Interview Prep VP Zoom Round Confidential

AI Tool Builder for FP&A — eBay

A warm intro from her boss — casual, relationship-first. Your job isn't to pass; it's to be someone she wants to champion. This guide gives you the fluency and POV to hold a great conversation — held lightly, deployed only when it's natural.

Format: Casual Zoom · warm intro from her boss Real goal: fit & rapport, not a grilling Your edge: grounding & verification Your ace: Req Room (eBay HC demo)
Read this first — it's a warm intro, not an interview

Her boss introduced you, which means you walk in already endorsed. She's partly doing him a favor, partly scouting — and she'll report back to him. So this isn't a test to pass; it's a conversation to click in. The whole guide below is your reserve: know it cold, deploy it lightly. Winning looks like her thinking “I like this person and I'd want them on the team.”

Dial UP

  • Genuine curiosity — ask about her world, her problems, what she's building. Let her talk.
  • Warmth & ease — it's a chat. Laugh, be human, react to what she says.
  • Authentic “why” — why you find trust-in-AI-for-finance genuinely interesting. Conviction, not a pitch.
  • Listening — follow her threads, build on them. The best signal of fit is a real back-and-forth.
  • Two-way framing — you're also deciding if you want it. That confidence reads well.

Dial DOWN

  • Reciting metrics — don't lead with eBay's take rate. Know it; mention it only if it's genuinely relevant.
  • The 90-day plan — do not offer it unprompted. It's over-eager for a casual chat.
  • Sounding rehearsed — no monologues, no frameworks-at-her. Answers stay under ~60 seconds.
  • Selling too hard — the endorsement already did the selling. Trying too hard undoes it.
  • Interviewing at her — it's a dialogue, not a Q&A round.

Conversation Starters — Get Her Talking

In a warm chat, the person who asks the better questions and listens hardest wins. These are tuned to feel human, not like an interview — and to make her do the talking, which is exactly what builds rapport. Pick a few that feel like you; don't run the list.

Open warm (first few minutes)

  • “Thanks so much for making the time — [boss's name] spoke really highly of you and the team. How do you two know each other / how long have you worked together?”
  • “Before we get into any of it — how's your week going? Is this a busy stretch for you, quarter-close kind of thing?”
  • “I'd love to just hear about what you do day-to-day — how would you describe your corner of eBay's finance world?”

Get her talking about the work

  • “What made the team start thinking about an AI-tools role in the first place? Was there a specific pain that kept coming up?”
  • “When you picture this going really well a year from now — what's different for your analysts day-to-day?”
  • “What's the part of the close or forecast cycle that everyone quietly dreads?”
  • “Have you tried any AI tools already? What landed, what didn't?”

Natural bridges to your world (only if she opens the door)

  • She mentions manual/repetitive work → “That's exactly the stuff I like pointing AI at — the trick is doing it so finance can actually trust the output. That's kind of been my whole thing.”
  • She mentions AI accuracy / trust worries → “Yeah — that's the real problem, honestly. I've spent a lot of time on the ‘make sure it isn't confidently wrong' side. Happy to nerd out on it if useful.”
  • She asks what you've built → “The most relevant one — I actually prototyped a headcount-planning agent against an eBay-style req workflow. Was a fun one because the hard part wasn't the AI, it was getting it trustworthy.”

Show you're a person, not a résumé

  • “What do you enjoy most about working there? What'd surprise me about the team?”
  • “How would you describe the culture on your team — heads-down, collaborative, scrappy?”
  • “What's kept you at eBay?” — people love answering this, and it's genuine.
Graceful close

“This was great — honestly made me more interested, not less. I'd love to keep talking. What do you think good next steps look like?”  Warm, confident, and it surfaces whether this becomes a real process — without any pressure. Then follow up with a short thank-you note the same day.

Your Story — In Your Own Words

The five questions a casual chat almost always reaches: who are you, what are you doing now, why finance, why AI, and what don't you know? These are the ones where sounding rehearsed hurts you most. The drafts below are a starting point, not a script. Rewrite them until they sound like you — the boxes are editable and save automatically.

