Books & Bots Advisory

Financial Controller-turned-builder. I automate month-end close for small businesses with AI + NetSuite. Faster closes, cleaner books, fewer late nights. 📚🤖

[Books & Bots Advisory @BooksAndBotsAI]

I have spent 21 years in accounting. Now I help small to mid-tier businesses put AI to work without the hype. Expect practical tips, real examples from the close, and the occasional lesson from getting it wrong. https://t.co/I8yb9cSxBo

[Books & Bots Advisory @BooksAndBotsAI]

@debs_obrien @bot Haha this is pretty cool just fired off excited to see what it comes back with.

[Books & Bots Advisory @BooksAndBotsAI]

@BoucherNicolas Having conversations with Claude in general and working with it as much as possible is great. It helps Claude build context on business flows. It just gets better and better.

[Books & Bots Advisory @BooksAndBotsAI]

@cbarmorecpa Making apps is a good idea I've never thought of. Typically I've been using it to design HTML dashboards for our team, but I'll have to give more thought to the app side as well.

[Books & Bots Advisory @BooksAndBotsAI]

Manual payment entry in AR usually means someone keying in every payment and applying it invoice by invoice. MCP Connectors for Microsoft, Netsuite and Plaid can semi-automate this process.

[Books & Bots Advisory @BooksAndBotsAI]

@__douglasgreen The biggest lesson I think is it doesn't have to happen all in one sitting. When Claude produces a balance sheet reconciliation, do your own flagging of issues and explain to Claude why it missed items. Over time it, it will get better.

[Books & Bots Advisory @BooksAndBotsAI]

Agreed, and context is what ends the guessing. The more time Claude Cowork spends inside your books, the more it learns how your organization actually works. Month one, you're spelling out every standard. By month three, it gets much better at it provided you are correcting along the way.

[Books & Bots Advisory @BooksAndBotsAI]

We've built about 50 Claude skills across our month end close. The hard part wasn't the AI, it was teaching it what wrong looks like in our books. Here's how we did it in two phases. https://t.co/ii96i6YmyL

[Books & Bots Advisory @BooksAndBotsAI]

@MayankWadhera True, and the quiet ones are different in every business. One company's is prepaids, another's is intercompany, another's is a vendor that bills in arrears. That's where Claude shines. You write your specific traps into the skill, and it adapts as new ones turn up.

[Books & Bots Advisory @BooksAndBotsAI]

Honestly, most of them don't need weekly tweaks. We adjust them as things come up: an edge case we didn't think of, a new account, or a step we can tighten. The first version gets it working usually and the updates are as hoc as needed. Many of the skill don't require tweaks at this point.

[Books & Bots Advisory @BooksAndBotsAI]

@coreyganim The outcome that sells best in finance is visibility. I use AI to build dashboards that show a business exactly where the money is leaking: vendor creep, duplicate spend, margin by product. Owners don't buy automation. They buy the moment they see the savings.

[Books & Bots Advisory @BooksAndBotsAI]

@williamkast_ Same on the finance side. I run most of month-end through Cowork now: recons, flux commentary, accrual support. And it keeps getting better, because every close adds more context about how our books actually work. Month 3 is a lot sharper than month 1.

[Books & Bots Advisory @BooksAndBotsAI]

21 years in accounting and the part of close I've always hated is the same stuff every month. Tying out recs, pulling flux explanations, chasing the same accruals.

I've been using Claude Cowork with NetSuite to knock that out first, so my time goes to the numbers that actually need a controller's judgment. The bots do the repetitive work, a human still signs off on every number.

[Books & Bots Advisory @BooksAndBotsAI]

As someone who is a NetSuite power user and uses Claude alot, the reporting aspect of Claude is extremely good. Mind blowingly good. We're using it to prepare all balance sheet account recs, month end documents, presentations, etc. I think when it comes to AI there's the reporting lane which it is very good at and then there is the transactional lane which it's getting better at but still work to do.

[Books & Bots Advisory @BooksAndBotsAI]

@xburak This post is so accurate. I spend my entire day automating different accounting and finance workflows. I think a lot of people think everyone just knows this stuff know and it wouldn't be further from the truth.

[Books & Bots Advisory @BooksAndBotsAI]

This is a pretty cool concept. Client owns the bot, firm sends instructions agent-to-agent, no portal password pile.

We’ve been thinking about the same boundary for Books & Bots Advisory: humans on money and exceptions, bots on gather / file / draft. Standing the client bot up is the hard product.

How are you handling the AP side for the client in this setup? Does one of the bots monitor an email address for incoming vendor bills and post them?

[Books & Bots Advisory @BooksAndBotsAI]

The manage / decide / do split is the useful part. Most people burn the expensive model on every yes/no, then wonder why the agent dies midweek.

For finance work that maps clean: code or a cheap gate for routing, a Bot for the clicking and filing, and a human for money, contracts, and anything irreversible. I’ve been building Books & Bots Advisory around that same boundary for close and controller work. AI clears the grunt. Judgment stays human.

Curious where you’ve seen the gate fail first in real ops.

[Books & Bots Advisory @BooksAndBotsAI]

This matches what I see on the controller and close side too. A lot of mid-tier firms already have books and a CPA. What’s missing is someone who can turn the ledger into cash visibility and decisions without a full-time CFO seat.

The $500k all-in cost is the real bind. Fractional (or fractional plus AI on the busywork) is how those companies buy judgment in doses they can actually afford.

I’ve spent 20+ years in accounting and finance. The AI piece helps most when it clears reconciliations and close grunt work so the human time goes to the forecast and the “what do we do next” questions.

Curious how you’re scoping first-90-day diagnostics versus ongoing retainer work.

[Books & Bots Advisory @BooksAndBotsAI]

A 13-week cash forecast for a mid-tier company gets hard fast. Open sales orders coming in. AR collecting. AP going out. All of it has to sit in one picture and the hard part isn't the early weeks. It's the handoff, when those balances roll off and the forecast has to carry the rest.

I've built these myself. Early weeks lean on open AR, AP, and open orders. Then those balances fall off, and the rest of the window hangs on forecast. That handoff is where most models quietly go wrong.

If you want help tying that forecast to real collections and payables, Books & Bots can walk through it with you. https://t.co/bdkIGCA48l