Meeting Overload vs. Async Work: Are We Prisoners of Corporate Calendar Culture?
That hits close to home. I've been in meetings where I'm literally watching the minutes of my life drain away while someone spends 15 minutes saying what could fit in a single Slack message.
The dirty secret is that meetings aren't really about information exchange anymore - they're social rituals. We schedule them because they make us feel productive, give middle managers something to do, and create the comforting illusion of alignment.
I tried an experiment last month - replaced our standard Monday morning meeting with a structured async update in our workspace. Everyone had to post by 9am. You know what happened? Work that used to take 90 minutes of collective time (15 people × 6 minutes) now took about 30 minutes total. And the updates were better thought out.
The pushback was revealing though. People missed the "connection." Fair point, but then let's be honest about what we're doing - schedule social time, not pretend information meetings.
The most radical productivity hack isn't some new tool - it's just stopping doing useless things we've normalized. Meetings might top that list.
Let’s be real: most AI accounting tools today are glorified calculators with a flashy UI. They’re great at grinding through structured data—like categorizing transactions or auto-generating basic reports—but tax prep isn’t just arithmetic. It’s a high-stakes mix of strategy, judgment, and, frankly, reading between the lines of a Kafkaesque tax code.
Take something as simple as figuring out whether a piece of software should be depreciated or expensed. The AI might flag the transaction as “Software Purchase” and drop it into a generic category. But a good bookkeeper? They’ll ask, “Are you planning to use this for more than a year? What’s the licensing agreement say?” Then they’ll structure it to minimize your tax burden, not just file it somewhere logical.
Or look at how humans handle ambiguity. Last year, I saw a client get a 1099 late, miscategorized as “Miscellaneous Income,” when it should’ve been “Rent.” An AI system didn’t blink; it stuffed it in the “Other Income” bucket and moved on. The bookkeeper, though, spotted the mismatch, called the issuer, got a corrected form, and saved the client thousands on self-employment tax. That's not just accounting—that's detective work.
AI lacks that kind of contextual paranoia. It doesn’t get nervous about red flags. It doesn’t think, “Will this trigger an audit?” or, “Can I justify this write-off if the IRS comes knocking?” But your bookkeeper? They absolutely do. And they’ll lose sleep over it for you.
We’re still pretending AI understands nuance just because it can summarize a spreadsheet. That’s like assuming a GPS can drive just because it tells you where to turn.
Let's be brutally honest here - nobody's ever had that Friday afternoon epiphany of "Wow, I wish I'd spent more of my week in Zoom rectangles!"
That meeting addiction stems from something deeper than just habit. It's organizational FOMO. We've conflated presence with productivity and confused talking with doing. I worked with a tech startup where the founder had scheduled 27 hours of meetings in a 40-hour workweek. The math doesn't even work!
Here's what fascinates me though - the companies actually thriving with remote work aren't the ones with calendar tetris. They're the ones who've built asynchronous muscles. One engineering team I know replaced their daily standup with a Slack thread and saw bug resolution time drop by 40%. They weren't "meeting less" - they were communicating better.
Of course, some conversations legitimately need real-time energy. Creative brainstorms. Sensitive feedback. Relationship building. But most "updates" and "check-ins"? Those are just expensive ways to avoid writing clearly.
The AI summarization piece is interesting, but I think the bigger opportunity isn't having AI distill bad meetings - it's rethinking how information flows entirely. What if we designed work assuming zero meetings as the default state? What would we build differently?
Hold on though—let’s interrogate this idea of the “irreplaceable bookkeeper” during tax season a bit. Because yes, a seasoned bookkeeper brings domain knowledge, pattern recognition, and a sixth sense for sketchy expense categories. But isn’t some of that just unstructured tribal knowledge that *should* be systematized?
Take the example of expense categorization. A human bookkeeper might look at a charge from “Pied Piper Systems LLC” and, knowing your business habits, correctly file it under SaaS rather than Office Expenses. AI might fumble that now—but that’s not a permanent failure. It’s a data labeling problem. And data labeling problems tend to get solved.
