AI-Wiki

This Is The Next Industry AI Will Disrupt

12 min read

Vitucci — This Is The Next Industry AI Will Disrupt (YC Root Access / The Breakdown)

AI is already transforming entire professions like software engineering and law. And accounting might be next.

In this episode of The Breakdown, YC’s Tom Blomfield and David Lieb sat down with Onshore founder Dominic Vitucci to find out just how AI is fundamentally changing one of the world’s oldest professions and what that could mean for the future of white collar work.

— channel description, YC Root Access

TL;DR

A ~33:56 YC Root Access The Breakdown episode (published 2026-03-07). Interviewers: Tom Blomfield (YC managing director, ex-Monzo / GoCardless founder) and David Lieb (YC partner, ex-Bump / Google Photos co-founder). Founder: Dominic Vitucci, founder/CEO of Onshore (YC W23) — an AI-driven corporate accounting and tax automation platform.

The substantive contributions are six.

1. The professional-services-disruption thesis. “I believe really saw what I was pitching there, but was confined by his circumstance” — Vitucci frames the Big 4 / top-20 accounting industry as a structurally AI-blocked industry, regardless of intent. The thesis (~22:02–24:25): the accounting industry’s revenue will go up in 10 years; headcount will stay roughly flat; revenue-per-employee will go through the roof. The shape of the surviving firm: small senior-partner layer for sales + a regulatory-compliance/expertise layer + some software-engineering — “I don’t think you’ll have the very fat bottom layer because I think you’ll have AI agents doing it. And I think the people at the top are mostly there for sales.” Tom Blomfield extends the thesis past law and accounting: “the contention that software programmers will no longer write code sounded crazy nine months ago and now is like orthodoxy almost. And my contention is the same thing is going to happen to accounting and law and audit… by the end of this year, I can see mass realization that the AI can do these knowledge work jobs just as well or better than humans, and yet still not that big a change in employment rate — that might be five or six or seven years to come.”

2. Why the incumbent Big 4 will not internally disrupt themselves (~7:58–11:14). Vitucci’s two reasons:

  • Value-perception inversion. “It is a fundamental shift in what they perceptionally drive as value. Right now they bill for hours of expertise, and as soon as the hour becomes a diminished unit of value, you no longer really have a whole lot of ground to stand on when you say hey certainly pay me $1000, $1500, $2000 an hour.”
  • Asymmetric senior-partner incentives. “You have to then as a senior partner — a guy who maybe has three to five years left in his career, has to plow a ton of money into something I will never reap the benefits from. And so you have this real asymmetric value prop. You got some 40-year-old partner says hey let’s use AI; some older guys 63, 64 years old say I’m going to retire in 24 months, I don’t care about this — I’m not going to see the value here in 5, 6, 10, 15 years. And so quickly the guys who make the decisions, the senior partnerships, say well what if we just don’t do that?”

Vitucci’s own observation from inside Grant Thornton: “the investment was purchasing $3 to $4 million of co-pilot licenses so that everyone could have co-pilot — what do you guys do with that? Nothing. It’s horrible. It doesn’t work.” The pattern: announcing Deloitte invests a billion dollars in AI as PR while the actual deployment is licence purchases with no engineering org behind them. The CFO question Tom Blomfield extracted from the AI-leadership at a Big 4 firm: “how many software engineers do they have working with them?” — answer: zero.

3. The two-year sell-to-the-firm failure, then the pivot to compete-with-the-firm (~11:14–17:15). Vitucci’s company in summer 2020 → summer 2022 sold AI tooling to accounting firms; got to “eight customers, mid-six-figures revenue” selling to top 20 / top 50 firms; “for two years I just thought… hey tech is better, AI is better, this is more interesting, and I would get push-back constantly. No no, my pivot table is better, this is an edge case, change your tech to fit this mold.” The reason: same misaligned-incentive trap from the inside — junior staff at the firm hear ‘automate 80% of the work you’re doing’ and think ‘I’m going to get fired’. Summer 2022: fired all the customers. Re-built the software for corporate taxpayers directly. Started cold-DMing 1000 people on LinkedIn. “Hey my name’s Dominic, I can do your R&D tax credit better because I use AI.” Closed a Miami customer; got into YC on the flight there.

