Lotus Li writes about how AI could change professional services and is building an AI-native accounting and tax advisory business. Their posts focus on using AI agents to deliver work differently, with human professionals reviewing the results.
Lotus Li is most useful for explaining AI-native professional services economics, delivery models, and client-data risks.
Best for
A B2B services software campaign for accounting and tax firm leaders on AI-native delivery.
Not for
Consumer tax apps seeking short-form video with guaranteed reach to first-time filers.
Key takeaways
A Big Four engagement was $250,130 across 785 hours and three months.
They ask whether that model can reduce to token costs and human review.
Uploading client cases to general AI agents commoditizes the firm's own expertise.
Check first: Confirm partnership interest, audience overlap, and post-level reach and response data.
Main ideas
Professional-services pricing should be rebuilt around token cost and human hours, not the staffing pyramid · Value shifts from how fast a person executes to how many agents they can orchestrate · Agents can replace the leverage, but a licensed human still signs off on the work
Lotus Li writes about how AI agents could reshape professional services and is building an AI-native accounting and tax advisory business.
Their posts connect the big idea—using agents to change how expertise is delivered—to practical questions about staffing, pricing, and client data.
In one post on AI-native professional services, they break down a Big Four engagement as $250,130 across 785 hours and three months, then ask whether firms can reduce that model to token costs and human review.
Read their work for a commercially grounded view of AI-native professional services, including both the proposed opportunity and a serious implementation concern.
Li’s distinctive contribution is to make the economics and tradeoffs of AI-enabled professional services concrete.
The Big Four engagement breakdown gives readers a specific picture of the labor and cost an AI-native model aims to change.
They also sort potential providers into three models: traditional firms adding proprietary AI, owners retrofitting acquired firms, and licensed AI-native firms where agents work alongside human reviewers.
In a post on agent orchestration, they argue that a professional’s value may increasingly come from coordinating a fleet of agents rather than executing every task personally.
The Claude-agents post adds a useful tension to that vision, questioning what happens to client-derived expertise when it is put into an AI tool.
Together, these posts help readers assess not just whether AI can speed up professional work, but how it could change pricing, delivery, and risk.
Lotus Li makes one argument in different keys: the hourly pyramid that professional services is built on is being undercut not by technology alone, but by clients who have seen work delivered and priced another way.
Read more
They are building an AI-native accounting and tax advisory practice around that thesis, while warning firms that handing client work to general AI agents gives away the expertise they currently sell. Across the three posts, the priorities are pricing that reflects token cost and human hours, human sign-off on AI-delivered work, and control over where client data and its benefits end up.
Professional-services pricing should be rebuilt around token cost and human hours, not the staffing pyramid
Lotus Li argues that professional services has been built and priced on one assumption: intelligence is human and scarce.
Read more
That is why clients wait weeks for a memo and pay for the pyramid of people behind it.
Their reasoning turns on client expectations, not technology: once clients have seen work delivered and priced another way, the old model loses its justification.
They sort firms into three futures: ceiling firms that keep the model and trade associate leverage for proprietary AI while charging more for senior judgment; AI-enabled owners that buy licensed firms with capital and retrofit them; and AI-native neofirms with human reviewers signing off.
They place their own venture in the third group.
A full itemized Big 4 engagement: 785 hours across seven staffing tiers, from partners at $950 and $1,050 an hour down to admin at $125; a $255,700 subtotal; a 10% client discount; $20,000 travel; and a $250,130 total over three months — plus at least 260 hours from the client's own staff on PBC lists, status calls, and meetings. Lotus Li asks whether the Big 4 can compress that table to two lines — token cost and human hours — and says they doubt it, while conceding that decades of trust and relationships are real.
Value shifts from how fast a person executes to how many agents they can orchestrate
Quoting a line from PitchBook's report on the AI Super-Cycle, Lotus Li describes an era in which a single professional scales by orchestrating fleets of agents, so market value is set by a person's orchestration capacity rather than their personal execution speed.
Read more
This is the mechanism behind the delivery model: the productivity claim is what lets a small team take on work that used to require a pyramid. The productivity figures come from the cited report, not Lotus Li's own measurement.
The shorthand comparison from the framing: tools produce roughly a 2X productivity gain, while a team produces roughly 100X. It closes with “an individual is a team” and the line that everyone will feel like a superman or superwoman.
