#AIChat is a Twitter chat on artificial intelligence.
My guest(s) give you actionable ideas you can immediately use at home or work.
On April 17th, 2019, my guests for #AIChat were @adlibsoftware and @Adlib_ScottM.
The topic? The Achilles Heel of AI (credit goes to team at @adlibsoftware for the topic title and their help with the questions).
Mistakes are mine.
———————
Here’s Q1 for @adlibsoftware and @Adlib_ScottM:
AI is the buzzword of the day – where are we on the hype cycle? Innovation trigger, inflated expectations, disillusionment, enlightenment or real productivity? #AIChat
— Nick Tang (@nickhtang) April 17, 2019
https://platform.twitter.com/widgets.js
A1a: All of the above or true, depending on each business’ unique situation, their industry and where they’re at in terms of adopting #AI (or considering it). #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A1b: Here’s a resource from @McKinsey regarding the rate of #AI adoption. They’re reporting that half of the organizations they surveyed have embedded #AI, with another 30% in the pilot stage: https://t.co/uLmi1vq6Fh #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A1c: Also, according to @McKinsey — by sector – firms in telecom, high-tech and FinServ are leading the way in overall #AI adoption: https://t.co/uLmi1vq6Fh #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
From my perspective, you can’t leverage what you don’t know. Critical to find out what’s in those untouched repositories of documents to learn from the past and inform the future leveraging #AI. #AIChat
— Nancy MacMillan (@NancyMAC29) April 17, 2019
https://platform.twitter.com/widgets.js
I think we’re coming out of disillusionment and into enlightenment. There are some real world examples of AI working. Retail, Insurance… #AIChat https://t.co/Pgruc1v8lQ
— Scott Mackey (@Adlib_ScottM) April 17, 2019
https://platform.twitter.com/widgets.js
Scott is correct, AI is a broad term that’s widely applied. There are already real world use cases with notable benefits, but there’s a sensationalized viewpoint in public perception. Educating the public is as much of a challenge as implementation #AIChat
— Colin McGuire (@realColinMac) April 17, 2019
https://platform.twitter.com/widgets.js
As you’re talking Q1, here’s the next question for @adlibsoftware and @Adlib_ScottM
Q2) Can AI work without structured data to drive it? #AIChat pic.twitter.com/yP3AG4bQUj
— Nick Tang (@nickhtang) April 17, 2019
https://platform.twitter.com/widgets.js
A2a: #AI requires massive amounts of data to train its algorithms. The more data it has, the more accurate it becomes. Structured data is essential for this! #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A2b: The challenge with data for #AI engines stems from unstructured content, which isn’t machine-readable such as emails, scanned documents, images etc. and it typically makes up about 80% of enterprise content. #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A2c: Another challenge with #UnstructuredData (beyond its existing volume) is that it’s growing at a rate of 26.8% annually, according to @IDC, making it imperative for #DataGovernance now versus later. #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
Sure it can work, but with 80% of data locked up in unstructured content, value will be limited. #AIChat
— Rob Schaafsma (@SchaafsmaRob) April 17, 2019
https://platform.twitter.com/widgets.js
Nope…it’s a blind spot. So much of it’s power comes from having the data to perform it’s magic. Data locked in #unstructured documents is a challenge! #AIChat https://t.co/ii6c7rmezZ
— Scott Mackey (@Adlib_ScottM) April 17, 2019
https://platform.twitter.com/widgets.js
Onto Q3 with @adlibsoftware and @Adlib_ScottM
Q3) To those thinking about AI – how would they tackle the unstructured data challenge? #AIChat
— Nick Tang (@nickhtang) April 17, 2019
https://platform.twitter.com/widgets.js
Start with the outcome in mind – what are you trying to achieve, then dig into the underlying elements that will feed that outcome. People, process, technology…and the data sources required. https://t.co/mobfnEgxHB
— Scott Mackey (@Adlib_ScottM) April 17, 2019
https://platform.twitter.com/widgets.js
A3a: To get started with #AI, businesses need to address their #DataStrategy, with an end goal of achieving better access to quality data. #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A3b: You can check out this blog post that speaks directly to *how* businesses can achieve #BetterData for #AI that is ✔ accessible ✔ timely and ✔ relevant: https://t.co/Hk0aOr0Gpe #AIChat pic.twitter.com/Br8VWrEdgz
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
#Content needs to be made machine readable, recognized as a particular type, then have data extracted according to that type. For example: EffectiveDate from a contract. #AIChat
— Rob Schaafsma (@SchaafsmaRob) April 17, 2019
https://platform.twitter.com/widgets.js
A3 Yes & No as #AI will simplify tasks as well as convert data into efficient streams as well as filter out the waste but there is still to what ends what is you use that’s often where the hard questions come in Utilization 🌐🤔 #AIChat
— BrainBlender🤔🌐 (@BrainBlenderTec) April 17, 2019
https://platform.twitter.com/widgets.js
As you’re learning from @Adlib_ScottM and @adlibsoftware during #AIChat here’s the last question for them:
Q4) Do businesses need to address their underlying data processes before applying AI – agree or disagree, and if so, how? #AIChat pic.twitter.com/6aqXZ74Ywk
— Nick Tang (@nickhtang) April 17, 2019
https://platform.twitter.com/widgets.js
A4a: If they want to see success with #AI, data standardization, extraction using #OCRsoftware and AI-based data classification are essential to achieve machine-consumable content. #AIChat
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
A4b: We also want to emphasize the need for an ongoing #DataGovernance strategy that automatically addresses content as it’s created, versus a one-time data-enrichment process. Here’s the 411: https://t.co/ytxXgDpjIf #AIChat pic.twitter.com/tsxS3f5vOj
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
Agree. You need to ensure the fuel you’re using for #AI is clean, broadly sourced and relevant. That said – you don’t have to wait, there may be use cases where you have good data and can get started. #AIChat https://t.co/maLRLz3GMZ
— Scott Mackey (@Adlib_ScottM) April 17, 2019
https://platform.twitter.com/widgets.js
A note before we wrap up! Prior to enhancing your #DataQuality, it’s essential to understand what content you’re working with. Here’s how you can kick-start this process: https://t.co/ZuM703Kgki #AIChat pic.twitter.com/ya2pc55fvo
— Adlib Software (@adlibsoftware) April 17, 2019
https://platform.twitter.com/widgets.js
If you take away just one thing from today’s #AIChat, let it be the importance of #DataQuality in digital business. Here’s a look at how Adlib delivers #BetterData: https://t.co/PUbw1WXFBQ pic.twitter.com/laXS2hvZHk
— Adlib Software (@adlibsoftware) April 17, 2019