Picking up the firm four years after the audit role split. Some context first, because a lot has happened.
The advisory practice completed its own role redesign about eighteen months after audit, on a different model than the audit split but with the same underlying principle. Advisory consolidated three previously distinct roles into two: a research-and-synthesis role where the bulk of the analytical work is AI-supported, and a client engagement role focused on diagnosis, framing, and delivery. The advisory redesign was less politically fraught than the audit one because the advisory practice had less institutional history to defend.
The tax practice finally moved into its own redesign about two years after audit, prompted by a combination of regulatory clarification on AI use in tax preparation, competitive pressure from peer firms that had moved earlier, and a generational turnover among the tax partners that brought in leadership less attached to the legacy model. The tax redesign is the most cautious of the three and is still ongoing.
The Managing Partner from the previous posts retired about eighteen months ago. He was succeeded by Rita, who took the Managing Partner role after spending two years building credibility outside her audit practice. Linda is still the CHRO. Marcus is still the CIO. Dana retired around the same time as the previous Managing Partner and was succeeded by her deputy, who was a less central character in the previous posts but who has been important in the firm's financial reorientation around the new operating model.
The firm has grown from 450 staff to about 600 over four years, mostly through hiring into the new role types, with some attrition from the legacy track that the firm chose not to backfill. Revenue has grown faster than headcount. Partner economics have improved meaningfully, though the partner compensation structure has been substantially revised to reflect the new operating model.
That's the state of things. The interesting question is whether the firm has reached the transformational stage the diagnostic post described, and whether the answer to that question is yes, no, or something more honest than either.
What the firm now does that it could not previously do:
The audit practice handles engagements that were structurally beyond the firm's capacity four years ago. Specifically, mid-cap clients with multi-entity consolidation requirements that previously required either a larger firm or an extended timeline have become standard work. The audit practice runs these engagements in roughly the same timeframes the firm used to require for simpler clients, with the AI-augmented review specialists handling the consolidation logic at a scale that would not have been feasible with the prior staffing model. The firm has won several clients from larger competitors specifically because of this capability. The pitch to those clients includes a credible claim that the firm can deliver Big Four-quality work at mid-market pricing because the firm's operating model produces that capability rather than just promising it.
The advisory practice has developed a category of engagement that didn't exist for the firm before: rapid-cycle strategic diagnostics. A client brings a question, and the practice delivers a substantive analysis with primary research, market scanning, and recommendation framing in two to three weeks rather than the eight to twelve weeks that comparable work used to take. The compression isn't from working faster. It's from the research-and-synthesis role being able to assemble the analytical substrate in a fraction of the time, which lets the client engagement role spend almost all of their time on the parts of the work that require human judgment. The rapid-cycle engagement type has become a real revenue line and a recruiting differentiator.
The tax practice's transformation is the slowest-moving of the three and the hardest to characterize cleanly. Tax preparation efficiency has improved measurably, but tax practice work depends on judgment about ambiguous regulations in ways that resist clean automation. The transformation in tax has been less about new capability than about freeing senior tax professionals from preparation work so they can focus on advisory, planning, and complex transactional work that the firm couldn't previously offer at scale. The tax practice's revenue mix has shifted away from compliance preparation and toward tax advisory in a way that was not part of the original plan but that has turned out to be the more important shift.
Across all three practices, the firm's clients are different than they were. More of them. Larger ones. Stickier engagements. Higher proportion of advisory and judgment work, lower proportion of mechanical compliance work. The firm's competitive position has shifted from "regional mid-market firm with good people" to "regional firm with capabilities that were previously only available from larger national firms." Several of the firm's larger competitors have started actively recruiting from the firm, which is a kind of validation the firm did not previously experience.
The new categories of work that now exist at the firm:
AI-augmented review specialist, as introduced in the previous post, is now an established role across audit and advisory with a clear career path and defined competencies. There are about forty of these people across the firm, with their own training program, their own promotion criteria, and their own partner-track pipeline. The role has its own internal community, with monthly meetings to discuss edge cases, share patterns, and develop new approaches. The role's external recognition is still emerging, the professional designation work that audit credentials have historically been built around hasn't caught up to what these people actually do.
AI operations engineer is a new role on Marcus's team. The role didn't exist when the audit project started. It exists now because the firm operates a significant AI-supported workflow surface that requires ongoing tuning, model monitoring, vendor coordination, and integration maintenance. There are six of them. The job didn't exist in any meaningful form three years ago.
Workflow design analyst is a new role that sits between IT and the practices. The role's job is to identify candidates for AI-supported workflow redesign, run the operational connection conversations with practice leaders, and shepherd projects through implementation. There are four of these people. They are, in effect, what Rita and Marcus were doing in stage three, formalized into a defined function so the firm doesn't have to rely on practice leaders happening to ask the right question.
Client AI advisory is the newest category. The advisory practice has spun up a small team that helps clients think through their own AI adoption, drawing on the firm's hands-on experience. The engagement type is new enough that the team is still figuring out how to scope and price it. The work itself is going well. The clients who buy this engagement are usually firms in the firm's market segment who are at stage one or two of the maturity progression and want help getting to stage three without making the mistakes the firm itself made.
None of these roles existed three years ago in any meaningful form. The firm did not plan to create them. They emerged from the work the firm was doing, and the firm has been catching up to them rather than designing them from the start. This is the pattern the maturity ladder predicts. The job categories that emerge from a technological cycle are not the ones the cycle's early enthusiasts predicted. They are the ones the work itself produced over time.
What the firm has not figured out:
The partner track for people who came up through the new roles is still an unresolved problem. The traditional audit partner track was built around a specific career progression: associate, senior associate, manager, senior manager, partner. The skills developed at each level were the skills required to be the next level. People making partner had typically managed dozens of audits and had developed judgment by doing that work for many years.
