In January 2025, Cisco held its inaugural AI Summit in Silicon Valley, the place tech leaders mentioned the alternatives and ethics round AI.
The occasion was purposely small to supply extra worth to attendees, primarily expertise executives from main manufacturers. Throughout his opening keynotes, Cisco CEO Chuck Robbins famous that 40 of the Fortune 100 had been represented on the occasion. The lineup was an interesting mixture of audio system from Scale AI, Glean, Credo AI, Cohere, OpenAI, Goldman Sachs and extra, in addition to many Cisco executives.
Attending the occasion modified my pondering in some areas of AI and confirmed different ideas. Under are my prime 5 takeaways from the Cisco AI Summit.
1. AI goes to displace jobs however create new ones
The most typical query individuals ask me about AI is the way it will have an effect on jobs. Many individuals consider it would change jobs, whereas others consider it would increase roles. Like all tech transitions, the fact is it would displace many roles. VoIP did away with voice engineers; PCs successfully eradicated mainframe directors; the cloud lessened the reliance on server directors; and AI could have an identical impact. Nevertheless, every expertise shift created extra jobs on the again finish of the transition.
In 2023, the World Financial Discussion board predicted that AI would get rid of 85 million jobs however create 97 million. Proper now, we aren’t certain what these jobs will likely be, which leads to the nervousness round AI we see right this moment. Throughout the summit, David Solomon, CEO of Goldman Sachs, acknowledged this transition.
“Huge expertise waves have introduced massive modifications on this planet for a protracted time period,” Solomon stated. “They destroy jobs, but we managed to create new jobs, and we drive development within the financial system.”
The important thing to transitioning from the pre-AI world to post-AI is embracing the expertise and adapting to vary.
2. We’re operating out of actual information to coach AI fashions, however artificial information can present aid
I first heard this assertion a couple of yr in the past, and I wasn’t certain the way it was doable, given how a lot information is created with AI. Throughout the summit, Scale AI CEO Alexandr Wang reiterated this sentiment and voiced his concern about AI with the ability to evolve on the pace we have seen.
It is important to comprehend that AI fashions are solely educated with publicly obtainable information, and far of the info being created is proprietary. As these programs develop in complexity, the quantity of high-quality, labeled information wanted to coach them successfully grows exponentially.
Artificial information provides an answer through the use of algorithms to generate synthetic information that mimic real-world eventualities, offering a nearly limitless useful resource for coaching AI fashions. This method can fill gaps the place actual information is sparse or laborious to gather, permitting for extra numerous coaching information with out compromising privateness or moral issues.
3. AI pricing fashions are nonetheless up within the air
One other widespread query about AI is how the pricing mannequin works. Most software program is presently offered with a subscription mannequin the place the enterprise pays per person, per 30 days, with cloud companies sometimes based mostly on utilization. How AI needs to be priced is a bit up within the air. CEOs and CFOs ceaselessly ask this query on earnings calls as a result of how an organization monetizes AI can have an effect on its valuation.
An organization with a subscription mannequin has a predictable income stream, whereas utilization-based pricing is extra unpredictable. Throughout a hearth chat on the Cisco AI Summit, Field CEO Aaron Levie supplied an alternative choice: outcome-based pricing. He gave an instance of an AI device designed to generate leads, and a vendor might cost based mostly on the standard of the leads generated.
There seems to be no consensus on how AI-based companies will likely be priced, besides that change is coming. IT leaders should press their distributors to make sure they totally perceive how and when change is coming.
4. AI will create an incremental step perform in productiveness and alter employee expectations
Most of the classes mentioned how individuals use AI. Field’s Levie mentioned how AI is shifting from aiding customers in doing their jobs to with the ability to do the work on behalf of somebody. In a separate dialogue, Arvind Jain, CEO of Glean, mentioned how the success of AI resides in its capacity to finish complicated end-to-end duties.
This implies it will not all the time change individuals however permit them to do extra. In a contact middle, an AI agent can summarize and log name summaries sooner and extra precisely than individuals. In a gross sales division, an AI device can analyze calls and emails and assist qualify prospects sooner. In healthcare, AI brokers can diagnose MRIs sooner, enabling medical doctors to spend extra time treating sufferers.
What do individuals do with the additional time? Do they take Fridays off? Do they get pleasure from longer lunches? Whereas this may sound interesting, we will doubtless anticipate greater productiveness ranges, and companies ought to regulate their targets accordingly. Within the examples above, a contact middle agent needs to be anticipated to deal with extra calls, whereas the salesperson can be anticipated to shut extra offers. As Solomon acknowledged, “We drive development within the financial system.”
We have seen this sample earlier than. Automation instruments raised gross sales quotas, and analytics made banking companies, akin to mortgage approvals, sooner. Using AI will allow staff to work sooner and smarter, and productiveness expectations will develop in consequence.
5. AI requires a safety rethink
The first goal of the Cisco AI Summit was to drive thought management in AI. Its secondary goal was to launch AI Protection, a safety device designed to guard corporations from the distinctive dangers related to AI.
Safety has all the time been a problem for corporations, because the tech atmosphere continues to evolve in ways in which broaden the assault floor. Cloud, cell, do business from home, SaaS and different traits have made the expertise atmosphere extra dynamic and distributed, which performs havoc with cyber groups.
Cisco’s AI Protection secures AI by offering visibility into AI purposes used inside an enterprise. It validates the assorted AI fashions for potential vulnerabilities. Organizations can use AI Protection to implement safety controls and safeguard the corporate from AI misuse and information leakage.
In a current dialog I had with fellow analyst Bob O’Donnell from TECHnalysis Analysis, he defined, “Cisco is utilizing observability instruments, akin to Splunk and ThousandEyes, to supply insights into what fashions are operating, the sorts of queries being run, the prompts getting used and lengthen the notion of observability into the AI period.”
He stated this transfer “helps place Cisco as an AI thought chief.” Extra thought management from the mainstream IT distributors is sweet for the business. Up to now, most AI evangelism has come from Nvidia and startups. Main IT distributors akin to Cisco will play a key position as AI scales.
Abstract
The Cisco AI Summit completed its main aim of offering a thought management platform for AI. The corporate acknowledged it might make the occasion an annual one, which I used to be glad to listen to, as tech wants extra business occasions slightly than ones overloaded with product data.