How long until we are managing hundreds of AI workers?

“Very soon, I think the valuation metric for a good manager will be: How many digital workers can you manage?” -Wang Guanchun.

The AI agents of the future aren’t the helpful intern chatbots we have now. They don’t improvise, they don’t get tired, and they can work 24/7. So what will it mean for everyone in finance when a lot of our employees are bots? A complete shift in the control model: you’re not just checking people’s work anymore—you’re designing, approving, and monitoring the machines that do it.

What is an AI Agent (in Plain English)?

A task-doer that can plan → act → verify across your stack (ERP, bank feeds, spreadsheets, email) inside guardrails you set. Think of a bank reconciliation that clears 100% of matches, a PBC collector that grabs evidence and names files correctly, or a fraud bot that runs duplicate-vendor tests all day.

This is in stark contrast to traditional controls where humans prepare and review samples. Agents review everything, but the risk shifts to configuration and oversight. The upside is: no theft, consistent rules, and full checks. The new danger is a bad scaling rule or mistake in instructions. The new job becomes: set good rules, watch expectations, and prove it all happened. Proving it happened is another risk as if the checks weren’t run at all, that makes it easier to commit fraud.

Example of the New Day-to-Day

  • Bank reconciliation: The agent clears the routine matches. The manager stops re-checking every tie-out and instead validates the rule quality and the data feeds, then works only the exceptions.
  • Close: Agents assemble workpapers and flux drafts; humans make final judgements and verify quality.
  • Fraud: Agents run tests continuously; humans investigate hits, tune thresholds, and add new scenarios.

This results in more control with less emotion, and managers freed from triple-checking human work that a machine now does consistently.

The Skill Shift for Leaders

Leaders now need agent literacy: the ability to write policies, understand prompts, tools, and failure mods. They need control design thinking to phrase rules for both machine and human alike. There is a lot more anomaly triage, rather than tick-and-tie repetition. Working with vendors means understanding data residency and updating notes and procedures. Leadership becomes a lot more technical.

It also becomes a lot more human.

As we’ve all become more comfortable with technology, we’ve seen the calls for soft skills. People become more comfortable with their computers and phones and lose some of their communication skills, especially younger generations like Gen Z. While leadership becomes more technical for AI, it also becomes more soft-skill based. When teams over-index on tools, junior employees learn less judgement and don’t develop the wisdom to work with others and make big-picture decisions. The new leadership skillset is someone who can use the tech, manage the people, supervise the bots, build the culture, and still communicate and act with clarity and empathy. That’s a big skillset. Here’s what it looks like in practice:

  • Explainability over wizardry. You can describe how an agent made a call—and translate it for auditors, boards, and operators.

  • Exception coaching. You turn flagged items into teachable moments, so the rules (and the people) get smarter.

  • Psychological safety. Teams will only escalate weird results if it’s safe to say, “this looks wrong.”

  • Servant leadership. Less “I’m the boss,” more “I support system design, clear blockers, uplifting people, and making good work visible.”

The Bottom Line

Agents mean more and better controls. Your advantage isn’t prompt magic; it’s the mix of sound controls, sharp exception handling, and the human skills to align people, machines, and stakeholders. That’s the new leadership stack.

We help teams shift from rechecking everything to managing exceptions—with the controls to back it up. If you want a second set of eyes on your stack, let’s talk.

All my best,

Alex
Alex Brokhin headshot in suit against a dark background
Alexander Brokhin

 

President & Founder

 

Call : 713-970-1163
www.imperial-texas.com

AI is Homogenizing Our Thoughts

Article on New Yorker.

Photo by Tim Mossholder on Pexels.

AI is slowly draining us of our creativity by pulling our thoughts closer to the average. AI systems are trained on immense amounts of data and optimized for consensus, creating a gravity back toward the most common or acceptable information. Creativity and judgement require deliberate divergence, but that’s more difficult when AI provides such easy (if middle-of-the-road) answers.