✎ These answers are editable

Click any answer box (across the whole guide) and type your own version — it saves in your browser automatically and survives reloads. A green left-border means it's your version; hit “Reset to suggestion” to bring the draft back. Text in [brackets] is a prompt for your real specifics — replace it.

“So — tell me a bit about yourself.”
Probing: the guaranteed opener. Not a résumé recital — a 60–90s arc that sets up everything else
Sure — [quick origin: I've been building software for ~X years, came up through Y]. For the last stretch I've been working independently, building AI agent systems — [a couple of concrete examples]. What I kept running into is that the interesting part was never getting a model to do something impressive; it was getting it reliable enough that someone would actually depend on it. So I ended up spending most of my time on the unglamorous half — grounding, verification, making sure the thing isn't confidently wrong. [One concrete: e.g. I built a headcount-planning agent where the lift came from verification, not a better model.] And honestly that's what pulls me toward something like this — I want to go deep on one hard problem with a real team, in a domain where reliability actually matters, instead of skimming across a lot of projects. Finance is about the purest version of that.
“So what are you working on these days? And — honestly — why would you leave that to join a team?”
Probing: the commitment question — your biggest unspoken risk. Her quiet worry: is this a consultant browsing for a gig who'll get bored in eight months? Answer it warmly and directly; do not dodge
Right now I'm working independently — building AI agent systems, mostly [what you're actually doing]. I've genuinely enjoyed it and learned an enormous amount. But the honest answer is that independent work makes you broad and shallow: you're constantly context-switching, and you rarely get to go deep enough on one problem to really solve it — or live with it in production long enough to learn what actually breaks. I want the opposite of that. One hard problem, one team, real data, something that compounds instead of ending when the engagement does. [Optional and very human: and frankly, I miss having colleagues.] So this isn't me looking for a gig — it's me wanting to go deep somewhere the work matters.
“So — why finance? You come from more of an AI/builder background.”
Probing: is this a genuine interest or just where a job opened up?
Honestly, it's the trust problem that pulls me in. I've spent my time building AI systems, and the hardest, most interesting version of that is a domain where being wrong actually costs something — where you can't just ship a plausible-looking answer. Finance is that domain in its purest form: a number's either right and traceable, or it's a real problem. I like working where the truth is checkable and it matters. [Add a personal thread — e.g. always liked the rigor of numbers / grew up around a family business / did X that got me into it / I like that finance is the one place AI has to actually be correct, not just impressive.]
“And why AI — what got you into building this stuff?”
Probing: genuine conviction vs. riding the hype
What hooked me wasn't the flashy demos — it was the gap between a demo and something you'd actually trust in production. Getting a model to look smart is easy; getting it to be reliable enough that a real team leans on it is the hard, fun engineering problem, and that's the part I kept gravitating toward. I like building the boring, load-bearing parts — the grounding, the checks — that make the magic actually usable. [Make it yours — e.g. the first agent I built that actually worked / the moment I realized I cared more about reliability than novelty / a specific project that clicked for me.]
“You don't come from a finance background — what do you think would be hardest for you here?”
Probing: self-awareness. A casual chat invites candour — deflecting here reads far worse than a clear-eyed answer
Fair — and I'd rather just be straight about it. I'm not going to walk in knowing your close calendar, your materiality thresholds, or how your category P&Ls are actually structured. That's real, and it's why I think my first month is mostly listening rather than building. What I'd lean on is that I've learned domains by embedding in them before — and the finance concepts themselves aren't really the hard part. The hard part is knowing which details matter to your team, and that only comes from sitting with them through a close. What I'd bring on day one is the other half: knowing how to make an AI system trustworthy enough for a domain like this. I'd much rather be honest that I'll need to learn the finance than oversell it and get something wrong in front of your CFO.
The move that defuses her biggest doubt

Notice your intro ends on the in-house pivot — “I want to go deep on one hard problem with a real team.” That's deliberate. It answers the commitment question before she has to ask it awkwardly, and it pairs with the Toronto-solo answer to close the loop: you're not a flight risk who happens to be unsupervised. Volunteering the answer to a concern reads as self-aware; waiting to be asked reads as evasive.

