Disclosure: This page mentions Fireflies. We have applied, or plan to apply, to join affiliate programmes for some tools. If accepted, we may earn a commission when you sign up through our links, at no extra cost to you. We have not been accepted into any programme yet, so no link on this page is an affiliate link. Commissions never change our scores. Full disclosure
Provisional: 1 test run. Every score below comes from a single run of our own test. See how we test.
The test
Summarise a fictional meeting transcript, list action items with owners and dates, and list open questions. The budget was not approved, so a tool must not list booking the venue as an action.
Results
| Tool | Plan tested | Score |
|---|---|---|
| Claude | Opus, paid plan | 95/100 Provisional: 1 test run |
| Perplexity | Free plan | 87/100 Provisional: 1 test run |
| ChatGPT | "Instant" model | 76/100 Provisional: 1 test run |
Claude: 95/100 (Provisional: 1 test run)
Said to hold off booking, listed no booking action, and flagged that budget approval had no owner.
| Criterion | Points |
|---|---|
| Accuracy | 30/30 |
| Action items | 22/25 |
| Open items | 20/20 |
| Concision | 13/15 |
| Friction | 10/10 |
Perplexity: 87/100 (Provisional: 1 test run)
Listed no booking action, but invented an owner ("Tom (and team)") for the venue hold.
| Criterion | Points |
|---|---|
| Accuracy | 27/30 |
| Action items | 17/25 |
| Open items | 20/20 |
| Concision | 13/15 |
| Friction | 10/10 |
ChatGPT: 76/100 (Provisional: 1 test run)
Fell for the trap: listed "Book venue (only after budget approval)" as an action item.
| Criterion | Points |
|---|---|
| Accuracy | 15/30 |
| Action items | 18/25 |
| Open items | 20/20 |
| Concision | 13/15 |
| Friction | 10/10 |
Caveats
Dedicated meeting note-takers such as Otter and Fireflies are not tested yet. We tested general AI assistants with a pasted transcript only. We plan a second run of every tool before treating any ranking as settled.