4. The lawyer-vs-accountant adoption-asymmetry (~13:37–15:55). Why did Lora and Harvey crack the legal vertical but the analogue hasn’t cracked accounting?

  • Liability premium. Lawyers’ mistakes carry jail/large-fine consequences; accountants’ usually only inefficiency consequences. The perceived legal time premium is still durable.
  • Industry forward-thinking-ness. “Lawyers are forward thinking — that might be the hottest take.”
  • Project-based billing maturity in law. Law has been moving to project-based billing for a decade; accounting still resists despite some project-pricing. “I’ll bet you if you divide the project by the number of hours that it takes that it gets real similar to their hourly rate.”

Tom Blomfield’s analysis: “Once you are billing by project, you have price competition and then people have a lot more incentive to reduce their cost.”

5. The Mark co-founder credibility argument (~18:33–21:14). Vitucci’s co-founder Mark was “my boss’s boss’s boss at Grant Thornton”, a senior partner near retirement. “Senior partner Grant Thornton comes over and co-signs this innovation, becomes somewhat more palatable than like junior guy that started doing this — and a wealth of expertise and knowledge.” Mark “passed away in the summer of 2024, about a year and a half ago, right on the precipice as we were raising our Series A.” Vitucci’s reflection on whether the credibility was load-bearing: “It’s possible we could have gotten it done — Lora is a great example. I think it would have been appreciably more difficult.”

6. The unit economics (~25:35–26:37). Onshore: $25M revenue today, targeting $100M by end of 2026. 60–100 employees end of year. “$1M of revenue per employee at least… my estimation [for Grant Thornton / Deloitte / KPMG] would be probably $100,000 maybe $150,000 — you’re an order of magnitude better.” Why the incumbent comp will not improve: “they will only get worse, especially now that we see instead of investing into technology, we see a lot of these big companies investing into overseas operations.”

Closing speculation on the next Excel. When asked “if you were going to start a startup today what idea would you choose?” (~28:36): “a better or different Microsoft Excel. I think spreadsheets and Excel in its original iteration as a spreadsheet functionality is great. I think that it has been kind of just like squished and torn and fit into a million different boxes for things it was never meant to do.” On why customers don’t just use Replit / vibe-code an internal tool: “if you ask somebody hey why don’t you vibe code a Replit app, they would ask what is Replit and what is vibe coding… the access and the democratization of that knowledge is less permeating society than we all might think it is.”

Caveats. YC channel marketing; Vitucci is mid-Series-A raise at filming (Mark’s death is named as “right on the precipice as we were raising our Series A”); every numerical claim is unaudited founder self-report. The “$1M revenue per employee at $100M ARR by end of 2026” is a stretch target — at $25M today and 60-100 employees end of year, the actual end-of-2026 trajectory depends on customer-acquisition rate Vitucci doesn’t disclose. The Big-4-will-stay-flat-on-headcount-while-revenue-rises prediction is a 10-year forecast that the historical record on professional-services-industry forecasting suggests we should hold loosely.

Why this matters in the corpus

The Vitucci ingest is the wiki’s first deep professional-services-AI-disruption anchor at industry-altitude (not company-altitude). The conversation lands four claims that no prior source articulates with equal directness:

  1. The structural-incentive-not-intent argument for incumbent blocking — Big 4 firms aren’t resisting AI on the merits; their senior-partner economics make AI investment a personal-financial-negative regardless of the business-merit case. This is the incumbent-defection-by-structural-incentive mechanism that complements Campfire’s AI-native safety inversion (which is the buyer-side mechanism that gives AI-native vendors air cover).
  2. The “you must compete with, not sell to, professional-services firms” doctrine — same conclusion Glasgow reached independently for ERP; Vitucci reaches it for accounting after two years of failed sell-to-firm attempts.
  3. The lawyer-vs-accountant adoption-asymmetry — useful for predicting which professional-services vertical-AI-startups will work and which won’t: liability premium + project-billing maturity are the two variables that predict adoption velocity.
  4. Revenue-per-employee as the killer comp — Vitucci’s order-of-magnitude framing ($1M+ per employee at Onshore vs ~$100-150k at Big 4 / top 20) is the most-specific articulation in the corpus of the AI-native-vendor-vs-incumbent unit-economics gap.