01 What She's Quietly Assessing

Even casual, she's forming a read — and she'll share it with her boss. Under the friendly surface she's answering one question: “Would I want this person building AI tools alongside my finance team?” The scoring below is what a warm chat still reveals; you just express it through good conversation, not a pitch.

What they're really scoring

  • Business judgment — do you reason in outcomes (faster close, better forecast accuracy) or in tech jargon?
  • Prioritization instinct — where would you point AI first, and why?
  • Trust & governance reflex — do you treat finance's rigor as sacred, or as friction?
  • Executive presence — concise, structured, a clear POV, still coachable.
  • Culture / “row the boat” — will finance analysts want to work with you?

How to show up

  • Lead with the business, close with the tech. Never the reverse.
  • Have a point of view — VPs distrust fence-sitters — but hold it loosely.
  • Answer in 60–90 seconds, then check in: “Want me to go deeper?”
  • Name the risk before they do. In finance, the person who raises the governance question first wins trust.
  • Bring one artifact you can reference live (Req Room).
The one-line frame

“I build AI tools that let a finance team move faster without ever asking them to trust a number they can't trace. In FP&A, trust is the product — and that's the exact problem my work has been about.”

02 Your Thesis — The Wedge Only You Can Claim

Every candidate will say “I can build AI agents.” Almost none will have spent their career on the part that actually blocks AI adoption in finance: making the output trustworthy enough to put in front of a CFO. That's your entire body of work. Lead with it.

The problem statement (say this early)

“The bottleneck for AI in FP&A isn't capability — models can already draft a variance narrative or write the SQL. The bottleneck is trust. A finance team can't ship a number to the board that an AI 'probably' got right. So the hard engineering problem isn't the generation — it's the verification, grounding, and audit trail around it. That's the layer I build.”

Three proof-points from your own work — translate them for finance

What you builtWhat it provesThe FP&A translation
Grounding & verification systems (Quorum, K=3 verify → 27.8%→0% false positives)You can drive an agent's fabrication rate to zero with independent checks“A variance narrative where every number is traced back to a GL line and cross-checked — not a plausible guess.”
Blind-judge / “compute grounding from checks that would fail”You don't trust a model's self-reported confidence — you engineer falsifiable checks“I never let the tool grade its own homework. Confidence comes from a check that would catch it lying.”
Req Room — headcount-planning agent (89.5% vs 47.4% baseline)You've already built an FP&A-adjacent workflow agent — for eBay's own domain“Headcount planning is core FP&A. I built an agent that nearly doubled the baseline on exactly this task.”

03 eBay Business — Know These Cold

You can't build finance tools for a business you can't explain. Memorize the shape of eBay's P&L and the levers FP&A actually forecasts. (FY2025 figures.)

~$80B
GMV (gross merchandise volume) — 51% international
$11.1B
Net revenue (FY2025)
13.94%
Take rate = net revenue ÷ GMV
135M
Active buyers · 2.5B live listings

How eBay makes money (the revenue bridge)

  • Transaction take rate — the core. Fees on GMV. FP&A forecasts GMV × take rate.
  • First-party advertising — the growth engine. Ad revenue as a % of GMV is climbing; this is a key FP&A driver and margin story.
  • Managed payments — eBay owns the payment flow (post-PayPal), adding revenue & data.
  • Financing / shipping programs (e.g., UK shipping ramp) — incremental take-rate levers.

The strategy FP&A supports

  • “Focus” categories — collectibles, motors parts & accessories, luxury (handbags, watches, sneakers), refurbished. Higher-margin, defensible vs. Amazon.
  • Focus vs. horizontal — investing in curated verticals while maintaining the broad marketplace.
  • Advertising penetration — the clearest margin-accretive lever.
  • AI product features — “magical” listing tools, shopping agents — eBay is already an AI-forward consumer company, which is why an internal AI-for-finance role exists.

04 FP&A 101 — Speak the Language

You don't need to be a CPA. You need to know the workflows, artifacts, and pain points well enough that a finance leader nods. Here's the working vocabulary.

