NetAppAI

Seed corn - use of DL to identify kernels of corn most likely to germinate increasing yields while driving down costs.

I love the inhaler example that Monty has talked about. Both my kids suffered for Asthma. I wish they had smart inhalers 10+ years ago.

I gotta go chatbots.

1. self driving cars; 2. diagnosing disease probabilities; 3. the future of warfare; 4. prioritizing responses to cyber attacks that have the highest impact; 5. predicting buyer patterns in retail to optimize sales

Social Chatbots; specific example I saw was for those going through severe depression and are seeking assistance.

automating customer care calls in call centers, drive-thrus ...

AI --> lowest energy manufacturing. Good for profit; good for the planet.

Autonomous transportation platforms. Intelligent robotics to power autonomous mobility for the disabled, elderly, etc.

Self Driving cars- just take me to work and home

virtual assistants

Swarming AI-powered intelligent coordinating devices, not just for warfare but for precision agriculture, architectural surveys, law enforcement, etc.

Drought planning

The most impactful in the next five years will be Autonomous Cars. The scariest is AI in warfare

I'd break out "seamless" in this context as automated and accelerated data analytics DevOps pipelines, in which AI workflows drive continuous ingest, modeling, training, and serving of AI models into business apps.

agree, managing data from edge to core to cloud is key

Agree, the AI work flow is complex and this is just the begining

@jameskobielus You are spot on. I would add a key element which is the need for seamless edge integration whether data ingestion from the edge to driving production models to the edge plus "rinsing and repeating."

Important point this. Organizations' data generally spans multiple environments/data centers. on-premises, cloud, and even edge devices in IoT apps. You want to easy data management.

comprehensive data management for AI is a must

For details on edge to core to cloud, read this white paper https://nt-ap.com/2v...
@SundarRanganat4

AI for social is a great NLP application using entity extraction and community detection with LM can enable AI to be more valuable for all systems and people

absolutely - we also see more enterprises using AI to mine insights from their call center recordings - can help with sentiment analysis, improving customer experience, reduce churn, etc.

we use ML for extracting insights from these chats and all Cube videos as we have full transcripts on all of our content. Soon we will make the leaderboard more about reputation and real influence not how many followers someone has

community and impact of content on social is our key focus on influence and reputation in context

@TonyPaikeday @MatthewButter lots of automation is coming that users actually want for support and soon other value activities. I'm with you guys

Well I know Southwest has 40+ people whose sole job is to monitor social media and monitor for customer support opportunities. Chatbots will eventually reduce those numbers drastically.

Southwest prides itself with it's customer service. I don't necessarily think that they will decrease the numbers of people working, but chatbots will allow them to scale their efforts to offer better and more personalized experiences for their customers.

Right. AI-driven multichannel customer contact infastructures are automating a lot of first- and second-level support, thereby reducing costs, speeding issue resolution, and capturing key customer/sentiment insights in real time.

Right. AI is driving both outbound customer service scenarios such as that, plus lots of inbound applications in digital channels to improve responsiveness, routing, resolution, experience, etc.

@furrier @TonyPaikeday @MatthewButter ML/AI crucial for social analytics moving forward to extract/overlay context and intent rather than just influence and reach

@dunntony I would love to get that kind of insight from my own team's efforts :)

.@plburris yes

Short answer is "Yes". It will help both. NVIDIA's Inception program which includes more than 3000 startups is a testament to both scenarios.

AI is being used to extend existing apps--most notably in data integration/curation, metadata-driven search engines, IT operations management, and security incident and event management

The tooling, method, metrics implications of legacy+AI and greenfield+AI are very, very important.

The key for AI business value is automation of business processes. Most business processes impact systems of record. New AI/Advanced applications will need to feed existing systems of record to achieve levels of impact.

many examples where machine intelligence can improve existing apps - take email for example, automated voice response systems, other customer service apps, service management apps, CRM, HCM apps...logistics and supply chain...

(part 1) that's one of the difficult questions a customer needs to ask up front. "what problem am I trying to solve". You can absolutely try and disrupt a new market while also remaking existing business processes.

AI within the datacenter will again be systems that automate decisions (e.g., take multiple alerts and take action as a result)

(part 2) If you can't answer the question of "what should I do first?" then that means you need some outside help.

Tooling is improving rapidly. We hear talent more frequently. Comments?

Do-it-yourself integrations are complex.

Two key challenges - data science talent and infrastructure. Fortunately having great infrastructure like ONTAP AI enables data scientists to be more efficient/effective.

Companies face the challenge of how you build out an optimized hardware/software stack for a wide range of AI workloads. Should there be different workload-optimized stacks? Which of these stacks are best deployed in managed public clouds vs. on-premises?

Enormous Compute and storage needs

Acceptance that moving the compute to the data is a long-term strategy.

Companies also face the challenge of how you benchmark the dizzying range of AI hardware/software stack permutations against each other, given the wide range of workloads, deployment scenarios, etc now and potential.

structured visibility across massive data sets, timely access to predictive analytics

data prep. The accuracy of your models is only as good as the data you feed it. I might have 10PB of data but if it hasn't been cleaned yet then its useless

AI for good is huge and it's not just a throwaway philanthropy initiative. There is real projects having impact here

The Opioid epidemic is a huge problem, and AI/ML can help solve that costly (lives, money) problem.

AI's social-transformative power is both in its ability to drive these insights, and its reliance on a deep open-source stack. AI for "social good" efforts thrive on open-source tools, platforms, data, etc. And voluntarism by data science subject matter experts.