The year ahead: Driving the adoption of AI applications

road_cars_electric2The concept of modern-day artificial intelligence might have been around since the 1950s – even earlier, according to some – but 2023 is likely to see AI finally hit the mainstream with increasingly innovative applications, although the appetite for change is still open to debate.

Earlier this year, an engineer for Google’s Responsible AI Organisation was sacked for saying the AI chatbot he was working on was sentient – describing it as having the equivalent perception and abilities to express thoughts and feelings as that of a human child.

Realistically, seeing sentient chatbots in our Google Nest, Alexa or Siri is unlikely in the next 12 months. However, customer service chatbots and other narrow applications of human-computer interactions will claim to be using AI that is no different to human interaction. So, no more “computer says no”, it might now say “maybe”.

The automotive industry is not known for advanced data analytics historically, yet it has made leaps in utilising AI, especially by trying to move to autonomous vehicles. The integration of AI into EVs will highlight automotive manufacturers’ AI capabilities, with intelligent cameras which can detect the drowsiness of drivers, and safety assist features which reduce the chances of a collision.

Next year we can expect to see the introduction of some form of personalised pricing based on the data that EV vehicles are capable of collecting, whether that be for insurance (partnerships between automotive manufacturers and insurance companies) or in leasing/finance costs.

shorful islamAnother area which is likely to dominate in the year ahead, due to Google’s latest extension to phasing out third-party cookies in Chrome to 2024, is the discussion around zero- and first-party data.

We shouldn’t expect a further delay as global privacy laws are making increasingly difficult to collect data without consent. Brands and advertisers will have to find new ways to both measure the effectiveness of advertising and to collect zero- and first-party data. Data management platforms will mature to incorporate the data traditionally collected by customer data platforms and offer brands an effortless way to collect data.

However, whether or not customers want to part with this data will be a significant test and we can expect to see many brands innovate how their digital properties are utilised. Some from pure brochureware to customer lead generation, and others will move to being more functional, offering a form of a transactional element, such as the move by automotive manufacturers to introduce ecommerce.

Publishers will have a more challenging time delivering a full subscription offering to all of their customers. Those that rely heavily on an ad-funded model will work towards a combination of collecting zero- and first-party data, as well as using more traditional approaches such as content-based targeting and even the use of panel-based measurement systems. Google is offering Topics to segment users, but how that will be delivered is unknown and publishers need to have a back-up plan.

If one thing is for sure, there will be greater transparency next year in how brands collect data, as the increasing need to both have zero- and first-party data, and to share with selected partners will force brands to be more open about what they are doing with the data due to the fears of falling foul of one or more regional privacy laws.

With third-party cookies being blocked, evidence from media agencies and data recruitment agencies shows an uptick in the demand for analysts with marketing mix modelling skills.

Brands may just decide that this worked well pre-digital marketing measurement, and with the challenges of third-party cookie blocking and complex privacy laws, perhaps MMM is something safe and trusted to return to.

Shorful Islam is chief data scientist at Tribal Worldwide

Print Friendly