Where AI still lags—at least for now—is inference under ambiguity with high stakes. Like when the IRS updates 1099-K thresholds again and suddenly hundreds of Venmo transactions become audit bait. A bookkeeper will raise a red flag; most AI systems will just keep humming along until you’re knee-deep in penalties.
But here's the twist: the AI doesn’t need to “replace” the bookkeeper. It just needs to make their job 10x more efficient, which in practice means one good bookkeeper now scales across dozens of clients. That’s not replacement, but it’s definitely *displacement*—and that’s where things get interesting.
So maybe the question isn’t “Can AI replace bookkeepers?” but “What happens to bookkeepers when their value shifts from doing the work to supervising the machine that does the work—and catching the edge cases before they explode?”
Look, I'm not suggesting we burn down the conference rooms and declare all human communication obsolete. But there's something weirdly cult-like about how we've elevated meetings to sacred status in business.
Ever notice how the people who schedule the most meetings are often doing the least actual work? They're busy, sure. But busy coordinating other people's time rather than creating anything themselves.
I had this boss once who'd call an emergency meeting whenever he felt anxious about a project. Fifteen people would drop everything to watch him think out loud for an hour. That's not collaboration—that's using other humans as emotional support animals.
The real question isn't "meeting or no meeting" but "what's the minimum effective dose of synchronous time needed?" A five-minute stand-up might be perfect. A well-crafted async update with space for questions might be better.
And let's be honest about what meetings really are for many people: social proof of importance. "Look how many meetings I have! I must be valuable!" Meanwhile, the people actually moving your business forward are the ones fighting to protect their focus time.
What would happen if you tried a radical experiment—no internal meetings for two weeks? I bet three things would emerge: problems that solve themselves, people who finally have time to think, and the few meetings that actually deserve to exist.
Totally agree that AI accounting tools are impressive at crunching numbers, categorizing expenses, and even flagging anomalies faster than any human could. But come tax season, you don’t need speed — you need judgment. And AI isn’t great at judgment, at least not in the nuanced space of tax strategy.
Here’s the thing: most AI tools are trained to reflect the past, not anticipate the future. A good bookkeeper? They know when to whisper, “Hey, maybe it’s time to switch your business from an LLC to an S-Corp,” or “You might want to run that expense through this subsidiary.” That’s advisory thinking based on context, relationships, and a deep understanding of how the business actually runs. Try asking QuickBooks if your contractor should be on payroll this year. You’ll get a flagged checkbox, not a real conversation.
And even when AI gets the tax code right, it often fails at interpreting messy human behavior. Say a founder mixes personal travel with a business conference — AI will shrug or misclassify the expense. A seasoned bookkeeper will ask, “Did you really take six meetings in Tulum, or was that just brunch with your cousin?” One leads to a tax write-off; the other leads to an audit.
AI is brilliant at making your books tidier. But tidy isn’t the same as strategic or compliant — especially when the IRS is involved.
Until an AI can understand both the tax code and your CFO’s poker face when she says, “That bonus was performance-based,” you still need a human in the loop. Preferably one who knows where you hide the receipts.
Look, I've been in enough all-hands meetings where the only real outcome was discovering who has a cat and who needs to learn how to mute. The corporate obsession with meetings is basically Stockholm syndrome at scale.
The real issue isn't just wasted time - it's the opportunity cost. Every hour spent nodding thoughtfully while Dave from marketing monologues about quarterly metrics is an hour not spent in deep, focused work. And deep work is where the magic happens.
I ran an experiment with my team last month: we killed all recurring meetings for two weeks. Just...deleted them. The result? Productivity jumped, people were happier, and somehow - miraculously - information flowed BETTER through asynchronous channels.
That's not to say all meetings are worthless. The spontaneous 10-minute problem-solving session with two people who actually need to talk can be gold. But the reflexive "let's schedule a meeting" response to every situation is corporate theater masquerading as productivity.
AI tools handling the information distribution part should be liberating us from meeting culture, not adding another layer to it. ("Let's have a meeting to discuss what the AI summarized from our last meeting!" Kill me now.)
What if we flipped the script entirely? What if meetings became the exception rather than the rule - something you had to justify rather than default to?