Agents can replace the leverage, but a licensed human still signs off on the work
For Lotus Li, the priority in AI-native firms is that automation changes the delivery and pricing model, not who is accountable.
Read more
Neofirms are licensed and AI-native, built on a new delivery and pricing model, with human reviewers signing off.
The accountability structure stays; the leverage underneath it is replaced.
That distinguishes their position from simply selling more software to the same pyramid.
Crosby AI, cited as the working example, has agents and lawyers review contracts together, billing by the page with a stated 58-minute median turnaround.
Uploading client work to general AI agents commoditizes the firm's own expertise
Writing to CPAs, consultants, lawyers, and other professional-services firms adopting Claude agents, Lotus Li asks what they are selling (human expertise) and what they have that Claude does not (real-world data and experience from client cases).
Read more
Once a client case is uploaded and used, the argument goes, it is no longer only the firm's: the model learns from it and serves the firm, its clients, and its competitors at the same time.
The firm automates the work while commoditizing its most valuable human insights.
The post offers no resolution.
It sits in tension with the practice Lotus Li is building: the objection is about where client data goes and who captures the benefit, not whether agents belong in professional services at all.
The post addressed to professionals adopting Claude agents ends with the claim that the firm is commoditizing its most valuable human insights while automating the work — a caution about a specific adoption pattern, not a rejection of AI.
This is the clearest statement of what Lotus Li is building and whom it targets.
Read more
Lotus Li argues that professional services have been priced on the assumption that intelligence is human and scarce, and that client expectations change when work is delivered and priced differently.
The post sorts firms into three futures: “ceiling firms” that keep the pyramid but trade associate leverage for proprietary AI and charge more for senior judgment (naming Kirkland & Ellis and Baker McKenzie); “AI-enabled owners” that buy licensed firms with capital and retrofit them (Thrive Holdings); and “neofirms,” which are licensed and AI-native, with human reviewers signing off.
The example is Crosby AI, where agents and lawyers review contracts together, billed by the page, with a stated 58-minute median turnaround.
At the post’s core is an itemized cost table for one typical Big 4 engagement: partner hours at $950 and $1,050, senior managers at $565, managers at $450, senior associates at $325, associates at $250, admin at $125, a $255,700 subtotal, a 10% client discount, $20,000 travel, and a $250,130 total across 785 hours and three months—plus roughly 260 more hours from the client’s own staff on PBC lists, status calls and meetings.
They ask whether the Big 4 can compress that table to two lines—token cost and human hours—and doubt it, while acknowledging that decades of trust and relationships are real.
The post closes with the venture’s starting scope: accounting and tax advisory services (R&D tax credits, income tax provision, technical accounting), and an open call for pilot clients and partners from firms.
For brands planning a campaign
For a co-branded webinar on AI-native delivery and pricing models, this post supplies the session’s spine: an engagement cost table ready to display, a three-firm taxonomy that helps the audience place themselves, and a defined service scope—R&D tax credits, income tax provision and technical accounting—that shows which areas the speaker can credibly discuss. The explicit “accepting clients for pilot and partners from firms” line makes firm leaders the intended audience.
This post explains where the productivity in the delivery model described in the other posts is supposed to come from.
Read more
Quoting a line from PitchBook’s report on the AI Super-Cycle, Lotus Li describes an era in which one human professional scales their impact by orchestrating fleets of agents, and argues that market value shifts from a person’s own execution speed to their “orchestration capacity.”
The post compresses the argument into a comparison attributed to the framing: tools give roughly 2X productivity gain while a team gives roughly 100X.
It closes with “an individual is a team” and the observation that “everyone will feel like a superman or superwoman.”
The tags are #digitalworkforce and #individualasplatform.
For brands planning a campaign
This is the thesis slide for a webinar aimed at accounting and tax advisory firm leaders.
It frames the argument as a shift in what firms should hire and reward—orchestration capacity rather than raw hours—and sets up the pricing discussion in the other posts.
Ask for Lotus Li’s own measurement before treating the cited productivity figures as evidence of results.
This post shows the skeptical question Lotus Li raises about AI adoption, helping marketers assess how candid they are about the downside of the model they are selling.