The FP&A calendar (the rhythm everything runs on)

CadenceWorkflowWhat it produces
AnnualAOP / Operating Plan & budget; Long-Range Plan (3–5 yr)The plan everything is measured against; targets by org
QuarterlyForecast / reforecast; earnings & board supportUpdated outlook; guidance inputs
MonthlyClose → variance / flux analysis (actuals vs plan vs forecast)“Why did this line move?” commentary; management reporting
ContinuousAd-hoc analysis, scenario modeling, decision supportAnswers to leadership's “what if…” questions

Terms to use naturally

  • Driver-based model — forecast built from operational drivers (GMV, take rate, headcount), not just last year × growth
  • Variance / flux — actual vs plan; the “why”
  • P&L bridge / walk — waterfall explaining a change quarter-over-quarter
  • Opex / capex — operating vs capital spend
  • Contribution margin — revenue minus variable cost
  • Reforecast — updating the outlook mid-period
  • Materiality — is a variance big enough to explain?
  • Accruals — recognizing cost/revenue before cash moves

The stack FP&A lives in

  • ERP (source of truth): Oracle / SAP — the general ledger
  • EPM / planning: Anaplan, Pigment, Oracle Hyperion/EPBCS, Adaptive — where models & budgets live
  • Data warehouse: the analytics layer (often Snowflake/BigQuery-class)
  • The last mile: Excel & slides — still where a shocking amount happens

Know these names so that when the VP mentions “our Anaplan models” you're not blinking.

Where the pain (and your ROI) actually is
  • Manual data wrangling & reconciliation across ERP / EPM / warehouse — hours before analysis even starts.
  • Slow close & flux commentary — analysts hand-writing “why did marketing spend move $2M” every month.
  • Forecast accuracy — and no fast way to run scenarios.
  • Ad-hoc leadership questions — “what's driving take rate in motors?” — that take a day to answer.
  • Model risk — broken links, stale assumptions, no version control, no audit trail.

05 Tools You'd Actually Build — Have 3 Ready

When the VP asks “where would you start?”, don't list ten ideas. Name a prioritization rule, then one flagship tool you'd ship first. Depth beats breadth.

Your prioritization rule (memorize this)

Start where three things overlap: high-frequency high-toil verifiable output.

“I'd deliberately not start with autonomous forecasting — it's high-value but low-verifiability and high-blast-radius. I'd start with monthly variance commentary and self-serve data retrieval: done constantly, universally hated, and every output is checkable against the ledger. Win trust there, then earn the right to touch the forecast.”

Flagship

Variance-Narrative Agent

Given actuals vs plan, auto-drafts the flux commentary — every figure traced to its GL driver and cross-checked before it's shown. Analyst reviews & approves, never re-derives.

Why it's the wedge: highest-toil monthly task, output is 100% verifiable, and it's exactly where hallucination would be caught — so it showcases your grounding work.

Fast follow

NL-to-Insight Data Copilot

“What's driving take rate in Motors this quarter?” → validated SQL over the warehouse → a checked answer with the query shown. Kills the day-long ad-hoc turnaround.

Guardrail: show the SQL and the row-count checks; never a naked number.

Then

Model / Close QA Agent

Anomaly & flux detection on the GL during close; catches broken links, circular refs, formula drift in planning models. (Your R7 bug-sweep pattern, pointed at spreadsheets.)

Later — earn it

Scenario / Forecast Copilot

Driver-based “what-if” generation and sensitivity analysis on top of the existing planning model. Advisory, human-in-loop — deliberately last, because it touches the number the board sees.

06 AI-in-FP&A Market — Show You've Read the Room

A few current data points let you speak to the trend credibly and, more importantly, position eBay within it.

54%
of CFOs make integrating finance AI agents a top 2026 priority (Deloitte Q4'25)
44%
of finance teams already deploying agentic AI in FP&A — up ~600% YoY
14%
have fully integrated agents into finance — the governance gap
40%
of enterprise apps to be agent-integrated by end-2026 (Gartner)

The take that makes you sound senior

“Everyone's deploying — but only ~14% have fully integrated agents into finance, and the gap is governance, not capability. So the winning move for an internal builder isn't chasing the most autonomous tool; it's building the trust and control layer that lets finance actually put these into production. That's a build-not-buy problem, because the moat is eBay's proprietary data and workflows — a Datarails or Pigment add-on can't touch that. I'd buy the commodity planning platform and build where the data is the differentiator.

