Depilot.
Depilot · AI advisory

The pilot was the easy part.

Getting an AI system from a promising demo into production, where it touches revenue and real customers and can no longer fail quietly, is a different discipline. Jay Sharma has practiced that discipline twice: for 300 million people at Indeed, and alone, as a founder with paying customers. Depilot makes it available to companies that are done experimenting.

Thirty minutes, free. Most people leave with a shorter list than they came with.
INDEED · AMAZON · MICROSOFT · PAYPAL · AGODA · BLOOMBERG
EXHIBIT 01

Pilot purgatory

95%
of enterprise generative-AI pilots show no measurable return, on $30–40B of spending. MIT, State of AI in Business, 2026.

Somewhere in your company is a pilot that impressed everyone in March and hasn't shipped since. It performs beautifully in a sandbox. No customer has ever touched it. Nobody will kill it, because it works, more or less. Nobody will ship it, because no one can say what happens when it is wrong at volume.

The pattern behind MIT's number is remarkably consistent, and it has almost nothing to do with which model anyone chose. Four habits, over and over:

The assistant that can't act. It chats. It cannot read the CRM, write to the ticketing system, or touch the order flow. Impressive in a meeting; useless on a Tuesday.

Launches on faith. No evaluation harness, no defined bar a change must clear, no way to know whether last month's update made things better or quietly worse.

One bot for everything. The general-purpose company assistant is almost always the wrong shape. What survives contact with production: narrow agents, specific workflows, a human somewhere in the loop.

Ownership by committee. AI lives with an innovation team, three vendors, and a steering group. Pilots multiply. Nothing ships.

What breaks the stall is not another pilot. It is judgment: someone who has shipped this before, looking at what you have and putting one of three words on it, in writing.

Kill

Some pilots cannot get there.

Wrong abstraction, no data path, no owner. Ending them early is the cheapest decision in all of AI. We put it in writing and say it to your face, in week two rather than quarter three.

Fix

Most are closer than they look.

The common verdict. A stalled pilot usually sits three or four specific blockers from shipping: evaluation, integration, governance. The report names yours and puts them in order.

Scale

A few are ready now.

Occasionally the foundations are sound and the hesitation is instinct rather than evidence. Then the job is speed: launch gates, guardrails, a kill switch, so the system stays trustworthy at ten times the traffic.

EXHIBIT 02

The 60-second pre-audit

Six questions from the first ten minutes of a real audit. Answer honestly; nobody is watching, and the verdict updates as you go.

Is any AI system you've built handling real work in production today — not a sandbox, not a beta?

Can your AI read from and write to your systems of record (CRM, ERP, ticketing) — or does it only chat?

Could anyone on your team prove, with data, whether your AI got better or worse last month?

Is there a defined bar an AI change must clear before it ships — and a kill switch if it misbehaves after?

Does a single senior executive own AI outcomes — not a committee, not "innovation"?

If your AI vendor tripled prices tomorrow, could you swap models without rebuilding?

Provisional verdict

Awaiting answers…

Six questions, no email required. The stamp appears when you finish.

EXHIBIT 03

What we do

Fixed scope and fixed fee, agreed before work starts. Every engagement is delivered by the person whose name is on the door.

01

Pilot-to-Production Audit

Two to three weeks inside your pilots: architecture, data paths, evaluation coverage. Each pilot receives a written verdict with the reasoning shown, plus a readiness scorecard and a 90-day sequence for the survivors.

Start here2–3 weeksFixed fee
02

Evaluation & Launch Governance

The flagship. An evaluation harness, defined launch gates, behavioral guardrails, and a kill switch, installed and running, with your team trained to operate them. Modeled on the machinery built at Indeed to decide what ships to 300 million people.

6–8 weeksFixed fee
03

Agent Architecture Sprint

For product companies putting agents into the product itself. Agent design, human-in-the-loop mechanics, build-or-buy decisions, and an architecture record your engineers can execute without us in the room.

3–4 weeksFixed fee
04

Fractional CPTO / Chief AI Officer

A senior operator inside your leadership team two or three days a week: roadmap, hiring bar, org design, architecture calls, board reporting. Two seats exist. The six-month minimum is deliberate; nothing real happens faster.

2–3 days/week6-month minimum
05

Board Advisory & AI Diligence

AI oversight for boards, and per-deal diligence for investors who need to know whether a target's AI claims survive a look at the actual architecture. They often don't, and it is better to learn that before wiring the money.

Retaineror per-deal
EXHIBIT 04

The receipts

Twenty-five years, condensed to what shipped. Context on any of these, gladly, on a call.

300M+

people served by the AI matching and launch-governance systems built at Indeed. The governance machinery still decides what ships there.

90%

reduction in bad matches after LLM + RLHF matching went live. Recommendation conversion rose by half.

$5B

in annual revenue carried through a hard regulatory deadline at Amazon Canada, on a catalog-wide ML compliance program.

2×

agentic AI shipped to production. Once with a 300-person org behind it, once entirely alone. Both have paying customers.

