AI in Media & Broadcast – Professional Essentials Guide

12

NEWSCASTSTUDIO.COM

viewers which is often used alongside the

viewers preferences when they first on-

boarded, to curate personalized content

recommendations. This includes what

the viewer searched for, what videos they

played and at what time of day, whether

they watched to the end, how long it took

them to watch, how long each viewing ses-

sion lasted, as well as video ratings and

popularity. Put all of this together and mix

in some clever algorithms, and video ser-

vices then know which movies and shows

the viewer will enjoy watching the most.

Unfortunately, it’s not quite that sim-

ple because the way that viewers engage

with content is incredibly nuanced. Differ-

ent viewers like different things, and this

goes much deeper than different genres

or types of content. For example, out of an

audience who all like a certain genre, let’s

say, action and adventure, the same trail-

er may not necessarily hook in all view-

ers. What appeals to one consumer may

not necessarily appeal to another. Video

providers can attract a wider variety of

viewers to a particular show if the trailer

is personalized to appeal to their individ-

ual tastes. This is an area that AI is driving

forward, with new software able to create

custom images, videos and trailers to sup-

port the decision making process.

For example, if a viewer sees a trailer

featuring their favorite actor, that will ap-

peal to them most, while another viewer

may not be interested in that actor but may

prefer shows that have a strong female

lead, so a different trailer may work bet-

ter for them. By showing different viewers

trailers that are personalized to appeal to

their individual preferences, providers are

much better able to encourage viewers to

watch content. Similarly, thumbnails can

also be personalized in much the same

way as trailers can, in order to have the

most impact. This process involves choos-

ing the single frame from a show or movie

that is most likely to resonate with the in-

dividual viewer and using it as the thumb-

nail. By selecting the frame that best cap-

tures their interest, the platform increases

the chances that viewers will click on the

title and start watching.

Impact of AI on personalization

Before the advent of AI, video services

relied heavily on static user profiles, lim-

ited metadata, and manual tagging of con-

tent for personalization. With AI, the pro-

cess is faster and more accurate, which

makes it both more efficient and effective.

Using AI and ML, providers can identify

nuanced patterns in user behavior, which

enable dynamic and adaptive personaliza-

tion that evolves with the user’s preferenc-

es in real time.

AI vastly speeds up the processes that

enable personalization. Let’s consider

tagging as an example. For content rec-

ommendations to be accurate, video pro-

viders need to understand and know the

content of shows and movies down to the

minutia detail. In the past, this would have

involved teams of staff, some in-house and

some freelance, sitting down and watch-

ing every show in detail then accurately

tagging it.

Yet these days, with AI, the entire key-

word tagging process can be automated.

And additionally, it can also add a great

deal of detail which would not have been

practical when the process was done man-

ually. Detailed keyword tagging helps vid-

eo services to better identify similarities

between content so that recommenda-

tions can be more diverse but still be ac-

curate and appeal to viewers. In this use

case, it’s easy to see that AI not only saves

time and effort, but also makes the process

more effective.

Beyond personalization, AI plays a cru-

cial role in enhancing the user experience

by optimizing streaming quality. Take Net-

flix, for instance, which leverages machine

learning algorithms to monitor network

conditions in real time and dynamically

adjusts video quality. By assessing factors

such as bandwidth, device type, and geo-

graphic location, Netflix ensures the best

possible video resolution while reducing

buffering and interruptions. This seam-

less performance not only keeps users

engaged but also encourages them to stay

on the platform because loading times are

fast and playback is smooth.

What will personalization mean in

tomorrow’s world?

Effective personalization is fundamen-

tal to a video provider’s success. Without

it, the viewer won’t be able to easily find

content they want to watch, and addition-

ally the content that they may enjoy may

not appeal to them at first glance. AI is en-

abling video services to deliver faster and

more precise recommendations with an

improved user experience. Already AI can

help providers to infer mood and deliver

emotionally and situationally appropriate

recommendations. The question is, how

might this develop and improve as time

goes on and AI technology continues to

advance?

It will most likely become even more pre-

cise and intuitive as AI systems are used to

analyze subtle behavioral cues like voice

tone and biometrics to understand the us-

er’s mood and anticipate their preferenc-

es. I expect the content creation function-

ality of AI will continue to advance, even

being able to create unique content expe-

riences, such as different opening scenes

or alternate endings, designed to provide

each viewer with the experience that they

will enjoy the most. I have a feeling that we

are only at the beginning of what AI-driven

personalization can deliver — it’s going to

be an exciting ride for sure.

A customer experience expert, Nikki

has been with Accedo for more than 8

years and has driven some of our largest

UX engagements across the APAC region.

She is skilled in UX & UI, Strategy and

Research. She leads Accedo’s team of

designers in Asia Pacific. When not at

Accedo you can find her on the Rugby

League field, passionate about playing and

coaching young women in the sport.

Continued from previous page

Yet these days, with AI, the entire keyword tagging

process can be automated. And additionally, it can

also add a great deal of detail which would not have

been practical when the process was done manually.

Industry experts interested in contribut-

ing columns on broadcast industry topics

can email editor@newscaststudio.com

with a brief pitch.

For more details and terms and condi-

tions of contributing, please see Page 2

of this magazine.

CONTRIBUTE A COLUMN