What does this mean for businesses? AI could draw thinking toward a middle. If you use AI for any length of time, you’ll start to notice how much online content is written (or at least edited) by bots. This can cascade out to business strategy converging into a central cluster of “right” answers. Junior employees may not gain the knowledge needed for their field, or the wisdom to lead.

Here are a few ways we can keep original thought:

  • Force alternatives by generating different takes on a topic.
  • Vary sources & models: Mix human research, client interviews, and at least two AI runs with different settings. Bonus points if those are completely different AI programs.
  • Set a “House POV”: Create a short doctrine of what you believe, and use AI to stress-test it, not replace it.
  • Assign a contrarian: One reviewer’s job is to argue the opposite and find blind spots. This technique was common in debates in school, but gets lost quickly in corporate work.

And if you’re using a chatbot, try these prompts to add variety to your queries:

  • “Give me 4 conflicting strategies with tradeoffs; label the risks of each.”

  • “List 3 minority/contrarian views on this topic and evidence for each.”

  • “Rewrite this draft from a skeptical investor’s POV; note what’s missing.”

Law firms and banks turn to the courts to get M&A fees paid

Article from Financial Times.

Photo by John Guccione on Pexels.

Finance leaders should reread their engagement letters.

Circle is challenging a 2020 agreement with FT Partners that—depending on how you read the trigger—could imply a 10% of enterprise value payout if Circle is sold. After its post-IPO surge, that math lands near $5bn. Elsewhere, Quinn Emanuel is seeking about $30m after litigation that pushed a reluctant buyer to close, and Wachtell is in arbitration over a $90m success fee tied to the Twitter/X deal.

Stressful M&A + fuzzy triggers can mean courtroom finance. Success and tail clauses can outgrow the perceived value of the underlying work. If you want to avoid a high fee later, make sure your contract is precise, set a base and cap, spell out payment mechanics, and require detailed billing transparency.

If your deal pencils out to $5 billion, maybe it’s time to reassess.

IPA Data Dive: Data Reveals The CPA License Is Losing Ground

Article from Inside Public Accounting.

Photo by Pixabay on Pexels.

The share of staff with CPA licenses fell from 56.0% to 48.4% between 2020–2024. At larger firms it’s even lower (41.5%). The drivers are fewer grads pursuing CPAs, the 150-hour hurdle, and firms expanding into advisory where different skills matter. States like Georgia and Minnesota are piloting alternative pathways, which are signals that the traditional model is straining and things are starting to change.

What this means for leaders:

  • Separate “must-have” from “nice-to-have.” Keep CPA-required roles crystal clear; open other seats to finance, data, and ops talent.
  • Support multiple pathways. Reimburse CPA prep where it matters; recognize alternative credentials where it doesn’t.
  • Build dual career ladders. One oriented to assurance/compliance; one to advisory/ops, both with real ceilings.

The CPA still anchors trust. But the winning teams blend licensed expertise with broader operational talent.

Other articles we are reading:

Pentagon should boost fees for rocket launch companies, audit says, from AL.com.

UK is now a top investment destination, Deloitte survey reveals, from The Times.

Move Over Private Equity, the IPO Rush Is Here (Maybe), from Going Concern.

Audit Prep

At Imperial, we don’t just support your audit—we help you succeed at it.

Our experienced team works behind the scenes to clean up financials, fix reporting issues, and build the documentation you need to get through audits smoothly. Whether you’re onboarding a new client, catching up before an audit, or trying to avoid future issues, we make sure your books are in order and your team is prepared.

🧾 Audit-ready books

🔍 Clean reporting and documentation

✅ Confidence going into your next exam

Let’s take the pressure off your audit.

Visit our Audit Preparation Services page to learn more, or give us a call: 713-970-1163.

Alexander Brokhin

Founder & CEO, Imperial

alex@imperial-texas.com

“I am extremely proud of the talent behind the firm. I work with an incredible group of professionals and look forward to the continued growth and success of our company. I am humbled by the loyalty and wonderful relationships we have with our clients. Thank you for allowing us to support you.”