07 Likely Questions — Strategy + Strong Answers

For each: the strategy (what they're probing) and a say-something-like-this answer. Every answer box here is editable too — rewrite them in your voice and they'll save automatically. Make them yours — don't recite.

“Why this role, and why eBay?”
Probing: genuine motivation & whether you get the business
I've spent my time building AI agents where the hard part is trust — grounding, verification, getting fabrication to zero. FP&A is the purest version of that problem: a domain where a wrong number is unacceptable. eBay is a data-rich marketplace where finance is under real pressure to do more with the same team — take-rate precision, ad-revenue forecasting, category discipline. That's exactly where an internal AI-tools builder pays for themselves. I want to build the tools, not just advise on them.
“Where would you point AI in our FP&A function first?”
Probing: prioritization judgment — the core of the role
I'd start where high-frequency, high-toil, and verifiable-output overlap — monthly variance commentary and self-serve data retrieval — not autonomous forecasting. Those are hated, done constantly, and every output is checkable against the ledger, so I can prove accuracy and build trust before touching the number the board sees. Win the close first, earn the forecast later.
“How do you get finance people to trust an AI tool?”
Probing: your governance reflex — the make-or-break for this role
You never ask them to trust it — you make it show its work. Every number grounds to a source row; there's an independent check that would fail if it were wrong, not a self-reported confidence score; the query and driver are one click away; and it drafts while the analyst approves. Autonomy is earned per-workflow after accuracy is demonstrated, and everything's reproducible for audit. Trust comes from traceability, not from the model being impressive.
“Tell me about something you've built.”
Probing: substance behind the résumé — go concrete
The most relevant one: I prototyped a headcount-planning agent — Req Room — against an eBay-style req workflow. It hit ~89% against a ~47% baseline. The lift didn't come from a better model; it came from decomposing the task and adding verification so the agent couldn't confidently produce a wrong answer. That's the same architecture I'd bring to a variance or forecasting tool here.
“What's the biggest risk of AI in finance? What keeps you up at night?”
Probing: maturity — do you see the downside clearly?
A confident, wrong number landing in a board deck or guidance model — because that's not a bug, it's an eroded-trust event that sets the whole program back a year. The subtle version is silent failure: the tool fabricates in the one field nobody checked. My whole approach is built around that — verify with checks that would fail, keep a human approving, and instrument for the errors you can't see. I'd rather ship a narrower tool I can vouch for than a flashy one I can't.
“How would you measure success?”
Probing: outcome orientation, not activity
Business outcomes first: days off the monthly close, forecast-accuracy delta, analyst hours redeployed from wrangling to analysis. Then adoption — are analysts choosing to use it unprompted — and a hard quality bar: the tool's error rate has to be at or below a human's on the same task, measured, not asserted. If adoption's high but I can't prove accuracy, that's a red flag, not a win.
“Build vs. buy?”
Probing: do you over-engineer, or think like an owner?
Buy the commodity — the planning platform, the ERP. Build where eBay's proprietary data and specific workflows are the moat, which is exactly the AI layer on top. A vendor add-on can't know your category P&L structure or take-rate drivers. So: don't rebuild Anaplan; do build the grounded agents that sit on top of it.
“You're technical — how do you work with a finance team?”
Probing: culture fit & humility
I embed and shadow before I build — sit through a close, watch where the hours actually go. I ship small and let their reaction steer, because they know materiality and the audit constraints better than I ever will. My job is to translate their pain into tools, not to hand them tech and hope. The best signal I'm doing it right is when an analyst asks for the next feature unprompted.
“Where does this go in three years?”
Probing: vision — can you see past v1?
The monthly flux mostly runs itself — the agent drafts, an analyst reviews exceptions. Reforecasts become continuous instead of quarterly fire drills. The analyst's job shifts from producing the number to interrogating it — more business partnering, less spreadsheet plumbing. But every step of that is gated on trust earned in the prior step; I wouldn't skip to autonomy.
“One thing I want to be upfront about — you'd be on your own here in Toronto. I'm in the UK, and the rest of the team is in India. Does that give you any pause?”
Probing: a genuine worry — will you thrive solo across time zones, or quietly struggle in isolation?
Honestly, not at all — I'm genuinely comfortable working on my own. I've worked with overseas and distributed teams before, so the time-zone spread is familiar territory, not something I'd be figuring out for the first time. I actually do some of my best work when I've got the autonomy to just run with something. The way I keep it from feeling isolated: I over-communicate async — clear written updates so nobody's ever guessing what I'm working on — and I protect a bit of overlap with each side, mornings with the India team and a window with you, for the things that are just better live. [Add a specific example — an overseas/distributed team or project you ran, and how you stayed in sync across the time gap.] So genuinely, the setup reads as a plus for me, not a concern.

