We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·Aug 12·5 min read
Scaling accounting capacity with OpenAI
TL;DR
Accounting firms face capacity constraints that limit their ability to take on advisory work and grow. Basis built AI agents using OpenAI's o3, o3-Pro, GPT-4.1, and GPT-5 models to automate accounting tasks.
✦ Why It Matters
Engineers can apply multi-model AI agent architectures to automate professional services workflows and measure productivity gains quantitatively.
Key Takeaways
- Deploy specialized AI agents using latest LLMs to automate domain-specific repetitive workflows in professional services.
- Measure productivity gains quantitatively; target 20-30% time reduction on manual tasks as a baseline metric.
- Design agents to free high-skill staff for advisory and growth work, not just cost reduction.
- Evaluate multi-model approaches (o3, GPT-4.1, GPT-5) for different task complexity levels within workflows.
Full Summary
Accounting firms traditionally struggle with time-intensive manual tasks that consume resources needed for client advisory services and business growth. Basis developed AI agents—autonomous software systems that perform tasks with minimal human intervention—leveraging OpenAI's latest large language models (o3, o3-Pro, GPT-4.1, and GPT-5) to automate routine accounting workflows.
The agents handle repetitive accounting processes, reducing the manual effort required per engagement. Field results showed a 30% reduction in time spent on accounting tasks, enabling firms to reallocate staff toward higher-margin advisory work and capacity expansion.
This approach demonstrates how specialized AI agents can address domain-specific bottlenecks in professional services by automating lower-value work.