Meer as die helfte van die Brasiliërs het al produkte gekoop wat deur KI beïnvloed is

Alredy no longer a novidade that artificial intelligences are increasingly present in the routine of Brazilians, whether to study, work or seek daily information. Now, the tools are also beginning to gain space at an important moment in the consumer's life: the purchase decision.

A new survey by Branddi, a company specializing in brand protection in the digital environment, reveals that 5 out of 10 Brazilians (54%) have already purchased some product or service directly influenced by recommendations made by artificial intelligences. Desses, 34% afirmam já ter realizado compras mais de uma vez com influência das ferramentas, enquanto 20% fizeram isso uma vez. Além disso, uma parcela significativa (27%) indicou que ainda não utilizou a ferramenta para esse fim, mas faria.

The habit also reveals a new behavior related to a process that precedes the purchase. Although search engines, like Google, are still the main source of information about products and services (72%), the use of artificial intelligences is already significant and has been gaining more and more space: 66% of respondents state they use AI to research products and services, with 38% doing so quite frequently and 28% stating they use it every time.

The movement indicates a gradual change in how consumers research products online, with artificial intelligences taking on a role that was previously concentrated in traditional search engines, sites of resensies and digital influencers. 

The purposes of use also vary: among the main objectives are to clarify doubts about product features (62%), seek personalized recommendations (54%), and compare prices and functionalities (48%).

According to Diego Daminelli, CEO of Branddi, the advancement of artificial intelligences in the purchase journey shows an important change in the digital behavior of consumers: “Today, many consumers already see artificial intelligences as a kind of shopping assistant. They help save time, compare options and gather information more quickly, which naturally makes them influence consumption decisions”.

“The point of attention is that this relationship is still under construction. Many people use the tools believing that all recommendations are completely neutral or accurate, when, in practice, there are limitations, biases and even risks involving incorrect information,” he explains.

Concerns and risks of scams

Despite the growing presence of artificial intelligences in the purchase journey, consumers still demonstrate some insecurities regarding the use of the tools in this context. Among the main concerns pointed out by respondents are the risk of receiving incorrect information (45%), misuse of personal data (36%), biased indications or disguised advertising (30%), and the dissemination of scams or frauds (28%)..

Die moeilikeheid om te identifiseer of ’n aanbeveling deur KI regtig neutraal is, verskyn ook as ’n relevante uitdaging. Volgens data van die navorsing, 45% van die respondente sê hulle het al probleme gehad om te verstaan of ’n voorstel deur die intelligensieartificiaal onpartydig of gesponsord was. Daarteenoor sê 29% dat hulle nog nooit hierdie probleem ervaar het nie.

“Die intelligensieartificiale is nog in ”n versnellende ontwikkelingsproses, en verbruikers moet ’n kritiese oog hou oor die aanbevelings wat hulle ontvang. Nie elke aanbeveling verteenwoordig noodwendig die beste beskikbare opsie of ’n betroubare bron nie. Die sorg moet nog groter wees in situasies wat baie voordelige promosies, eksterne skakels of versoeke vir persoonlike data behels. Bedrieërs volg digitale tendense en pas hul strategieë vinnig aan by die nuwe gewoontes van verbruikers,” waarsku Daminelli.

“In die komende jare is die neiging dat die intelligensieartificiale nog meer invloed op koopbesluite sal hê. Daarom sal deursigtigheid oor advertensies, verantwoordelikheid in die hantering van data en verificasiemekanismes van kardinale belang wees om die vertroue van verbruikers in hierdie nuwe digitale scenario te versterk,” sluit hy af.

Metodologie

– Publiek: 500 Brasiliane van alle state van die land, insluitend vroue en mans, van 18 jaar en alle sosiale klasse is ondervra.

''' Versameling: die data van die studie is via ’n aanlyn navorsingsplatform ingesamel. Die betroubaarheidsindeks was 95%, en die foutmarge was 3,3 persentasiepunte.

''' Versamelingsdatum: uitgevoer op 12 Januarie 2026.