Read more
Writing to CPAs, consultants, lawyers and other professional services firms that adopt Claude agents, Lotus Li walks through a chain of questions: what are you selling (human expertise), and what do you have that Claude does not (real-world data and experience from client cases)?
The argument is that once a client case is uploaded and used, it is no longer only the firm’s: Claude learns from it and serves the firm, its clients and its competitors at the same time.
The firm thus commoditizes its most valuable human insights while automating the work.
For brands planning a campaign
Use this as the tension or objection segment of a webinar: it names the data-confidentiality and differentiation risks firm leaders will raise, and lets the speaker raise them first. Pair it with the delivery and pricing argument from the other posts rather than presenting it as a standalone position.
Format: short argument posts built around a number or a quoted line3
Building an AI-native accounting and tax practice1
Caution about putting client work into general AI agents1
The 'individual as platform' view of AI leverage1
Building an AI-native accounting and tax practice
Lotus Li writes as the founder of an AI-native accounting and tax advisory business targeting work the Big 4 bill by the hour: R&D tax credits, income tax provision, and technical accounting.
Read more
In the longest post, they break down a typical Big 4 engagement—785 hours across seven staffing tiers, discounted to roughly $250,000, plus at least 260 hours of client-side work—and ask whether that table can be compressed to two lines: token cost and human hours. The plan is to start with accounting and tax advisory, take on pilot clients, and recruit partners from established firms.
Builds the case from a firm-level cost table and named market examples, dividing professional services into three emerging types: ceiling firms that trade associate leverage for proprietary AI, AI-enabled owners that buy licensed firms and retrofit them, and AI-native neofirms with human reviewers signing off.
For brands planning a campaign
For a co-branded webinar on AI-native delivery and pricing, this offers a ready-made spine: a side-by-side comparison of pyramid pricing and agent-plus-reviewer pricing, presented by someone making that bet rather than commenting from the outside. The three-firm taxonomy gives the session a simple structure and a natural place to include a software vendor's tooling without turning it into a product demo.
Caution about putting client work into general AI agents
In a post addressed to CPAs, consultants, lawyers, and other professional-services firms adopting Claude agents, Lotus Li argues the trade is less one-sided than it looks.
Read more
The firm sells human expertise, differentiated by real-world experience from client cases; once a case is uploaded, the model learns from it. The conclusion is blunt: the firm is commoditizing its most valuable human insights while automating the work.
This is a caution about a specific adoption pattern, not a rejection of AI. The writer is building an agent-based practice with human reviewers signing off, so the objection is about where client data goes and who benefits from it—not whether agents belong in professional services.
For brands planning a campaign
This is a debate hook for a webinar aimed at firm leaders: what stays out of third-party models, and what can safely be automated. Vendors pitching data control or private deployment can use the argument directly; those selling open model access should expect this objection and come prepared with an answer.
Quoting PitchBook's report on the AI Super-Cycle, Lotus Li argues that one professional now scales by orchestrating fleets of agents, shifting market value from personal execution speed to orchestration capacity.
Read more
Their shorthand: tools give a 2X productivity gain, teams give 100X, and everyone will feel like a superman or superwoman.
A short, quotable framing built around a borrowed analyst line rather than original analysis. The post's force comes from the line itself, not evidence assembled by the writer.
For brands planning a campaign
A strong one-line thesis for a webinar title or opening slide, particularly for tools sold to small and mid-sized firms that cannot add headcount. Ask Lotus Li for a before-and-after example to make the claim concrete for an audience of firm leaders.
Format: short argument posts built around a number or a quoted line
Across the three available posts, the pattern is consistent: compact LinkedIn arguments, roughly one to a few hundred words, opening with a claim, a direct question to the reader, or a quoted line.
Read more
One carries a full engagement cost table; another turns on a single sentence from an analyst report; the third is a short question-and-answer sequence aimed at a named professional audience. The tone is declarative, with one idea per post.
For brands planning a campaign
Co-marketing assets should match this register: text-first posts, slide-ready statistics, and one sharp claim per piece rather than long-form thought leadership. The engagement cost table is the most portable asset: it works as a carousel, a slide, or a standalone post, and carries the webinar's pricing argument on its own.
Lotus Li's posts disclose no sponsorship, affiliate, employment, or investment relationships.
