Time for customer-tech’s next wave of easier, applied AI claims to be tested – starting with Salesforce’s Copilot-killing, hallucination-busting autonomous agents
If you’re in a marketing team still attempting to crack how AI can be deployed safely, effectively, and efficiently with some scale and real ROI scenarios, saddle up to an industry peer in the next six months who’s road-testing the “no code” drag-and-drop templates and process-collapsing, easy-build, AI capability claims launched last month by Salesforce across its multiple cloud offers. The customer tech giant this week is on the Australian leg of its “World Tour” to 3000 delegates and such is the pitch - that it has simplified and revolutionised “applied AI” for businesses that have struggled so far to cut through their hallucinogenic concerns and deployment complexity - that all industry eyes and ears in the next six months should be on whether early users agree with the bold claims pasted on the Salesforce can.
Salesforce has either set itself up for a fast sweep through the customer tech market share battlefield or a fast wake-up call that its well-oiled marketing engine is too far ahead of what real-world marketing, commerce, sales, and service teams on the ground can actually deliver.
Upwards of 2,000 delegates will descend on Sydney’s ICC this week to hear first-hand what Salesforce says the future of AI should look like in business applications – no code, limitless iterations of templates, and custom-built autonomous agents that, in the case of marketing, set about building campaigns from scratch - segmentation and audience targets, strategic messaging and creative execution drafts included. In the Salesforce playbook, marketing teams can now connect up their agents to the commerce, customer service, and sales teams to create company-wide customer lifecycle journeys. These same agents rip out traditional, mundane process tasks like building multiple customer journey maps, which have been the specialist capability of many in CX and marketing teams.
The global lead on Salesforce Marketing Cloud, Steve Hammond, is adamant his firm’s AI rollout of its “Agentforce” autonomous agents addresses the trials and tribulations of marketing teams, specifically, where and how to apply generative AI with less complexity and pain and more impact and ROI certainty.
That deployment complexity and fear among marketers around AI’s hallucinogenic accuracy is one of the biggest themes to emerge from the latest State of Marketing Report from Salesforce – along with campaign performance and ROI measurement. The other ongoing frustration is data stuck in divisional silos.
“The biggest challenge for marketers is that unless they have their data pulled together into a place where they can run AI on top of it, they can't really create the advantage of AI really working for them,” Hammond told Mi3 during last month’s annual Dreamforce event in San Francisco. “If a marketer has 15 different silos of data, it's a real challenge for any kind of AI system to be able to understand what all that data means in these different systems, and then a marketer isn't really in a position to be able to have that AI give them an advantage of working across them all.”
This argument is central to Salesforce’s AI engine and its autonomous agents; that a company’s own data, across all its operations, is what AI agents are best trained on to take personalised customer action with tangible business improvements versus generic LLM models and complex, costly and custom AI experimentation.
The competitive response from cloud rivals including Microsoft, Adobe, Google, and best-of-breed martech and CX vendors will be one to watch.
Hammond and the broader Salesforce machine say what Salesforce’s Agentforce and cloud stack can do “almost exclusively” is help redefine marketing beyond top-of-funnel advertising activities like getting someone to a website or driving a clickthrough or search query.
Engagement
“Marketing now has the opportunity to extend into an engagement where someone says to a call centre, ‘Hey, look, I bought this product, the call center agent can see the history of someone's marketing activity, the products they've looked at, videos they've looked at, campaigns they've responded to and then in that environment, continue to further personalise that engagement with that customer because they have all that together.
“We've built a system that lets an organisation look at the broader problem of how they create this ideal customer journey across the entire company and for marketing to be able to continually reengage that customer with the best-personalised content and actions to keep the brand, loyalty, and engagement for that customer all the way through the cycle.”
An upside for marketers in this rollout is partly less mundane, process-driven time spent on CX for teams – Marketing Cloud’s
“Agentforce Campaign” product, for instance, allows teams to build an AI-assembled brief using natural language prompts and tactical recommendations like audience segments and channels for messaging, based on a company’s own business data.
“It’s taking multiple steps out of campaign creation,” Hammond says. “More broadly, that ability for marketing to have AI customer data pull from across the organisation and be used in real-time [with autonomous agents] is almost exclusively with Salesforce.”
That assertion over the next six months should deliver some early results and industry sentiment: Has Salesforce created what it claims - a breakthrough AI moment for customer experience, revenue growth, and efficiencies versus its competitors – or another wave of selling business dreams, this time with reasoning, autonomous business agents?