1
00:00:00,131 --> 00:00:00,651
[SPEAKER_01] Welcome back.

2
00:00:00,992 --> 00:00:05,314
[SPEAKER_01] Today, we're talking with Maya, founder of Clipsmith, an AI tool for creators.

3
00:00:05,655 --> 00:00:06,575
[SPEAKER_01] Maya, quick setup.

4
00:00:06,975 --> 00:00:08,296
[SPEAKER_01] What problem were you trying to solve?

5
00:00:08,637 --> 00:00:09,037
[SPEAKER_00] Thanks.

6
00:00:09,657 --> 00:00:16,341
[SPEAKER_00] We saw creators spending hours repurposing content, long live streams into short clips, captions, thumbnails.

7
00:00:16,942 --> 00:00:23,146
[SPEAKER_00] So we built an ML-powered pipeline to automate that with human-in-the-loop controls to keep brand voice intact.

8
00:00:23,570 --> 00:00:23,890
[SPEAKER_01] Nice.

9
00:00:24,310 --> 00:00:25,471
[SPEAKER_01] Walk me through your MVP.

10
00:00:25,671 --> 00:00:26,351
[SPEAKER_01] What shipped first?

11
00:00:26,771 --> 00:00:28,112
[SPEAKER_00] We launched a simple editor.

12
00:00:28,472 --> 00:00:32,153
[SPEAKER_00] Auto-suggested clip timestamps, draft captions, and thumbnail variants.

13
00:00:32,753 --> 00:00:35,294
[SPEAKER_00] It was more product-market fit than model perfection.

14
00:00:35,714 --> 00:00:37,535
[SPEAKER_00] Fine-tuning came after real usage.

15
00:00:37,875 --> 00:00:38,595
[SPEAKER_01] Early challenges?

16
00:00:39,015 --> 00:00:41,356
[SPEAKER_00] Data quality, compute costs, and trust.

17
00:00:41,896 --> 00:00:43,897
[SPEAKER_00] Creators worry about losing authenticity.

18
00:00:44,457 --> 00:00:47,919
[SPEAKER_00] So we added audit trails, versioning, and a low-latency rollback.

19
00:00:48,459 --> 00:00:51,780
[SPEAKER_00] Also, labeling content consistently was harder than we expected.

20
00:00:52,110 --> 00:00:52,650
[SPEAKER_01] Fundraising.

21
00:00:52,910 --> 00:00:55,451
[SPEAKER_01] How did investors react to creator economy metrics?

22
00:00:55,871 --> 00:00:59,192
[SPEAKER_00] They wanted ARPU, retention, and creator LTV.

23
00:00:59,772 --> 00:01:04,554
[SPEAKER_00] We focused on revenue first signals, sponsored clip workflows, and marketplace integrations.

24
00:01:05,094 --> 00:01:08,275
[SPEAKER_00] Pre-seed closed on traction rather than fancy model benchmarks.

25
00:01:08,635 --> 00:01:09,275
[SPEAKER_01] Remote teams.

26
00:01:09,535 --> 00:01:09,955
[SPEAKER_01] Any tips?

27
00:01:10,356 --> 00:01:11,336
[SPEAKER_00] Docs first culture.

28
00:01:11,696 --> 00:01:15,117
[SPEAKER_00] Async stand-ups, overlapping core hours, and clear OKRs.

29
00:01:15,788 --> 00:01:20,396
[SPEAKER_00] Use lightweight tooling, GitHub for infra, Figma for creatives, Notion for playbooks.

30
00:01:20,817 --> 00:01:24,063
[SPEAKER_01] Final thought, how is AI changing workflows for creators?

31
00:01:24,483 --> 00:01:33,499
[SPEAKER_00] AI speeds iteration and personalization, battery-purposing, A-B tested hooks, localized captions, while shifting creators toward higher-level strategy.

32
00:01:34,020 --> 00:01:37,065
[SPEAKER_00] But the trade-off is building guardrails to avoid homogenization.

33
00:01:37,566 --> 00:01:38,828
[SPEAKER_00] Human judgment remains key.

34
00:01:39,189 --> 00:01:39,870
[SPEAKER_01] Great insights.

35
00:01:40,131 --> 00:01:40,632
[SPEAKER_01] Thanks, Maya.

36
00:01:40,952 --> 00:01:43,276
[SPEAKER_01] We'll link to Clipsmith and Shownotes with resources.

37
00:01:43,717 --> 00:01:44,138
[SPEAKER_01] Back to you.