AI Ready
ArchiveDaily Brief
Subscribe
Anthropic just put Claude inside Excel
Daily Brief

Anthropic just put Claude inside Excel

Anthropic shipped ten ready-to-use finance agents and put Claude inside Excel and PowerPoint. Here's what changed.

By Haroon Choudery·May 6, 2026·8 min read

THE AI BRIEF

Today's signal: Anthropic shipped ten ready-to-use finance agents and put Claude inside Excel and PowerPoint. Here's what changed.

In today’s issue:

  • Main story: Anthropic just shipped the work the consultants were going to bill you for

  • Also worth knowing: OpenAI/Anthropic acquisition talks, Gemma 4 speed gains, Chrome's silent AI install, and browser agent pricing

  • Free webinar: Build Notion Agents to Automate Complex Tasks, a 30-minute Lightning Lesson, May 14 at 2 pm ET. Save your seat →

THE READ

Anthropic released ten ready-to-use agents for finance teams this morning. They cover the work a CFO's office actually does: pitchbooks, KYC checks, month-end close, ledger reconciliation, valuation review, audit, model building, earnings review, market research, and meeting prep. Claude also now runs inside Excel, PowerPoint, Word, and Outlook, so an analyst can use it without opening a new app.

This lands a day after Anthropic and OpenAI announced their joint ventures with private equity to staff AI rollouts at large companies. Yesterday, the labs hired the people who do the install. Today, they shipped the install itself, prebuilt.

A Big Three consultant would scope a pitchbook tool as a six-figure project, sell it once, and hand over a binder. Anthropic is selling the same finished tool as a $20-a-month seat, and they sell it again every month to every customer. That is a different business than consulting, and it is now competing for the part of the work that consultants used to own.

AI READY IN 21 DAYS · FREE

A 21-day AI program. In your inbox. Then it ends.

One email every morning at 7 am ET. Each one is a short read and one real thing to try before lunch. By Day 21, you have nine concrete capabilities, including prompting that hits your bar on the first draft, AI workflows that take five hours a week off your plate, and the language to lead the AI conversation at your company.

Built for ops leaders, COOs, chiefs of staff, founders, and team leads at mid-market companies. The four weeks move you from foundations (effective prompting, catching hallucinations, learning AI with AI) to connect (a personal context layer, daily automation, five hours a week back), to build (reusable skills, agent design and debugging), to lead (tool and model evaluation, leading the conversation at work). Twenty-one mornings, no upsells, no community Slack.

Not sure if it is for you? Take the 2-minute diagnostic.

Where this could be wrong

The biggest objection is that real finance work does not fit a template, because every bank does things differently and every PE fund builds models its own way. A general-purpose pitchbook agent will hit a wall the first time a partner says but we do it this way.

That is fair. It is also why Anthropic shipped these as starting points with named connectors into FactSet, S&P Capital IQ, MSCI, PitchBook, Morningstar, LSEG, and Daloopa. Whether starting from a template and adapting it actually costs less than starting from scratch is the question every firm has to answer for itself. The math probably favors the template for most mid-market firms. The math probably still favors a custom build for the largest banks. Where the line falls is the thing to watch.

What this means if you run operations

Two things worth tracking, not two things to do today.

First, if your finance team has been waiting on a budget for a six-figure AI install, the offer space just changed. Whether the templates work well enough to replace a custom build is genuinely unclear right now, but the option of putting Claude on month-end close for the cost of one new seat now exists where it didn't yesterday. Worth knowing it exists before someone walks you through it.

Second, if you bought the six-figure custom install last year, the renewal conversation is going to be different. Whether your vendor has a real answer for why custom is still worth the markup, or whether they're going to wave at brand and security and hope you don't push, is the actual signal worth watching. The question is not whether to renew. The question is what your vendor says when you ask them about the templates by name.

SPONSORED BY CLUTCH

Hire secure AI teammates that work 24/7.

Hire pre-built AI teammates. Give your engineers and operators a platform to ship their own AI apps. Stop losing sleep about what is running where.

Clutch is the platform behind both: pre-built agents for the workflows your ops team should automate first, plus the integration plane your team's vibe-coded apps and Claude Code projects plug into. One platform. Real production. Visible and safe by default.

Built for ops, engineering, and security teams that are tired of the shadow-AI surface area inside their own company.

ALSO WORTH KNOWING

The same Anthropic and OpenAI ventures may buy AI consulting firms outright
Reuters says both groups are in early talks to acquire implementation firms instead of building deployment teams from scratch. The labs see staffing as the bottleneck before the new money becomes useful.

Google made Gemma 4 run three times faster
A new drafter model lets Gemma 4 generate three times the words per second with no loss in quality. For any team running open-source AI on its own hardware, this is the upgrade that turns a slow local agent into one that actually keeps up.

Chrome quietly downloaded a 4GB AI model onto user laptops without asking
A privacy researcher caught Chrome installing Google's Gemini Nano model on every machine running the browser. The file comes back if you delete it. If your IT team manages Chrome at scale, this is the first AI question your security lead is going to bring up.

Browser-based AI agents cost 45 times more than the simpler kind for the same task
Reflex tested two ways to do the same admin job. The agent that clicks around like a human took 14 minutes and 500,000 tokens. The agent who called the underlying tools directly took 8 calls. Useful price tag the next time a vendor pitches you an AI that can use any software.

P.S.

If your finance team is the one feeling the most pressure this quarter, hit reply and tell me which workflow your CFO would hand to an agent first: close, KYC, valuation review, or something else. I read every one.

FORWARD THIS

If a colleague is about to scope an AI install for finance, forward this. The offer space changed this morning, and the scope they had on Friday is not the scope they have today.

Back tomorrow,
Haroon

Enjoying this issue?
Get the next Daily Brief as it lands.
Free. Four sends a week.
Keep reading

Get every issue, as it lands.