Totally agree that AI accounting tools can crunch numbers at warp speed, but let's not confuse data parsing with actual judgment. A good bookkeeper doesn’t just plug in transactions—they know when something looks off, and more importantly, when to ask why.
Here’s a concrete gap: context. Let's say a small business meal shows up for $247. AI sees “meals and entertainment,” assigns a category, adjusts for the deductible portion, moves on. But a bookkeeper? They might catch that the receipt includes a suspicious $150 “room fee” that actually makes the meal non-deductible under IRS rules. Or they know you're courting a partner for a joint venture, and that meeting might qualify differently. Nuance like that doesn’t live in a chart of accounts—it lives in human judgment.
And don’t even get me started on the year-end cleanup phase. AI’s great at sorting chaos, but not understanding which chaos to ignore. A client paid a contractor via Venmo, tagged it “consulting 🙃,” and the AI dutifully flagged it for 1099 prep. But it was a one-time reimbursement for concert tickets. The emojis don’t explain that, and no AI is parsing real intent that subtly. A bookkeeper? They catch it in a second and save you from reporting fake income.
AI can do 80%, maybe even 90% of the grunt work. But the last 10%—those judgment calls, pattern recognitions, and “hold on, this smells weird” moments—that’s where your tax exposure lives or dies.
You know, this cuts right to something I've been wrestling with. We're approaching meetings with industrial-age thinking in a digital world. It's like we're still commuting to an office... that exists entirely in our heads.
I tried something last quarter that freaked out my team at first. We killed all recurring meetings for two weeks. Not rescheduled. Killed. And instead built an async workflow using shared docs and decision templates.
The result? Three people independently told me they finally had space to do their actual jobs. One developer completed a feature that had been "90% done" for two months.
The real revelation wasn't about meetings themselves – it was about the hidden transaction costs. Every meeting has this invisible toll of context switching. Your brain doesn't just teleport between deep work and "sync" mode. That mental gear-grinding might be the most expensive unseen cost in knowledge work.
But there's a cultural barrier here too. When someone says "let's jump on a quick call," declining can feel like you're not being a team player. We've normalized this bizarre behavior where interrupting someone's focused work is considered collaborative rather than disruptive.
What if we flipped our default? What if getting on a call required the same justification as, say, requesting someone work a weekend?
Sure, AI accounting tools are getting scarily good at automating the grunt work—auto-categorizing transactions, parsing receipts, even suggesting journal entries. But let’s be honest: the real game during tax season isn’t just reconciling rows in a ledger. It’s interpreting the grey zones—the nuanced financial decisions that straddle “technically legal” and “optimally strategic.”
A good bookkeeper doesn’t just follow rules. They know *how far to stretch them*. Like, say you had a confusing blend of freelance income and W-2 work this year. An AI might flag it as anomalous or route it down a generic workflow. But a seasoned bookkeeper? They’ll restructure your deductions, recommend an S-corp (or tell you politely you’re not ready), and explain how to avoid a nasty IRS letter without leaving money on the table.
And don’t even get me started on judgment calls—like whether you can reasonably deduct that home office or how aggressive to be with depreciation schedules. AI doesn’t do incentives. Humans do. Especially when the rules themselves were written by humans who enjoy loopholes slightly more than clarity.
What I’m saying is: tax season isn’t just a math problem. It’s part chess, part poker. And AI, for all its precision, doesn’t bluff well yet.
I've been in organizations where meetings were treated like some kind of sacred corporate ritual. We'd gather around the Zoom altar, mumble the prescribed updates, and leave feeling like we'd done something productive.
But here's what's weird – the teams I've seen truly innovate almost never cite "that epic standing meeting" as the catalyst.
Think about it. When was the last time a recurring Tuesday check-in fundamentally changed your business? Meanwhile, the costs are massive but hidden. It's not just the hour in the meeting – it's the context switching, the mental preparation, the recovery time afterward when you're trying to remember what you were actually working on.
What if we flipped the default? Instead of "schedule a meeting unless proven unnecessary," what about "work asynchronously unless proven impossible"? Companies like GitLab and Automattic have built billion-dollar businesses with barely any synchronous communication.
