Draft a quote from these components"
▼
Knowledge Basepolicies · supplier terms · past quotes · fees & conditions
▼
Skills build the outputQuote · Confirmation · Itinerary
▼
Verification layer checks itflights · fees · dates · rooms · consistency
▼
Consultant reviews and approvesthen it goes to the client or supplier
04
What we build
Feature by feature.
Each capability ships on its own and delivers value the moment it lands.
01Phase 1Centralised Knowledge Base
One searchable source of truth for supplier terms, fees, templates and history. Live pricing is pulled at request time, never stale.
02Phase 1AI Interface
A single plain-language way to ask questions and search across every consultant's email and documents. Answers always cite their source.
03Phase 1Email Triage
Classifies the 100+ daily emails, de-dupes repeated supplier promos, and surfaces the specials genuinely worth flagging.
04Phase 1Quote Assembly
Pulls property, room, rate and inclusions into QCostings and generates the QQ quote. Entered once, not twice.
05Phase 2Fees & Conditions
States cancellation, baggage and seat fees per component, and flags when a higher fare with free seating beats a cheaper one.
06Phase 2Confirmations & Itineraries
Populates the KK confirmation and assembles itineraries with consistent formatting. No more copy, reformat, repaginate.
07Phase 3Reconfirmation & Doc-Chasing
Tracks the 30-day window, chases suppliers for late documents, and inserts confirmation numbers into itineraries.
08FoundationHosting, Security & Delivery
Managed cloud, single-tenant, data stays in-region, least-privilege access, automated deployment and full audit logging.
05
The Differentiator
The agent that checks the agent.
Speed is only useful if it is safe. Before any output reaches a consultant, an independent pass verifies it against the real source. Flights against Sabre. Hotels against the supplier. Every fee, and every document, in agreement.
Where it finds a mismatch, it does not silently fix it. It flags the discrepancy, names the source it checked, and routes the decision back to the consultant. Anything it cannot verify is marked unverified, never passed off as confirmed.
06
Delivery
Start with a proof of concept.
You're experimenting, so we start small: a focused PoC on a sample of your data, no big commitment. If it earns its place, we expand into the full phased build.
Step 1 · Now
Proof of Concept
- Knowledge base on sample docs
- Plain-language AI Q&A
- Email triage demo
- ~2 to 3 weeks
Then · Phase 1
Foundation
- Live data + quote assembly
- Integrations mapped
- Production hosting
Phase 2
Accuracy
- Fees & conditions
- Confirmations
- Verification layer
Phase 3
Close the loops
- Reconfirmation
- Doc-chasing
- Full verification
07
Investment
Start small. Prove it. Then scale.
A fixed, low-cost proof of concept to try it on your own data, with no big commitment. You only invest in the full build if the PoC earns it.
| Stage | Covers | Investment |
|---|
Proof of Concept
start here | Working demo on a sample of your data: knowledge base + plain-language AI Q&A + email triage. No live integrations, hosted by us. ~2 to 3 weeks. | $3,500 to 4,000 |
Full build
later, optional | The phased platform: live integrations, quote assembly, fees & conditions, confirmations, the verification layer | from $9,500 |
If it goes live: managed service from $1,500 / month · hosting, run, monitoring, support, model usage
08
Why this works
Try it small. Scale only if it works.
Start with a low-cost proof of concept on your own data. No big commitment, no lock-in. If it proves its value, we expand it into the full build and run the whole thing for you.
100%
outputs verified before a person sees them
0
systems for the agency to maintain
Muhammad Asim · muhammadasim@codeaza.org · codeaza.com
09