The AI-augmented review specialist track does not produce the same kind of person. The people excelling in those roles have different skill profiles. Their judgment is sharper in narrower domains. Their managerial experience is different in shape. The firm has, in the last year, promoted its first AI-augmented review specialist to partner. The promotion was contested in the partner vote, not because the candidate was unqualified but because the partner group did not have a shared mental model for what an audit partner trained in the new track is supposed to be good at. The vote went through. The questions remain open.
The compensation structure has been substantively revised but is still being calibrated. The two tracks were initially compensated at equivalent levels at each rank, which felt right at the time. Over four years, it has become apparent that the work the two tracks produce is not equivalent in revenue contribution per hour, and that the original equivalence was producing distortions in retention and recruiting. The firm is now in the third revision of its compensation matrix in eighteen months. It will probably revise again.
The recruiting pipeline is healthier than it was during the transition but is not yet stable. Universities are still catching up to what the firm now wants from new graduates. The accounting curricula that have been the firm's traditional pipeline still mostly produce people who would have fit the legacy senior associate role. The firm has had to develop its own substantial onboarding investment to bring new hires up to either of the new tracks. The investment is real and the firm absorbs it because the alternative is worse. It is not yet a sustainable model at scale.
The tax practice has not fully completed its redesign and may never do so in the way audit and advisory did. The structure of tax work resists clean splitting in a way that the other practices' work did not. The tax practice's transformation has been real but more incremental. Whether this is acceptable to the firm long-term is a question the new Managing Partner is still working through.
Other vantages on the firm at this stage:
A tax senior associate, four years into the transition. She is, in 2030, doing work that her predecessor in 2026 was doing, but the work has changed in texture. The mechanical preparation that used to fill her week is now mostly handled by AI tools that flag her attention when something needs human judgment. Her time is spent on tax planning, transactional work, and supervised client communication. She is, in her own assessment, doing work she would have been doing two ranks higher in the old model. The traditional career ladder has not yet caught up to her experience, and there is a small generation of tax associates in the firm who are professionally further along than their titles suggest. The firm's HR function is aware of this and working on it.
A new partner. The first AI-augmented review specialist to be promoted to partner, six months into the role. Her partner work is structured differently than her peers'. She handles fewer engagements at higher complexity. Her contribution to firm strategy includes regular input on the AI operating model in ways her peers' contributions don't. She is the prototype of what an audit partner from the new track looks like, and she is figuring out the role in real time alongside the firm. She has been told, in confidence, that several of the firm's larger competitors have reached out to her about partner positions at their firms. She has not seriously considered any of them. Yet.
A client board meeting. The firm's Managing Partner, Rita, is presenting to a client board on the firm's capabilities in the context of a competitive pitch the client is running. The pitch deck includes specific descriptions of how the firm delivers work, with AI capability described as foundational to the operating model rather than as a feature of certain projects. The client's board members ask questions that would not have been askable four years ago: how the firm's AI tooling handles specific compliance scenarios, what the firm's training program for AI-augmented work looks like, how the firm thinks about AI's implications for the client's own audit committee questions. Rita's answers are specific and credible. The firm wins the engagement.
The training program. A multi-week immersion for new hires in either of the new tracks. The program was built over the previous three years and has been substantively revised six times. The current version is still imperfect. The training producers are constantly trying to compress the time required to bring new hires to productive output, against the constraint that the work itself requires real depth. The current state is that the firm's new hires reach baseline productivity in their respective tracks in about half the time their predecessors required in the legacy model. The training producers expect to compress this further. The training program itself is a new functional area of the firm that didn't previously exist.
The firm has substantively transformed. The work it does is different in kind from the work it did four years ago. The people it employs are organized into roles that didn't exist when the project began. The client base has shifted. The competitive position has shifted. The firm's pitch to new clients depends on capabilities that AI made possible and that the firm built deliberately over four years of uneven, contested, expensive work.
The firm has also not finished transforming. The partner track for new-role staff is still being worked out. The compensation matrix is still being calibrated. The tax practice is still incomplete. The recruiting pipeline is still unsteady. New questions about the firm's operating model emerge faster than the firm can resolve them.
This is what stage five actually looks like, which is to say: not what the vendor pitches and conference keynotes describe. The transformation is real and the work is ongoing. The firm has reached a state where AI capability is foundational to its operations, where removing AI would mean abandoning categories of work it now does, where new professional roles exist that didn't exist four years ago, and where the competitive position has shifted in ways that are visible to the market.
The firm has also paid for that transformation in ways the previous posts described and in ways the firm is still paying. Several capable people left during the transition. The institutional culture took a meaningful hit and has rebuilt around different assumptions. The partnership track is structurally different from what it was, in ways that are still being worked out. The firm's leadership spends a meaningful portion of its time on questions that did not exist five years ago.
Stage five is not a destination the firm arrived at. It's a state the firm is now operating in, with all the ongoing work that any operating state entails. The transformational stage in the diagnostic post is, in this honest portrait, less an achievement than a description of a firm that has gone through the previous four stages and is now living with the results.
That's the case study. The progression is meant to be recognizable. The pace is meant to be realistic. The costs are meant to be visible. The gains are meant to be specific. Most organizations reading this series will see themselves clearly at one of the earlier stages, and that's fine. The point of the maturity progression is not that every organization needs to be at stage five. The point is that every organization is somewhere on it, and most are earlier than they think.
The next posts in this series will move away from the theoretical case and into specific patterns I have observed in actual work: what makes operational connection conversations succeed, what makes systemic redesign fail, where the IT and HR partnership most often breaks down, and where it most often holds. The case study has set the table. The pattern observations are the meal.
— Chris