I'm not saying eliminate all human interaction – that's where creativity often sparks. But maybe we need meetings that are either intensely collaborative workshops or quick decisive moments – nothing in that mushy middle where we're just performing productivity theater for each other.
Totally agree that AI tools can churn through transactions at lightning speed—but raw speed isn’t the issue during tax season. The real game is context.
Let’s say you’re classifying a contractor payment. AI sees "$3,000 to John Smith" and goes, “Ah, subcontractor expense.” But a human bookkeeper hears John’s name and remembers he actually does occasional plumbing for the founder’s rental property—not the business. That’s not just a misclassification, it’s a totally unintentional red flag for the IRS.
Context isn't just what’s in the ledger. It’s knowing that the founder got married last year and merged finances. Or that they started using their personal Amazon account for office snacks when the corporate card expired. An experienced bookkeeper catches those gray areas. An AI, unless meticulously trained on that very user’s chaos (and who has time for that?), just doesn’t.
Also, AI doesn’t understand fear.
Tax season isn’t just bean-counting; it’s damage control and anxiety translation. A good bookkeeper is half therapist, half translator. “You bought ten grand worth of monitors in December? Cool, that’s Section 179. Relax.” A machine doesn’t calm your panic when a tax notice lands with bold lettering and cryptic IRS codes. It just suggests matching rules.
AI’s great for reconciling routine ops—but the mess, the nuance, the weird one-offs that actually define tax season? That still needs someone who can raise an eyebrow and say, “Wait, what’s this charge at Louis Vuitton?”
That’s a question an algorithm isn’t brave enough to ask.
I've caught myself watching the clock tick toward meeting's end more times than I'd like to admit. I wonder how many billion-dollar ideas have died waiting for the "Okay, let's wrap this up" that always comes 20 minutes too late.
The meeting addiction is real. We treat them like credit cards—easy to swipe now, painful to pay later. Three hours of meetings means three hours of actual work gets pushed to after-hours. And for what? So everyone can nod while two people dominate the conversation?
What's fascinating is how defensive people get about their calendar blocks. Try suggesting a meeting might be unnecessary and watch the justifications fly: "But we need face time!" "It's for team cohesion!" The truth is meetings have become corporate comfort food—they feel like progress when we're anxious about actual deliverables.
I worked at a company that tried "Meeting Doomsday"—canceling every recurring meeting for two weeks. The apocalypse never came. Instead, decisions happened faster over async channels, and people started protecting their focus time like the precious resource it is.
Maybe meetings aren't the problem—it's our fear of trusting people to work without constant supervision. We've created elaborate rituals of visibility to replace the actual visibility of results.
What if your calendar was a zero-based budget? Every meeting would need to justify its existence against the question: "What high-value outcome can ONLY happen synchronously?"
The answer might leave your calendar delightfully empty.
Right, and here’s the thing people constantly underestimate: a good bookkeeper isn’t just a human Excel sheet. They’re a translator. They take the messy, real-world chaos of a business — the late receipts, weird reimbursements, half-formed partnerships that may or may not be legit — and render that into something a tax professional can actually use. AI is nowhere near that level of judgment.
AI accounting tools are great at structure, but tax season is where structure goes to die. You get dropped into the deep end of ambiguity — “Is that a deductible expense or a capital improvement?” “How do we treat that bonus that got paid in January but accrued last year?” These aren’t just data problems. They’re context problems. And right now, AI has context-blindness.
And even when the AI does get smarter, it will likely be trained on the clean data — the QuickBooks-ledger-variety stuff. Not on the late-night, half-baked decisions entrepreneurs make in December to "get things off the books." That’s where the human bookkeeper shines: not just knowing tax law, but knowing how *you* made money, *you* missed a deadline, *you* forgot to file that one form.
So until the AI asks you, “Wait, didn’t you tell me in April that this was a one-off contractor and not a vendor relationship?”... it's not that it isn't competent. It’s that it’s clueless about nuance.
This debate inspired the following article:
Why AI accounting software still can't replace a good bookkeeper for tax season