Interview Prototype · 2-day buildEdition 01

Wego Earnings Digest

Quarterly competitive intelligence — built from public earnings calls, filtered through Wego's strategic lens.

Why this exists

Built over two days as part of preparing for my interview with Wego's Strategy & Planning team. Edition 01 is a working proof of concept — one company, the full nine-lens framework — to show how I'd approach the role's actual output before walking into the conversation.

Edition 01 · 2026 Q1 · Booking Holdings · Prepared May 2026

Read the brief (PDF)View source on GitHub

Built by Subhankar Shukla. Independent project — not affiliated with Wego.

The problem

Earnings calls hold the signal. Reading all five every quarter doesn’t scale.

Every quarter, Wego's closest comparable competitors host earnings calls. Inside those four-to-six hours of transcript is real strategic signal — pricing posture, AI distribution moves, capital allocation conviction, MENA disclosures. Most of it is irrelevant. Some of it is wedge-able into a Wego decision.

The job of a strategy team isn't reading transcripts. It's making decisions from them. So the question this project tries to answer is simple: can the reading be done by a system, in a way that’s honest, auditable, and tuned to Wego specifically — leaving the strategy team to do the strategy?

How it works

A public transcript becomes a Wego brief in five steps.

The system is small on purpose. Two of the five steps are human; the other three are AI doing exactly what humans tell it to do, in a shape that can be checked against the source.

  1. 01
    Input

    A public earnings call goes in.

    Every quarter, public companies host an earnings call where the CEO and CFO walk through results, talk about pressures on the business, and answer analyst questions. The transcript is published the next day. That transcript — verbatim, public, anyone can pull it — is the only input to the system. Nothing scraped, nothing leaked, nothing private.

  2. 02
    Human

    The lens is defined by hand.

    Before any AI runs, a human decides what “relevant for Wego” actually means. That decision becomes the nine-lens framework — MENA, India, SEA, AI distribution, Capital, Marketing, Partnership, Look-to-book, Forward guidance. The lens is the strategic work. Everything downstream is the AI executing against it.

  3. 03
    AI

    The AI reads the transcript with the lens in hand.

    The system feeds the transcript and the lens into a language model and asks it to do one thing: pull every passage in the call that touches one of the nine lenses. Think of it as an analyst with a highlighter — except the analyst only highlights what matters to Wego, and never gets tired. The output is structured: each passage tagged with the lens it hit and the speaker who said it.

  4. 04
    AI

    Each passage becomes a strategic bullet.

    The AI takes those highlighted passages and writes one strategic bullet for each. Two parts: a verbatim quote from the executive, and a one-line implication for Wego. The model is instructed to never write an implication that isn’t anchored to a real quote. If the quote is missing, the bullet doesn’t get written.

  5. 05
    Output

    The bullets are stitched into a brief.

    A final pass assembles the per-company bullets into a single digest, generates a five-line executive summary on top, and outputs a clean markdown file. From there it’s rendered to PDF and to a web reader — same content, two surfaces.

Step 02 in detail · The lens

Nine dimensions the AI is told to look for.

The lens is the only place the human does strategic work. It tells the model what counts as signal and what counts as noise. Everything downstream — extraction, synthesis, composition — runs against this list. Change the list, and a different system pops out the other side.

  • 01

    MENA

    Middle East exposure, conflict drag, regional growth signal.

  • 02

    India

    Outbound corridor, MMYT comparables, GCC↔India travel flow.

  • 03

    SEA

    Southeast Asia distribution, Traveloka and Agoda comparables.

  • 04

    AI / Distribution

    LLM partnerships, agent-layer threats, Claude and OpenAI defaults.

  • 05

    Capital

    Buybacks, guidance shifts, balance-sheet posture into disruption.

  • 06

    Marketing / SEO

    Performance marketing, organic decay, AI Overviews impact.

  • 07

    Partnership

    Bedbank, supply, distribution deals, network density.

  • 08

    Look-to-book

    Funnel conversion, direct vs metasearch economics.

  • 09

    Forward guidance

    Quarter-ahead disclosure, what comps are pricing in.

See it in action

A short walkthrough of the live system.

~75 seconds, no audio. The video below tours the live reader — front page, methodology, executive summary, per-company briefs, and back. The on-screen type is small at full zoom; the guide beneath the video tells you what to look for at each step.

Walkthrough · Edition 01 · Click play, or use the guide below.
What you’re seeing
  1. 01

    Front page · the masthead.

    The reader opens on the dark hero — “Quarterly competitive intelligence, written for Wego Strategy.” One-line subtitle explains what each edition does and how hallucinations are auto-rewritten before publish. Sets the tone before any content.

