How AI is Making it Easier for Marketers to Create Videos

Campaign MarketingTechnology

There’s no ques­tion that arti­fi­cial intel­li­gence (AI) is con­tribut­ing to the evo­lu­tion of video con­tent. Video is arguably the go-to plat­form that the vast major­i­ty of busi­ness­es use for con­tent mar­ket­ing — and its growth will con­tin­ue to sky­rock­et.

That said, this pow­er­house con­tent type isn’t with­out its chal­lenges. For starters, pro­duc­ing qual­i­ty video con­tent requires a sig­nif­i­cant invest­ment in time, tal­ent, and costs of video pro­duc­tion — costs that can eas­i­ly eclipse oth­er con­tent devel­op­ment process­es and plat­forms.

Sec­ond, with 60 per­cent of con­sumers con­sid­er­ing brand-gen­er­at­ed con­tent as point­less clut­ter, con­tent cre­ators are under tremen­dous pres­sure to deliv­er engag­ing and high-con­ver­sion cus­tomer expe­ri­ences to an audi­ence who may or may not be so recep­tive.

For­tu­nate­ly, advances in AI and machine learn­ing are empow­er­ing mar­ket­ing teams to stream­line and accel­er­ate video pro­duc­tion work­flows and cre­ate more tar­get­ed con­tent, in less time and with few­er resources. Here are some of the ways Adobe’s AI-infused video cre­ation fea­tures are dri­ving suc­cess for mar­keters:

#1. Eliminating wasted time with intelligent search

For video cre­ators, the sheer vol­ume of footage col­lect­ed can become unwieldy. With giga­bytes of footage, wad­ing through and watch­ing video files to find the per­fect shot can take hours, if not days or even weeks. To lim­it this incon­ve­nience, many con­tent cre­ators sac­ri­fice hours of time man­u­al­ly tag­ging videos (with attrib­ut­es that describe the footage) — time that could be spent on con­tent devel­op­ment.

AI can help. At Adobe, our Video Auto Tag Adobe Sen­sei ser­vice auto­mat­i­cal­ly tags attrib­ut­es for each seg­ment of each video — think col­ors, faces, loca­tions, actions, prod­ucts, or cor­po­rate logos. AI pro­vides a search­able tax­on­o­my for find­ing not just the right video file, but the right cut of the video, in a frac­tion of the time.

Video Smart Tags for user gen­er­at­ed con­tent.

The Video Auto Tag process allows for nat­ur­al lan­guage search of video files, even when mul­ti­ple con­di­tions are present. For exam­ple, if you searched for a video con­tain­ing “two dogs play­ing in the water at the lake,” AI-pow­ered tools would retrieve videos meet­ing all five of those con­di­tions, pro­vid­ing an unprece­dent­ed lev­el of speci­fici­ty for video­g­ra­phers and edi­tors on tight time­lines.

#2. Predicting greater video memorability

If you’re cre­at­ing mar­ket­ing videos, you’re like­ly look­ing at insights from past videos to try to pre­dict future per­for­mance — or, pos­si­bly, you go with your gut to come up with a con­tent strat­e­gy that will get you the most views.

How­ev­er, both approach­es can be frus­trat­ing and time con­sum­ing, while yield­ing unpre­dictable results. This, again, is a chal­lenge that Adobe Sen­sei can help with — specif­i­cal­ly, the tech­nol­o­gy that we “sneaked” at Adobe Sum­mit last March: #VideoAd­AI

#VideoAd­AI is an explorato­ry tech­nol­o­gy that uses a deep neur­al net­work to auto­mat­i­cal­ly gen­er­ate meta­da­ta, such as cat­e­gories and tags for video ads. To do this, AI “reads” the ad’s video, audio, or oth­er avail­able infor­ma­tion and iden­ti­fies demo­graph­ic seg­ments and sub­jects rel­e­vant to a giv­en plat­form. This makes it easy to line up the right video ad with the right ad plat­form. #VideoAd­AI then employs a sec­ond deep neur­al net­work to assign a “watch­a­bil­i­ty” score to the ad.

Sim­i­lar­ly, the AI-dri­ven Auto Curate fea­ture pow­ered by Adobe Sen­sei in Pre­miere Ele­ments can view an entire video cat­a­log, iden­ti­fy the best footage, and sur­face it for use in a project, sav­ing video­g­ra­phers and edi­tors the has­sle of brows­ing through end­less file fold­ers of footage.

