When it comes to embedding AI and machine learning into an organization, the advice of ‘don’t run before you can walk’ is worth heeding.
That’s the view of Jay Kulkarni, founder of Theorem, who believes retail and ecommerce brands – and their marketing agencies – must stop and think before rushing to imple-ment too many new technologies at once. Organisations will drown in data and fragmentation if they dive in too fast, he warns.
He acknowledges that automat-ing mundane and tactical tasks can bring efficiencies, freeing up time for more human thinking and crea-tivity. But to succeed, leaders must define their business problems and goals first and then apply technology solutions to them.
“When organisations lead with the tech, it fails because you’re not fixing a problem,” he explains.
According to Kulkarni, harnessing robotic processing and automation can achieve “more from less” at a time when budgets are tight, talent is scarce, inflation is sticky and costs and prices are rising. It can also demon-strate a measurable ROI to marketers looking to streamline workflows.
But there must be a coherent approach to infrastructure, without silos, he says, describing technology as “a commodity”.
“Based on your business objectives, you find the platform and the tools to help you achieve them. Work back-wards, what will your marketing activ-ity look like two to three years out? This is a multi-year project, continu-ally optimising as you go,” he adds.
Accept it won’t all work the first time
Kulkarni believes brands and agencies are now more comfortable experi- menting, especially with automation accessible directly from the cloud through AWS or Microsoft Azure; this makes it cost-effective.
Theorem works with global publish- ers and B2C brands to create flexible marketing solutions and its efforts to experiment internally first has helped win client confidence by demonstrat- ing tangible successes.
“We automated reporting with a 50% reduction in human labour,” Kulkarni explains. “If you are a media planner or campaign manager, you want a holistic view of data and in the past that was a very manual process.
“Now we have a set of bots download- ing campaign or client data, cleansing it and creating reports for us.”
Admitting robotic process auto- mation is not yet perfect, he cites Theorem’s approach as ‘humanising bots’, reducing 100% human involve- ment with people still on hand should a bot fail or there is data ambiguity.
“Marketing is half art and half sci- ence,” adds Kulkarni, “and automation helps with the science. The art is where we still have a pair of human eyes.”
Automating the data reporting that underpins any campaign is a low-friction way to begin this jour- ney, he advises. End-to-end campaign deployment could then come next, using machine learning alongside Salesforce integration to help plan and action insertions.
“The human element will always stay as the art of marketing, it will never go away because of the machine,” believes Kulkarni. “The trick is that balance.”
Full transparency is critical
Theorem’s own internal programme has ensured career paths for those employees displaced by automation. Kulkarni suggests clients do the same while ensuring full transparency of their actions. “If people feel insecure about automation, it will not succeed. This is an organisational challenge, not a technological one.”
There are then demonstrable upsides; saving human resources on reporting can have a huge knock-on effect on the business. For example, Kulkarni suggests time is better spent scaling brand loyalty programmes.
With the cost of acquisition for new audiences now going up significantly, the ability for a retail or ecommerce brand to cross-sell and upsell to exist- ing audiences is, according to him, “good business”.
Giving employees more time to develop and deploy ideas that build tighter rela- tionships with customers is critical at a moment when consumer spending is falling, he says. “The ability for brands to maximise loyalty programmes digitally is a lot more cost-effective. We have seen this get the highest share of volume rel- ative to acquisition of new audiences on social media.”
He adds: “Freeing up teams to think of more innovative ways to create market- ing campaigns gives them time for the art rather than the science.”