Read more
They discuss Anthropic's Claude as software professional-services firms adopt and announce that they are building an AI-native accounting and tax advisory practice taking pilot clients and firm partners. Ask for the business name and their formal role.
Claude (Anthropic)
Mention
Li writes about what happens when CPA, consulting, and legal firms adopt Claude agents, framing Claude as a tool those firms run and warning that client cases uploaded to it also train the model.
Read more
The post comments on adoption and states no payment, endorsement, or working relationship with Anthropic; it does not name the company directly.
Explore Lotus Li's work through three LinkedIn posts that build one argument: professional services is still priced on a human-hour pyramid that clients no longer accept.
Read more
Start with the long post that itemizes a typical Big 4 engagement—785 hours across seven staffing levels, roughly $250,000 after discount, plus at least 260 client-side hours—and sorts firms into ceiling firms, AI-enabled owners, and AI-native neofirms with human reviewers signing off.
Then read the shorter post, which borrows PitchBook's “individual as a platform” line to argue that orchestrating agents matters more than personal execution speed.
Finish with the warning to CPAs, consultants, and lawyers that uploading client work to general AI agents can teach the model and serve competitors, commoditizing the expertise the firm sells.
They are building an AI-native accounting and tax advisory practice while warning about where client data and its benefits go.
Use the tags to move between AI-native delivery, professional-services pricing, and client-data control.
Similar creators share the professional-services audience but focus on different levers: firm economics, rainmaking, or in-house finance operations.
Both write for leaders of professional-services firms and argue that results depend on how work and economics are structured, using concrete firm examples rather than generic advice.
Read more
Fotsch focuses on founder dependency and economic engagement—giving employees a stake and line of sight to profit—while Lotus Li focuses on AI-native delivery, agent orchestration, and pricing against the Big 4 pyramid.
Both address leaders of professional-services firms and treat client relationships and go-to-market as things a firm can deliberately design, drawing on named firms and research rather than personal opinion alone.
Read more
Channer's lens is rainmaking behavior and client engagement, backed by a book and multi-firm research; Lotus Li's is AI-native delivery economics and control of client data.
Both write about AI changing accounting and finance work, using concrete operational details—contract parsing, revenue recognition, invoicing—to show what changes in practice.
Read more
Hu writes from inside a company's finance function about CFO workflows and fintech product strategy; Lotus Li writes about professional-services delivery and pricing models and the risk of commoditizing client expertise.
Lotus Li’s LinkedIn posts address professionals and decision-makers considering how AI could reshape professional-services work.
Read more
The topic is especially relevant to firms exploring AI-native accounting and tax advisory, though their posts do not say how many followers work in those fields.
In a post about AI agents and client expertise, they warn that uploading client cases to an AI tool could erode the knowledge professionals rely on to distinguish their services.
A post on AI-native accounting makes the business case more concrete, contrasting a traditional Big Four engagement’s time and costs with a proposed model built around AI and human review.
Lotus Li's three available LinkedIn posts do not describe sponsored placements, paid collaborations, media kits, rate cards, audience packages, or a partnership inquiry route.
Read more
The only partner-facing request runs the other way: in the September 2026 post, they say the accounting and tax advisory business is "currently accepting clients for pilot and partners from firms" for R&D tax credits, income tax provision, and technical accounting work.
This is a call for clients and delivery partners for their own firm, not an offer of sponsored content or creator collaborations.
Buyers considering a co-branded webinar, sponsored post, or similar placement should ask whether partnership opportunities are available and what terms apply.
Campaign need: A B2B services software company wants to reach accounting and tax advisory firm leaders with a co-branded webinar on AI-native delivery and pricing models.
Read more
Decision: Strong subject fit, but the partnership is unconfirmed and audience overlap unverified.
Why it fits: All three available posts target the campaign’s intended audience.
The strongest fit is the accounting and tax post: Lotus Li is building an AI-native accounting and tax advisory business focused on R&D tax credits, income tax provision, and technical accounting.
They lay out a full Big 4 engagement cost table, including partner hours at $950 and $1,050, senior managers at $565, managers at $450, senior associates at $325, associates at $250, and admin at $125, reaching a $255,700 subtotal and a $250,130 total across 785 hours and three months.
They then ask whether that table can compress to two lines: token cost and human hours.
