AI tools cut the drafting time on comps memos and market summaries by half, but none of them will catch a broken formula reference in your own model, so you still check every number by hand.
ChatGPT, Claude, Perplexity, and Excel Copilot were tested on core financial analyst tasks, from comps pulls to model review. This AI for financial analysts verdict was issued in August 2026.
- ✓First-draft company summaries from 10-K filings and earnings calls
- ✓Fast comps pulls with cited multiples via Perplexity
- ✓Formula explanation and writing help through Excel Copilot
- ✓Market and macro summaries condensed for morning notes
- ×Model auditing misses hardcoded values and circular references
- ×Financial figures cited without period or source attached
- ×No cross-session memory of model assumptions or conventions
Scorecard
Drafts are fast but drop nuance, and cited figures lack period context
Fits the research and drafting stage well, not the review stage
Every session starts from zero with no recall of prior assumptions
Quick to adopt for summaries and formula help, payoff is immediate
Errors look polished and reviewed, making silent failures hard to spot
The missing tool is persistent model memory: a system that attaches assumptions, prior review comments, and source citations to the workbook file itself, so each session inherits full context instead of starting blind.