The future of enterprise software isn’t a better dashboard. It’s a conversation.
This week, Bret Taylor — former co-CEO of Salesforce, former CTO of Facebook, and now CEO of Sierra AI — released something that should make every finance leader pay attention. It’s called Ghostwriter: an AI agent that builds other AI agents. No forms, no menus, no configuration screens. You describe what you want in plain English, and it creates it.
Want a customer service agent that handles returns, speaks French, and connects to your Shopify backend? Have a conversation. Done.
This isn’t a demo. It’s a signal.
The Pattern That Should Worry Software Vendors
Ghostwriter is part of a pattern that’s been accelerating since late 2025. Tools like Codex and Claude Code have already transformed software engineering. Developers no longer write every line of code — they orchestrate, review, and direct. The output is the same. The process is fundamentally different.
Taylor’s argument is that the same transformation is coming for all enterprise software. Not just development tools — everything. ERP, CRM, accounting platforms, treasury management systems. The entire stack.
His exact words: “Rather than every enterprise app having a web app for humans and an API for automation, every software platform’s UI will be an agent that can do the work on your behalf.”
Read that again if you’re a CFO still commissioning twelve-month ERP implementations.
What This Means for Finance
Let me make this concrete.
Today, your month-end close looks something like this: your finance team logs into three or four different systems. They run reports, download spreadsheets, reconcile manually, chase variances, prepare packs, and upload to the group reporting tool. The process takes five to ten working days. Most of the time is spent navigating software, not thinking about the business.
Now imagine the same process with an agent layer sitting on top of your existing systems. You say: “Run the month-end close. Reconcile the bank against the ledger. Flag anything over five thousand that doesn’t match a purchase order. Prepare the board pack in the usual format. Highlight the three biggest variances and draft commentary.”
The agent does it. You review, adjust, approve. The same output, a fraction of the time, and your team spends their energy on judgment — not data wrangling.
This isn’t science fiction. The individual components already exist. Bank reconciliation APIs, accounting platform integrations, document generation, natural language processing. What’s been missing is the orchestration layer — the thing that ties them together without requiring a human to click through each step.
That’s what Taylor is building. And he’s not alone.
The CFO’s New Job Description
This shift creates a new archetype: the CFO as orchestrator.
The traditional CFO role has always included a significant amount of operational work. Processing, reporting, compliance, controls. Important work, but fundamentally mechanical. The strategic work — capital allocation, commercial pricing, risk management, M&A evaluation — often gets squeezed into whatever time is left after the operational grind.
Agent-driven finance flips that ratio.
When the operational layer is handled by AI agents that you direct rather than do, the CFO role becomes primarily about judgment, strategy, and decision-making. You become the conductor, not the first violin.
This has implications for team structure too. The finance team of 2028 won’t necessarily be smaller, but it will be shaped differently. Fewer people doing transactional processing. More people doing analysis, business partnering, and — critically — managing and training the agents.
Because here’s the thing about AI agents: they’re only as good as the instructions they receive. A well-configured agent with clear guardrails and good data will outperform a team of ten doing manual work. A badly configured one will produce confident-looking nonsense at machine speed. The quality of the orchestration matters enormously.
The PE Angle
If you’re in private equity, this should be front of mind for every portfolio company review.
The finance functions in most PE-backed businesses are already lean. They’ve been through cost reduction, shared services optimisation, and system consolidation. There isn’t much more fat to cut through traditional means.
Agent-driven automation is a different lever entirely. It doesn’t just reduce headcount — it compresses cycle times, improves data quality, and frees up the finance team to do actual value-add work. A portfolio company that can close in three days instead of ten, produce real-time margin analysis instead of monthly snapshots, and have its FD spending 80% of their time on commercial strategy instead of reconciliation — that’s a more valuable business.
The sponsors who build agent capability into their value creation playbook will see it in their returns. The ones who treat it as “an IT thing” will wonder why their portfolio companies feel slow.
The Uncomfortable Question
Here’s the part nobody wants to talk about: if an AI agent can build another AI agent through a conversation, and that second agent can run your month-end close, what exactly is the long-term role of the mid-level finance professional who currently does that work?
I don’t have a comfortable answer. The honest one is that some roles will be eliminated, others will be transformed, and new ones will be created that don’t exist yet. The transition won’t be smooth, and it won’t be evenly distributed.
What I do know is that the finance professionals who learn to work with agents — to orchestrate, direct, and quality-check their output — will be dramatically more valuable than those who compete with them on operational tasks.
The CFO’s job isn’t going away. It’s being elevated. But only for those who recognise the shift and lean into it.
The Bottom Line
Bret Taylor didn’t just launch a product this week. He articulated a vision for how all enterprise software will work within the next three to five years. The web app era — with its forms, menus, dashboards, and configuration screens — is giving way to the agent era, where you describe what you want and the software figures out how to do it.
For finance leaders, the question isn’t whether this will happen. It’s whether you’ll be the one directing the agents, or the one being replaced by them.
Start learning now. The orchestrators are going to win.
Navigating the shift from operator to orchestrator? Whether you’re building an AI-ready finance function or evaluating agent capability in portfolio companies — get in touch.