That is a ready-made webinar spine, not a generic AI lecture.
The post also says the business is currently accepting clients for pilot and partners from firms, showing that Lotus Li is already speaking to firm leaders and recruiting partners from established firms.
A second post offers a useful objection-handling segment.
Lotus Li warns CPAs, consultants, lawyers, and other professional-services firms that uploading client cases into Claude agents commoditizes a firm's own case experience while automating the work.
For a software vendor, this cuts both ways: it is credible tension that firm leaders will recognize, and it creates a natural place to discuss data handling, private deployments, and human review if the vendor's product addresses those points.
A third post offers the framing thesis: one professional orchestrating fleets of agents shifts value from personal execution speed to orchestration capacity.
It is a short secondhand reading of a PitchBook report, so use it as a hook, not measured proof.
Conditions and gaps: Lotus Li has about 3,275 LinkedIn followers as of October 11, 2026, but follower count does not establish whether accounting and tax firm leaders are among them.
Before booking, ask for audience demographics and recent engagement data, and confirm webinar terms directly with Lotus Li; no partnership or sponsorship program, media kit, rate card, or inquiry route is mentioned.
When to choose a specialist
Campaign need: A consumer tax-preparation app wants a short-form video campaign with guaranteed reach to first-time individual filers.
Read more
Decision: Poor fit based on the available information.
Why it fails: All three available posts address professional-services firms and their leaders, not individual taxpayers.
They cover firm delivery models, engagement pricing, leverage, and client-data risk.
There is no consumer filing content or tax-season consumer advice, and the available information does not indicate an audience of first-time filers.
Lotus Li has about 3,275 followers, a small base for a high-volume reach campaign; the following is not broken down by taxpayer status, age, income, or geography.
There is also no video, short-form, or platform-performance data to show that Lotus Li can deliver this campaign.
A campaign requiring guaranteed reach and video execution would be buying unverified capabilities.
The larger problem is the topic mismatch: the posts focus on how accounting and tax firms should price, staff, and protect client work—a different audience and buying decision from first-time individual filers choosing a tax app.
If the app instead wanted to reach accounting firm leaders as channel partners, that would be a different campaign requiring reassessment.
A worked campaign example
Campaign angle: The 785-hour engagement versus two lines. A co-branded webinar on what AI-native delivery and pricing actually change inside accounting and tax advisory firms.
Why this angle: It builds directly on Lotus Li's strongest post, which already contains the exhibit, the argument, and the firm-leader audience.
It gives firm leaders a concrete before-and-after rather than an abstract AI promise.
Lotus Li can speak as a practitioner building this model, not as a commentator.
Proposed structure: (1) Open with the Big 4 engagement table and the 785-hour, three-month, $250,130 example, plus the roughly 260 client-side hours.
(2) Lotus Li walks through the three futures in the post: ceiling firms that keep the pyramid and charge more for senior judgment, AI-enabled owners that buy licensed firms and retrofit them, and AI-native neofirms with human reviewers signing off.
Use a live poll to let attendees place their own firm.
(3) Cover data control using the Claude agents warning: what happens when client cases become training input, and who else the model serves.
The software vendor explains how its product handles client data, private models, human reviewer sign-off, and audit trails, if those features exist.
(4) Work through an example in R&D tax credits or income tax provision, the service lines Lotus Li names as the starting scope.
(5) Close with a pilot offer: a scoped review of one engagement process with defined human review points and a pricing comparison.
Format and assets: A 45-minute live webinar with 30 minutes of discussion and 15 minutes of Q&A.
Use the engagement cost table as the opening slide and the three-firm taxonomy as a self-placement poll.
A one-page follow-up should include the table and a simple decision checklist for firm leaders.
Co-branded registration copy should lead with the cost comparison, not a product demo.
Before the session, Lotus Li should publish a short text-first post stating the same contrast and driving registration.
Fit conditions: This is a proposed campaign, not a documented Lotus Li offer.
Before booking, confirm directly whether Lotus Li is available and on what terms, since no partnership or sponsorship route is mentioned.
Ask for follower demographics and recent post engagement to assess the share of accounting and tax firm leaders.
Verify that the software vendor can demonstrate its claims about data handling, private deployment, and human review, since Lotus Li's post raises those concerns.