  2. 02

    Methodology · how the system works.

    Click into Methodology to see the plain-English explanation of the pipeline — input, lens, extract, synthesise, source discipline. Designed to be readable by anyone, not just an analyst.

  3. 03

    Executive Summary · the five takeaways.

    Back on the home view: the boxed Executive Summary card shows the five highest-signal takeaways from the edition. Each one names a competitor disclosure and the specific Wego implication.

  4. 04

    Per-company briefs · the source evidence.

    Click each company tag at the bottom — BKNG, EXPE, MMYT, TRIP — to drill into the per-company brief. Each bullet pairs a verbatim CEO or CFO quote with a one-line strategic implication for Wego.

  5. 05

    Back to home · the loop closes.

    Return to the home page. The same five takeaways now read with the source evidence behind them. That’s the loop — competitor signal, Wego-specific filter, auditable output, repeat next quarter.

Why it doesn’t make things up

Four design choices that keep the output honest.

The biggest risk with any AI tool that writes about a business is that it sounds confident while saying things the source never said. The system avoids that by design — not by hoping the model behaves.

  1. 01

    Every claim is anchored to a verbatim quote.

    The bullet structure forces it: every entry on the page has a quote block from the actual earnings call. If the AI can’t find a quote to back a claim, it doesn’t write the claim. There is no paraphrase mode and no “model speculation” mode.

  2. 02

    The strategic logic is in the prompt, not the model.

    What makes the digest Wego-specific isn’t the AI — it’s the human-written lens and the synthesis instructions sitting on top. Swap the lens, and the same machinery produces a digest for a different company or industry.

  3. 03

    Anyone can audit the output against the source.

    Each bullet names the speaker and quotes them directly. A reader who wants to verify a claim opens the original transcript, searches for the quote, and reads the surrounding paragraph. Nothing is hidden behind a black box.

  4. 04

    The pipeline is reproducible from any laptop.

    No proprietary infrastructure. Public transcript in, markdown digest out. Roughly five minutes of compute and under a dollar in API cost per edition. If the lens needs to change, edit a single prompt file and re-run.

What it produced

Edition 01 · Booking Holdings · Q1 2026.

These are the five executive-summary takeaways the system landed on after one run against Booking's April 28 earnings call. Every line traces to a verbatim CEO or CFO quote in the full PDF.

  1. 01

    MENA is now publicly priced into global travel earnings.

    Booking quantified a 2pp Q1 drag and 3pp Q2 forecast drag from MENA conflict — the first granular disclosure from a major. Wego's 65% MENA mix is no longer narrated in a vacuum.

  2. 02

    Booking's MENA exposure is 4–7%.

    That's the new comp anyone evaluating Wego will use — open question whether Wego wants to frame its own MENA share as concentration rather than exposure in the next investor or partner conversation.

  3. 03

    The AI-distribution stack is going public.

    Booking named four LLM partners (OpenAI, Google, Anthropic, Amazon) and flagged its Claude Live flagship placement on the call. The window to be Anthropic's regional travel partner — MENA + India outbound — likely closes in 6–12 months.

  4. 04

    Direct-booking is the moat against horizontal AI agents.

    Booking is explicit: own the conversion or lose it. Metasearch-only Wego is structurally more exposed; defensibility shifts onto MENA brand and GCC user trust, not funnel ownership.

  5. 05

    Capital conviction at the trough.

    Record $3.6B buyback into a guidance cut. The board signal: lean into MENA share now, not away from it.

Read the full brief (PDF)Browse the source

Coverage

Edition 01 scope, and what comes next.

CompanyTickerQuarterStatus
Booking HoldingsBKNG2026 Q1Synthesized
Expedia GroupEXPE2025 Q4Planned — Edition 02
Trip.com GroupTCOM2025 Q4Planned — Edition 02
MakeMyTripMMYT2026 Q3Planned — Edition 02
TripadvisorTRIP2025 Q4Planned — Edition 02

Edition 01 is a proof of concept. Edition 02 broadens the panel to the full five-company set when remaining Q1 2026 calls land in May–June 2026.

About

Subhankar Shukla.

Third-year Rotman Commerce student at the University of Toronto, CFA Level 2 candidate. Builds AI-powered finance tooling on the side. Most recent production project before this: AXIOM — an institutional valuation platform with multi-stage DCF, comps, RAG over SEC EDGAR 10-Ks, and AI-generated investment theses.

Open to extending the framework — partnership ICP screening, M&A target landscape, market entry analysis. Reach out anytime.