Uniphore, Sprinklr and Drift execs discuss enterprise AI opportunities in CX
One of Annie Weckesser’s customers is on a mission. It’s a firm whose CEO has called in the customer success manager and told them to reduce the headcount of their department from approximately 2,000 people down to 20. The CEO believes that technologies like generative artificial intelligence should make that level of downsizing both feasible and quick.
For Weckesser, CMO of Palo Alto, Calif.-based conversational AI and automation provider Uniphore, the story shows just how quickly some large businesses are starting to double down on the technology, and the urgency at which other should follow suit.
“What will happen if Wall Street or the markets gets wind of that – that customer success, customer experience or customer service can be run with one per cent of the headcount?” she asked during a panel discussion at the Collision technology conference in Toronto recently. “It’s going to change how we do modelling. It’s going to change fundamentally how we think. So run back to your organization and make sure that you have someone on your executive team who is in charge of your generative AI strategy and move with it.”
That said, Weckesser and other vendor executives are cognizant that there will likely be big differences in the adoption of enterprise AI versus the way tools like ChatGPT are being explored among consumers.
Rady Thomas, founder and CEO at New York-based experience management platform provider Sprinklr, noted that it can be amusing when a consumer asks ChatGPT a question and gets back an incorrect or strange answer. For those working in a large corporation, not so much.
“In the enterprise, you cannot afford to have unpredictable outcomes,” he said. “Anyone can have a developer make an API call and put something out that looks impressive, but you have to understand the models. You have to have algorithms, and you have to have large quantities of data that’s bound by data you can verify.”
This explains why Sprinklr had around seven pages of details about its use of AI when it filed an initial public offering, Thomas added. “It’s fascinating to see everyone wake up to this five years later.”
Drift, a Boston, Mass.-based vendor best known for providing chatbots on company web sites, has been building up AI capabilities to serve enterprise clients since 2018 and began rolling out integrations with ChatGPT this past March, its CEO Scott Ernst said. He recommended large organizations look beyond the traditional use cases for gen AI like customer service or sales, and consider the full scope of opportunities the technology offers them.
“No longer do we have to create playbooks that predefine what a customer will do when they interact with you. We can use our AI and use the guardrails that we create with our platform to make sure that we’re delivering that unique experience no matter where you are in that lifecycle.”
Weckesser said Uniphore is seeing enterprises contend with a host of challenges in getting ready to fully take advantage of AI. First is establishing an accurate total cost of ownership, given the compute power that may be required. Second is the likelihood of increased regulation of AI, and the risk of finding your organization in legal crosshairs as a result of how it has been deployed. There are also unanswered questions around whether enterprises can train their models on publicly available data or use industry-specific data.
“Even when you’ve done all that, how do you make sure that you can find a provider that has the AI ready for scale?” she wondered. “(If I’m an enterprise), how can I build my legacy systems onto that AI?”
Ernst suggested companies will have to “self-regulate” until legal authorities catch up to AI, possibly by creating their own AI codes of conduct.
“We’re going to need a new definition of corporate compliance in a world where we don’t have this control,” he said.
O f course, any form of self-regulation should be part of a broader internal enterprise AI strategy, Weckesser added, and there are still significant numbers of large organizations that haven’t gotten started yet.
“I would urge the CX leaders in the room and the technology leaders in the room to go back to your organizations and if you don’t have a generative AI strategy, you cannot wait a month,” she said. “You cannot wait a quarter. You need to move now.”
Thomas suggested the best way to kick-start that conversation is by looking deep within the enterprise before choosing any AI tools.
“I am not a fan of starting with technology first,” he said. “I think you have to start with where are you spending a lot of human labor? Where can creativity or more productivity help your company the most? So start with the business problem, and then figure out how AI can help you solve it.”