2000–09

Bloomberg

Software engineer to engineering manager. Nine years building systems where being wrong costs money.

2008–12

KodeSphere (co-founder)

Took Proofhub, a SaaS product, from nothing to $2M in annual revenue. Product, engineering, and sales in one seat.

2012–20

PayPal · Microsoft · Agoda

Fraud detection at PayPal's transaction volume. Tripled engagement on Bing's knowledge graph. Ran payments for one of Asia's largest travel marketplaces: 90+ local payment methods, cross-border rails in five countries.

2020–23

Amazon, CTO scope, Canada Marketplace

Founded the technology org and grew it past a hundred people. ML compliance across millions of listings. Launched BNPL at marketplace scale.

2023–25

Indeed, Sr. Director, Search & Recommendations

LLM + RLHF matching for 300 million people, and the governance system that still gates every AI release. Also switched off a channel earning $24M a year because it was quietly damaging the marketplace. The replacement performed better on every measure that mattered.

2025–

One More Million (founder) · Depilot

A live agentic fintech product with paying customers, designed, written, and operated end to end with AI agents, by one person. Depilot is the practice built from what both scales taught.

EXHIBIT 05

The Launch Gate method

Gate 01 — Triage

Thirty minutes, free.

Bring the stuck pilot or the roadmap. You'll get a plain reading of where you stand and whether an audit is worth your money. Some callers need one. Plenty don't, and we say so.

Gate 02 — Prove

The audit.

Two to three weeks of evidence: architecture, data paths, evaluation coverage, governance. Every pilot leaves stamped, with reasoning you could hand to your board unedited.

Gate 03 — Ship & Govern

The install.

Harness, gates, guardrails, kill switch, running in your stack and operated by your team. Where it fits, a fractional seat follows, until you no longer need one.

Terms of engagement

You work with Jay. There is no bench and no handoff after the sale. This limits how many clients Depilot can take, which is the point.

Engagements end with something running. A harness, a gate, a system your team operates. Documents describe the work; they aren't the work.

Prices are fixed and quoted up front. Nobody here bills you for thinking slowly.

Bad news arrives early. If the right answer is to kill the project, you hear it in the first two weeks, while it's still cheap.

The practice runs on AI agents. Research, drafting, analysis, with the judgment kept human. Ask to see how it works; it doubles as a demo.

EXHIBIT 06

The operator

Personnel fileDP-001
Jay Sharma
Founder & Principal
CareerBloomberg → KodeSphere → PayPal → Microsoft → Agoda → Amazon → Indeed → founder
ScopeCTO-scope exec · 100+ person orgs · product + engineering in one seat
DomainsFintech & payments · marketplaces · HR tech · proptech
EducationMBA, Michigan Ross · BE Computer Science
StatusAdvisory practice · 2 fractional seats · remote, US hours

Jay Sharma wrote his first production code at Bloomberg in 2000 and spent the next twenty-five years as the person difficult systems got handed to. He founded Amazon Canada's technology organization, grew it past a hundred people, and carried five billion dollars of revenue through a regulatory deadline that had no extension in it.

At Indeed he ran search and recommendations for 300 million people. His teams shipped the LLM matching that cut bad matches by ninety percent, and built the evaluation and launch-governance machinery that still decides what ships there. He also switched off a channel earning twenty-four million dollars a year because it was quietly damaging the marketplace. The replacement made better matches and more money. That decision is this firm in miniature.

In 2025 he left to test a private conviction: that one operator with disciplined AI agents could build what used to take a team. One More Million, a live fintech product with paying customers, is the result. Depilot is the practice built on what both scales taught him.

"Most AI advice fails in the gap between a demo that impresses and a system you'd bet revenue on. I've crossed that gap twice. The crossing is the practice."

EXHIBIT 07

Fair questions

Who is this for?
Mid-market and PE-backed companies with stalled pilots, and Series B–D product companies building agents into what they sell. If you're pre-revenue, or shopping for a team to build an app, we're the wrong call, and happy to tell you who the right one is.
Do you actually build, or just advise?
The audits and sprints produce work your team executes. The governance install ends with a running system. A fractional seat means Jay operating inside your organization. Development at volume is the one thing we don't sell; we'll help you pick and manage that partner instead.
How is this different from a big firm?
A large firm sends a partner to sell you and a team to learn on you. Here, the person who shipped this at 300-million-user scale does the work himself, at a fixed price, and will tell you to kill things. That last part is rarer than it should be.
What does it cost?
Audits start in the low five figures. Installs and sprints are quoted per engagement. Fractional seats are a monthly retainer with a six-month minimum. Every number lands before the work starts.
Where do you work?
Remote, with standing US East Coast hours. Recent work spans the US, Canada, and Asia-Pacific.

Bring the stuck pilot.

Thirty minutes with someone who has shipped what you're attempting. You'll leave knowing why it stalled and what shipping would take. That holds whether or not we ever send you an invoice.

Book the triage call →
or write directly: jay@depilot.ai
LinkedIn: /in/speakwithjay