08 Amazing Questions to Ask Her

In a warm, casual chat your questions are the impression you leave — more than your answers. The ones below are built for this conversation: she builds AI-for-finance herself, Anitha connected you, and it's a get-to-know-you, not a grilling. Pick 3–4 that feel like you, ask them, then shut up and follow the thread — the follow-up (“say more about that”) is where it gets real.

Open her up Get her talking about the real problem

  • “When you picture this role really working, what's the first thing that gets easier for your team — the close, the forecast, or the endless ad-hoc ‘why did this move' questions?” Why it lands: makes her name the actual pain in her own words — and tells you exactly where to aim first.
  • “Where has AI already let you down in finance — something that looked great in a demo and then didn't survive real data?” Why it lands: invites candor, and signals you know the demo-to-production gap is the whole game.
  • “What's your team still doing by hand that they'd be almost embarrassed to still be doing in two years?” Why it lands: surfaces the highest-toil work — your best first target — and it's a genuinely fun one to answer.

How she thinks Reveal her judgment

  • “How do you personally draw the line between ‘AI drafts, a human approves' and letting something run on its own — and where would you never let it cross?” Why it lands: her governance philosophy is the make-or-break of this role, and it mirrors your whole thesis.
  • “If you had to bet — is the bigger unlock better forecasts, or faster and cheaper ones? Where's the pain actually worse?” Why it lands: forces a real opinion, and her answer reframes what you'd build first.
  • “What would a tool have to do for you to trust its number enough to put it in front of the CFO?” Why it lands: hands her your exact wheelhouse and lets her define the bar you'd be building to.

Peer, not rival Since she builds this too

  • “Anitha mentioned you're building in a similar space — I'd genuinely love to hear where you've landed, and where you're still wrestling.” Why it lands: collaborative, not competitive — positions you as a peer comparing notes, never someone one-upping her work.
  • “You're closer to this than almost anyone — what do you think most people get wrong about applying AI to FP&A?” Why it lands: flattering and revealing — her “what people get wrong” is a window into her entire philosophy.

Fit & human Scope — and warmth

  • “Is this a builder embedded inside finance, or a finance-literate person on an AI team? Honestly, the best version of me looks a little different depending which.” Why it lands: shows self-awareness and gets you the single most important scoping fact.
  • “Six months in, what would make you look back and say ‘that hire was clearly the right call'?” Why it lands: her success criteria in her own words — echo them for the rest of the chat.
  • “What's kept you at eBay?” Why it lands: people love answering it, it's genuine, and it ends things on warmth.
The three rules
  • Ask 3–4, not all. A firehose of questions reads as a checklist; a few good ones you actually dig into reads as a conversation.
  • The follow-up is the skill. Her first answer is the setup; “say more about that” is where you learn something real and she feels heard.
  • Avoid: anything Googleable (revenue, org size), comp this early, and yes/no questions that let her off in one word.

09 Your 90-Day Plan — Offer It Unprompted

If there's a natural moment, sketch this. VPs love a candidate who's already mentally started. Keep it to four beats.

PhaseFocusOutput
Days 1–30 · ListenEmbed with FP&A. Sit through a close. Map where the hours go. Learn the stack & the calendar.A ranked pain-map; one workflow chosen (high-freq × high-toil × verifiable).
Days 30–60 · Ship v1Build one grounded, human-in-loop tool — likely variance commentary or data retrieval.A working tool with a real analyst using it; accuracy instrumented from day one.
Days 60–90 · ProveMeasure against the human baseline. Fix trust gaps. Get an unprompted second user.A number: X hours saved / accuracy at-or-below human error; a case for tool #2.
Beyond · ExpandTemplatize the trust architecture so each new tool ships faster than the last.A repeatable pattern, not a pile of one-offs.