The boil-the-ocean tension with Tan-CS153 is captured as a typed contradicts relationship — not because either source is wrong, but because the wiki should record that the productive 2026 founder doctrine carries opposite directional advice depending on stage: at the engineering-capability level boil the ocean is the inversion of legacy don’t boil the ocean business advice (Tan); at the wedge-into-an-industry-go-to-market level be great at one thing really early and don’t try to do everything (Vitucci). Both are right at different altitudes.

What was actually ingested

The full ~33:56 transcript was read end-to-end. Vitucci is a high-substance founder — every chapter contains substantive content. The transcript is auto-generated, ASR-cleaned, and renders almost everything cleanly; the most consequential ASR slips are “co-pilot” (probably Copilot, GitHub) and “Grunt Thornton” (Grant Thornton).

Linked entities and concepts

Entities promoted by this source:

  • Y Combinator — channel; already entity. Bumps source-count.

Dangling — single-source mention, deferred:

  • Onshore — YC W23; AI corporate accounting + tax; $25M revenue today, $100M target end of 2026. First wiki mention; promote on second source.
  • Dominic Vitucci — founder/CEO of Onshore; CS + accounting at University of Illinois; ex-Grant Thornton. First wiki mention.
  • Mark (surname not stated; Vitucci’s co-founder; senior partner at Grant Thornton) — passed away summer 2024. First wiki mention.
  • Tom Blomfield — already on the Y Combinator entity-page Dangling list from Garg 2026’s reference to “Pete and Tom and Gary”; this is the second substantive-source mention by full name (Tom Blomfield) — promotable to an entity page in a follow-up ingest. Vitucci-channel-description names him as ex-Monzo / GoCardless founder.
  • David Lieb — YC partner; ex-Bump / Google Photos co-founder. First wiki mention.
  • Grant Thornton — Vitucci’s pre-YC employer; top-20 US accounting firm. First wiki mention.
  • Lora — legal AI startup cited by Tom Blomfield as a Harvey-equivalent on a tear. First wiki mention; second-source promotion candidate as the legal-AI cluster grows.
  • Harvey — legal AI startup cited as the canonical lawyer-vertical-AI-success-story. First wiki mention.

Concept pages touched:

  • ai-employment-effects — major update opportunity: Vitucci’s Big-4 industry revenue up, headcount roughly flat, revenue-per-employee 10× prediction is the wiki’s most-specific articulation of the substitution-without-elimination pattern at industry-altitude. Tom Blomfield’s “five to six to seven years to come” employment-rate-impact-lag prediction is the wiki’s first explicit founder-vantage on the capability-recognition-precedes-employment-effect time-lag.
  • automation-vs-augmentation — the Big 4 firm of the future shape (small senior partners / regulatory-expertise layer / software engineers + agents doing the work) is the wiki’s first articulation of augmentation at the senior-tier persisting while automation collapses the junior tier.
  • enterprise-ai-adoption — adds the sell-to-firm fails because of misaligned senior-partner incentives + compete-with-firm succeeds mechanism, with Vitucci’s two-year worked example as the empirical anchor.
  • ai-deskilling / durable-skills — Lora/Harvey crack legal because lawyers are forward-thinking + already moved to project-based billing gives the wiki two variables (industry forward-thinking-ness + billing-model maturity) that predict adoption velocity. The “durable skill” claim flips here — the durable skill in accounting is not expertise (which the AI can match or exceed at junior-mid level today) but sales, regulatory-compliance positioning, and software engineering — the three layers Vitucci names as remaining in his future-accounting-firm.

Source quality

  • Channel: YC Root Access — The Breakdown series; analytical conversational format (vs Founder Firesides interview format).
  • Format: ~34-minute three-way conversation; founder + two YC partners; substantive analytical pressure from the interviewers (David Lieb’s “if the metric is revenue per employee, yes I imagine that is going to go through the roof” and Tom Blomfield’s extension of the thesis to all knowledge work in 5-7 years are partner-side contributions, not just founder soliloquy).
  • Empirical anchors: $25M revenue today, $100M target end of 2026; 60-100 employees end of year; the $3-4M of co-pilot licences line about a specific large firm is checkable in principle (though Vitucci wisely doesn’t name which firm).
  • Bias / motive: Series-A-raising founder competing with the Big 4; treat the Big-4-will-flatten prediction as motivated, but the analysis of senior-partner-incentives is grounded in Vitucci’s direct insider experience at Grant Thornton.
  • Transcript provenance: youtube-transcript-skill (Playwright path) — succeeded on first attempt; one of two videos in this batch (with the Stanford CS153 lecture) that did not require the yt-dlp fallback. ASR clean enough that no special post-processing was needed beyond standard rolling-caption dedupe (which was a no-op on this transcript shape).