#3. Streamlining the post-production process

The tra­di­tion­al video pro­duc­tion process is filled with repet­i­tive tasks that can be eas­i­ly auto­mat­ed with AI. For exam­ple, many con­tent cre­ators devel­op a series of dif­fer­ent thumb­nails, flash cards, and sum­maries to match the vary­ing require­ments for dif­fer­ent video plat­forms like Face­book or Insta­gram.

With #Pro­jectSmooth­Op­er­a­tor, anoth­er exper­i­men­tal tech­nol­o­gy that we pre­viewed at Adobe MAX, mar­keters and edi­tors can auto­mate plat­form-spe­cif­ic video opti­mi­sa­tion. This elim­i­nates the need to man­u­al­ly opti­mize con­tent for dif­fer­ent social media plat­forms, mobile, tablet, and oth­er devices.

What’s more, using Adobe Sen­sei tech­nol­o­gy, Project Smooth Oper­a­tor exam­ines the video con­tent to gen­er­ate a heat map that syncs with each sub­ject, and its action, in frames. From here, a glob­al cam­era path is deter­mined, auto­mat­i­cal­ly shift­ing the frames to cap­ture the core action of each shot.

AI can also help with col­or cor­rec­tion, anoth­er time-con­sum­ing and cost­ly part of the video edit­ing process. Adobe Premiere’s AI-pow­ered Col­or Match fea­ture applies col­or cor­rec­tion set­tings from one shot to all the oth­ers in a scene, tak­ing spe­cial care to iden­ti­fy faces and match skin tones. Edi­tors, though, can still man­u­al­ly adjust col­or cor­rec­tion set­tings using the col­or wheels to ensure the right col­or every time.

#4. Seamlessly integrating video and audio

Few parts of the tra­di­tion­al post-pro­duc­tion process are as tedious as sync­ing audio with video. Whether it’s man­u­al­ly adjust­ing sound lev­els, tweak­ing the tim­ing of dia­logue to match actors’ mouths, or try­ing to bal­ance musi­cal and sound effects, the process can be extreme­ly time con­sum­ing and resource-heavy. Here, also, Adobe Sen­sei is improv­ing the process, while simul­ta­ne­ous­ly improv­ing video qual­i­ty.

The Morph Cut fea­ture in Pre­miere Pro auto­mat­i­cal­ly elim­i­nates unwant­ed jump cuts, removes unwant­ed paus­es or words, and makes each shot’s tran­si­tion seam­less. Sim­i­lar­ly, Premiere’s audio Auto Duck­ing fea­ture intel­li­gent­ly reduces the vol­ume of sound­track ele­ments to avoid drown­ing out the dia­logue.

More video, faster, at higher quality with Adobe Sensei

Thanks to Adobe’s cur­rent AI-pow­ered solu­tions, video cre­ators can keep pace with cus­tomers’ insa­tiable appetites for qual­i­ty con­tent. These tech­nolo­gies stream­line and sim­pli­fy video edit­ing with­out a major learn­ing curve, so edi­tors can focus on the cre­ativ­i­ty and let the tech­nol­o­gy help with the time-con­sum­ing, often menial, tasks.

Where we are with AI right now is just the begin­ning. Some of these Adobe Sen­sei pow­ered tech­nolo­gies are still in explorato­ry stages. There will cer­tain­ly be more inno­va­tions emerg­ing, poised to dri­ve video con­tent cre­ation even fur­ther. It’s a piv­otal moment in the video indus­try, and an excit­ing time for the cre­ative minds pow­er­ing it — and we’re here to keep inno­vat­ing, keep advanc­ing, and help ensure our cre­ative cus­tomers can see their visions real­ized.

Learn more about Adobe Sen­sei here, and read more arti­cles on the Adobe Blog about Adobe Sen­sei here.

 


Campaign Marketing, Technology
Tatiana Mejia

Posted on 05-16-2019


Comments

  • By Ekim Saribardak - 11:15 AM on May 18, 2019   Reply

    I can attest to this arti­cle tbh, AI or rather the automa­tion of a lot of process­es were very help­ful for me to cre­ate videos for my work.
    I use most­ly Adobe soft­ware to cre­ate mar­ket­ing videos and tuto­ri­als and I have to say, I real­ly love the direc­tion video tech­nol­o­gy is head­ing.

Join the discussion