10 Story Bank — Map Your Real Work

Have these loaded so you're never reaching. Each is one project → the finance-relevant point it proves.

Lead with this Req Room

An eBay headcount-planning agent, 89.5% vs 47.4% baseline. Proves: you've already built an FP&A-adjacent workflow agent, and the lift came from decomposition + verification, not a bigger model.

Also live FP&A console

A working FP&A console with a verification/checks layer built in [confirm what it best demonstrates before you cite it]. Proves: you've not only theorized about grounded finance tooling — you've shipped a version of it. Pairs naturally with Req Room as “here's the thinking, here's it running.”

Quorum

K=3 verification drove false positives 27.8% → 0%. Proves: you can engineer trust to the zero-error bar finance requires.

Grounding & blind-judge philosophy

“Confidence from checks that would fail”; “fabrication lands in the unchecked field.” Proves: a hard-won, specific POV on why AI outputs fail silently — and how to catch it.

Production agent systems (FieldAgent, Aegis)

Real, deployed, measured systems. Proves: you ship and operate, not just prototype.


Framing rule for every story: state the business outcome first (“a headcount plan the team could trust”), then the metric, then — only if asked — the how. The VP cares about the first two.

11 Practicalities — Don't Get Caught Flat

Casual chats drift into logistics. None of this should be volunteered — but you want a calm answer ready if she raises it.

⚡ Demo readiness — do this before the call

  • Actually load the link. Confirm Req Room is up and working. Your ace is worthless if it 404s while you're sharing your screen.
  • Have it open in a tab — not something you go hunting for mid-call.
  • Have the 20-second verbal version ready. In a casual chat, describing it well usually beats derailing the conversation into a screen-share.
  • Don't force it. Only offer if she opens the door. If she asks: “I can show you properly sometime — or honestly, easiest is I just send you the link after this?” Sending after is often the better play — it keeps the rapport and gives her something to forward to her boss.

Bonus: a link she can forward is how you get advocated for in rooms you're not in.

💰 If comp comes up

In a warm intro this is almost always informational, not a negotiation — she's checking you're in the right universe. Don't anchor yourself low, and try not to name the first number.

The graceful turn-around:

I don't have a hard number in mind yet — honestly I'd want to understand the scope and level first. What range is the role budgeted at?

If pressed for a number, give a range, tie it to scope, and note you'd want to understand the whole package. Never a single precise figure.

📋 Logistics worth knowing / asking

  • Which entity employs you? Toronto-based, boss in the UK — is this eBay Canada, FTE vs. contractor? Genuinely worth clarifying.
  • CAD or USD, and is comp benchmarked to Toronto or to a US band?
  • Level / title — and is this net-new headcount or backfill?
  • Remote or hybrid — is there a Toronto office expectation, or fully remote given the team's spread?
  • Start timing — and how you'd wind down current commitments cleanly.

Ask at most one or two of these, late, and lightly. This is a first chat, not an offer call.

12 One-Screen Cheat Sheet

Glance-able during the call. Keep this section open in a tab.

Numbers

eBay FY25: ~$80B GMV · $11.1B net rev · 13.94% take rate · 135M active buyers · 51% GMV intl · ad revenue = the growth/margin lever.
Market: 54% CFOs → agents a top-2026 priority · 44% teams on agentic AI (+600% YoY) · only 14% fully integrated (governance gap).

Your three lines

  • Thesis: “In FP&A, trust is the product — and grounding/verification is exactly what my work has been about.”
  • Prioritization: “Start where high-frequency × high-toil × verifiable overlap — variance commentary & data retrieval, not autonomous forecasting.”
  • Ace: “I already built an eBay headcount-planning agent — Req Room, ~89% vs ~47% baseline.”

Don't

  • Don't lead with tech. Don't ramble past 90 seconds. Don't oversell autonomy.
  • Don't treat governance/audit as friction — treat it as the design constraint that makes you credible.
  • Don't skip your questions. This round is won on them.