Also confirm that the vendor is comfortable with Lotus Li's cautious framing of third-party AI agents, which may not match a generic open-model pitch.
Reader task test: A firm leader should finish the session able to compare their current delivery and pricing model with an AI-native alternative, name one data-control question to ask any vendor, and take a specific next step, such as a pilot scoping call.
For the same co-branded webinar decision, Bill Fotsch is the stronger choice when the software company's buyer is a boutique professional-services firm CEO and the product depends on changing how the whole firm sees and shares economic outcomes.
Read more
Fotsch's posts describe a founder-dependency trap in which every major decision returns to the owner, and present economic engagement as the fix.
The Adams Beasley Associates example gives a named case: sales doubled in year one, profits grew faster, and there were no layoffs.
The posts also point to a free May 5 session capped at 10 firms.
That offers more concrete detail on firm economics and owner bottlenecks than Lotus Li's posts.
Fotsch's posts do not address AI-native delivery or pricing.
A webinar built around Fotsch would shift the topic to founder dependency, employee ownership and economic engagement, so the software vendor would need a product story that connects to firm economics rather than AI cost compression.
The audience is described as boutique consulting firm CEOs, not specifically accounting and tax advisory firm leaders.
For the same webinar decision, Rory Channer is the stronger choice when the campaign needs book-backed, research-led authority on client relationship building and rainmaker behavior across professional services, including accounting firms.
Read more
Channer's posts describe The Activator Advantage, published by Harvard Business Review Press on May 20, 2025, and cite more than 18 months of research with over 100 global professional services firms, sponsored by Intapp.
The posts speak directly to partners and professional-services leaders about how clients buy and how firms develop client relationships.
That gives the webinar a more established, research-anchored platform than Lotus Li's posts.
Channer's posts focus on business development, client engagement and rainmaker behaviors, not AI-native delivery or pricing.
A webinar would shift from Lotus Li's Big 4 cost table to relationship-building and go-to-market change.
The posts are dated January and May 2025 and offer no AI cost or delivery model.
Ask for evidence of audience overlap with accounting and tax advisory firm leaders, and for commercial terms.
For the same webinar decision, Steve Hu is the stronger choice when the software company's buyer is a CFO or finance team and the demonstration should show AI working inside finance operations.
Read more
Hu's posts give a specific first-90-days workflow: an agent parsed every company contract into JSON, a routine handled new contracts, and a revenue-recognition layer ran a GAAP process alongside the accounting team.
A second post examines Ramp's entry into accounts receivable and explains why invoice creation, reconciliation and revenue recognition are harder to justify as standalone tools.
That offers more concrete detail on finance-team execution and software buying than Lotus Li's firm-level pricing argument.
Hu's audience is CFOs and finance operators, not accounting and tax advisory firm leaders acting as external advisers.
The posts do not cover selling professional services, firm leverage or pricing an advisory engagement.
A webinar with Hu would focus on finance operations and AI implementation, not a co-branded discussion of AI-native delivery and pricing for advisory firms.
Ask for evidence of audience overlap and for commercial terms.
Three complete LinkedIn posts from Lotus Li's account are available here, and the full text of each is readable rather than just titles or excerpts.
The posts describe an AI-native accounting and tax advisory practice scoped to R&D tax credits, income tax provision, and technical accounting, including an itemized Big 4 engagement of 785 hours and about $250,130; a reading of PitchBook's AI Super-Cycle report that argues a professional's value shifts from personal execution speed to how many agents they can orchestrate; and a warning that uploading client cases to general AI agents lets the model learn from that work and serve competitors, commoditizing the expertise a firm sells.
None of the three posts carry a visible publication date.
A LinkedIn profile viewed October 11, 2026 lists about 3,275 followers.
For the creator-comparison section, short samples of posts by three other writers — Bill Fotsch, Rory Channer, and Steve Hu — provide context for Lotus Li alongside people addressing the same professional-services audience from different angles; those posts date from 2025 and 2026 and are not Lotus Li's own work.
The page does not represent Lotus Li's venture as a verified business, and the posts here do not describe sponsorship formats, a media kit, a rate card, or an inquiry route.
This is a close reading of three undated posts plus a follower count, not a full posting history, an audience breakdown, or a commercial offering.
