1
00:00:00,720 --> 00:00:03,900
What are small businesses actually
doing with generative AI

2
00:00:03,930 --> 00:00:06,180
and what changes when they do?

3
00:00:06,270 --> 00:00:11,490
In a 2024 OECD survey of over
5,000 SMEs in seven countries,

4
00:00:11,520 --> 00:00:16,430
30.7% reported work-related use
by the respondent or a colleague.

5
00:00:16,500 --> 00:00:18,820
Among those users, text leads:

6
00:00:18,850 --> 00:00:21,200
91.6% generate text,

7
00:00:21,230 --> 00:00:25,580
34.9% images, 16.9% video,

8
00:00:25,610 --> 00:00:28,110
and 14.7% audio

9
00:00:28,140 --> 00:00:29,300
Multiple answers allowed.

10
00:00:29,730 --> 00:00:31,490
The examples are practical.

11
00:00:31,520 --> 00:00:33,940
Restaurants prepare menus and edit images.

12
00:00:33,970 --> 00:00:37,980
Construction firms produce video
simulations of architectural models.

13
00:00:38,010 --> 00:00:42,600
Information and communication firms
develop, debug and simplify code.

14
00:00:42,630 --> 00:00:46,160
But only 28.7% use it in core activities.

15
00:00:46,620 --> 00:00:48,860
Reported benefits centre on people:

16
00:00:48,900 --> 00:00:52,440
65.1% say employee performance improved.

17
00:00:52,680 --> 00:00:58,160
45.2% saved money, and
25.9% report higher revenue.

18
00:00:58,290 --> 00:01:01,890
These are self-reported outcomes,
not proven causal effects

19
00:01:01,920 --> 00:01:05,100
65.1% is the share reporting an improvement,

20
00:01:05,130 --> 00:01:08,100
not a 65.1% productivity gain.

21
00:01:08,550 --> 00:01:11,050
The picture is uneven but concrete:

22
00:01:11,220 --> 00:01:15,020
many small businesses already use
generative AI in everyday work,

23
00:01:15,050 --> 00:01:18,330
while adoption and reported
benefits vary widely.

24
00:01:18,360 --> 00:01:20,270
That gap is the story.