Relationships

Supports

  • Stanford CS153 Frontier Systems | The AI Native Company: How One Founder Becomes a 1000x Engineer — Onshore is the wiki's first **deep professional-services-industry-disruption anchor** that complements Tan & Hu's CS153 *forward-deployed-engineer wedge* prescription with a worked example *one industry deep*. Hu names the wedge pattern generically (Salient / Happy Robot / Reducto); Vitucci is the wedge-pattern told as a 20-something-CPA-leaves-Grant-Thornton story over 33 minutes. **Productive tension on *boil-the-ocean***: Tan-CS153 inverts the *don't boil the ocean* dictum (*'let's boil the ocean, you can do the work of 500 to 1000 people'*); Vitucci asserts the conservative version (*'the idea of trying to boil the ocean all at once is very challenging; it has been an incredible benefit for our business to be great at one thing really early'*). The wiki carries both — they are both defensible at different stages and the contradiction itself is informative.
  • The ERP for the AI Revolution is here — Twin **AI-native-vendor-disrupting-incumbent-service-business** anchors on the YC Root Access channel. Glasgow / Campfire competes with [[NetSuite]] for tech-company ERP; Vitucci / Onshore competes with the **Big Four (Deloitte / KPMG / EY / PwC) + top-20 firms (Grant Thornton et al.)** for corporate tax and (eventually) audit. Both articulate the same incumbent-defection mechanism: incumbents have **misaligned senior-partner incentives** that block AI investment (Glasgow's *AI-native safety inversion*; Vitucci's *senior-partner-3-years-from-retirement-won't-plow-money-into-AI*). Both founders began by selling-to-incumbents and pivoted to compete-with-incumbents (Glasgow implicitly from his Adobe ERP background; Vitucci explicitly after 2 years of *banging his head against the wall* selling AI software to accounting firms). Both arrive at **revenue-per-employee** as the killer comp (Campfire: implied 10× of NetSuite's per-employee revenue; Onshore: explicit *'$1M+ per employee at $100M ARR target — Deloitte / KPMG estimated at $100-150k per employee, an order of magnitude better'*).
  • This $1.5 Trillion Industry Still Runs on Paper and Fax Machines — Twin **2026 founder-fireside on vertical AI-native operator** ingests on the YC Root Access channel, two months apart. Both founders work an industry vertical that is structurally hostile to AI investment for *misaligned-incentive* reasons (Vori: 220k+ independent grocers running on paper, pencil, fax — no in-house IT to deploy software; Onshore: Big 4 / top-20 senior partners whose pension and 3-5-year-runway-to-retirement makes AI investment a personal-financial-negative). Both arrive at **wedge-first** as the disciplined way in (Vori: inventory-reorder app; Onshore: R&D tax credit). Both name **fired-all-our-incumbent-customers / fired-all-our-accounting-firm-customers** as the pivot moment that unlocked the real opportunity.

Contradicts

  • Stanford CS153 Frontier Systems | The AI Native Company: How One Founder Becomes a 1000x Engineer — Direct *boil-the-ocean* contradiction. Tan-CS153 names *boil-the-ocean* as the pro-AI inversion of conservative business advice (*'let's boil the ocean — you can do the work of 500 to 1000 people'*); Vitucci-Onshore names *don't-boil-the-ocean* as the disciplined survival strategy (*'the idea of trying to boil the ocean all at once is very challenging; we can come and show we can do an R&D tax credit better, faster, more affordably with better traceability — it's way better than showing up saying we can do all your stuff'*). Both are right at different stages: pre-revenue capability-claim (*how much can one engineer ship?*) vs revenue-stage wedge-discipline (*how do you win one CFO conversation against the Big 4?*). The wiki carries the disagreement explicitly — there is no synthesis that makes them agree, and the productive read is that Tan and Vitucci are arguing about different things.

Graph View

Backlinks