Who is really adopting AI in their industry?

artem-sapegin-180146-unsplash

Bitvore looks at which industries are adopting AI, how many projects are being worked on, and how well they are performing.

During the space race of the 50s, 60s, and 70s, there was a reason so many monkeys and chimpanzees were shot up into space.  Getting into space is relatively easy.  Getting back safely, is the hard part (How the Spaceship Got Its Shape, credit starc@) 

One particular difficulty involves atmospheric re-entry.  If a spacecraft’s speed and angle is too steep, the deceleration forces will generate too much heat.  If the angle is too shallow, other unpleasant things can happen.  Contrary to popular opinion that the spaceship will “bounce off the atmosphere like a flat stone skipping off the water surface of a pond,” what typically happens is the craft doesn’t lose enough velocity in the dense atmosphere, may miss its mark, continue on its orbit, or be exposed to heat flux for much longer periods of time.  

 

Adopting Technology

Science aside, re-entry has become a commonly used metaphor for all sorts of business activities. Employees will bounce out of companies if they aren’t a good fit.  Technology adoption is the same way.  Rapid uptake of technology can result in what Gartnercalls a “hype cycle,” leading to individuals burning out on new technology due to disappointment of unfulfilled promises and over-reach.  Technology without enough momentum or deep adoption will lose support fairly quickly if it doesn’t help solve enough real problems and “bounce out.”

 

Identifying the Movers and the Shakers

At Bitvore, we wanted to determine which industries' artificial intelligence as a technology is making a dent.  By querying our production system for artificial intelligence or machine learning projects, implementations, rollouts, deployments, or integrations, we were able to identify around 1.7 million news stories.  Of those, approximately 8,500 were what Bitvore calls precision intelligence.  Translation: the record matches our “AI project” concept criteria, and a high confidence that one of the companies we track is associated with the record, with text that describes a situation where something important happened. In our system, we call this a signal.

 Try Bitvore's Interactive Data Tool

 

So, what did we find?

A lot of interesting things.  First, non-precision news has a lot of AI and machine learning projects described in the media.  Most of those either aren’t associated with a particular company, are simply research summaries, or are experimental recitations of work done with the technology.   We characterize these stories as the AI industry talking about AI.  Examples include governments, trade groups, unions, industry associations, or research groups discussing the need for better AI adoption or competitiveness. Most of the world already knows they have to do something with AI, but we are far more interested in actual adoption into an industry.   For that reason, the “otherwise unclassified” AI project bubble was excluded from the summary graph below.

 

2019-03-04_11-03-04

 

Each bubble represents an industry.  The size shows the number of AI projects we found as correlated to the industry. 

The color shows the average sentiment across all projects in that industry.  The range of sentiment is between 0.0614 (weak positive) and 0.6159 (strong positive). Orange represents above average sentiment within their industry. Blue represents below average in their industry compared to other vendors. One thing to note is that we are using NAICS Industry codes for this analysis and they are hierarchical and distinct.  Software Publishers (511210) is a specialization of Software Publishers (Generic) (5112). You can see the full list of specializations for Software Publishers here

 

Top 15 Industries doing the most AI projects going well:

  1.    Software Publishers (511210)

2.     Hardware Manufacturing (332510)

3.     Computer Systems Design and Related Services (54151)

4.     Software Publishers (5112)

5.     Finance and Insurance (52)

6.     Automotive Parts and Accessories Stores  (441310)

7.     Retail Trade (44-45)

8.     Aerospace Product and Parts Manufacturing (3364)

9.     Electrical Equipment Appliance and Component Manufacturing (335)

10.   Depository Credit Intermediation  (5221)

11.   Pharmaceutical Preparation Manufacturing  (325412)

12.   Electronic Shopping and Mail-Order Houses  (45411)

13.   Wireless Telecommunications Carriers (except Satellite) (517210)

14.   Professional Scientific and Technical Services (54)

15.   Investment Advice  (523930)

 

Top 15 Industries with the most AI projects struggling below average:

1.     Aerospace Product and Parts Manufacturing (3364)

2.     Depository Credit Intermediation  (5221)

3.     Internet Publishing and Broadcasting and Web Search Portals (519130)

4.     Medical Equipment and Supplies Manufacturing (3391)

5.     Other Computer Related Services (541519)

6.     Automobile Manufacturing  (336111)

7.     Pharmaceutical and Medicine Manufacturing (3254)

8.     Personal and Household Goods Repair and Maintenance (8114)

9.     Other Investment Pools and Funds (5259)

10.   Specialized Freight Trucking (4842)

11.   Discount Department Stores  (452112)

12.   Marketing Research and Public Opinion Polling (54191)

13.   Water Transportation (483)

14.   Lessors of Other Real Estate Property  (53119)

15.   Scientific Research and Development Services (5417)

 

Top 15 Industries hottest on experimental adoption of AI:

1.     Other Nondepository Credit Intermediation  (52229)

2.     Telephone Call Centers (56142)

3.     Scenic and Sightseeing Transportation (487)

4.     Real Estate (531)

5.     Portfolio Management  (523920)

6.     Support Activities for Animal Production (11521)

7.     Publishing Industries (except Internet) (511)

8.     Jewelry Luggage and Leather Goods Stores  (4483)

9.     Bed-and-Breakfast Inns  (721191)

10.   Motion Picture and Sound Recording Industries (512)

11.   Collection Agencies (56144)

12.   Newspaper Publishers (51111)

13.   Couriers and Express Delivery Services (4921)

14.   Radio Broadcasting (51511)

15.   Directory and Mailing List Publishers  (511140)

 

Top 15 Industries dreading the day their AI projects get finished:

1.     Waste Management and Remediation Services (562)

2.     Oil and Gas Field Machinery and Equipment Manufacturing  (333132)

3.     Cafeterias Grill Buffets and Buffets  (722514)

4.     Tobacco Manufacturing  (312230)

5.     Other Financial Vehicles  (525990)

6.     Specialized Freight Trucking (4842)

7.     Biological Product (except Diagnostic) Manufacturing  (325414)

8.     Other Insurance Related Activities  (52429)

9.     Motor Vehicle and Motor Vehicle Parts and Supplies Merchant Wholesalers  (4231)

10.   Public Relations Agencies (54182)

11.   Grocery and Related Product Merchant Wholesalers  (4244)

12.   Apparel Accessories and Other Apparel Manufacturing (31599)

13.   Lessors of Other Real Estate Property  (53119)

14.   Semiconductor and Related Device Manufacturing  (334413)

15.   Aircraft Manufacturing  (336411)

 

A Deeper, Quick Look

Obviously there are too many categories to do a full analysis in one blog post, but we wanted to figure out why Waste Management and Remediation Services, and Oil and Gas Field Machinery and Equipment Manufacturing were having such a difficult time with their AI adoption. 

Below is a string of stories on how algorithms and artificial intelligence are taking over the oil fields, including Weatherford International.  One company is perfecting the use of the Internet-of-Things (IoT) and AI in the oil patch.  It appears that it’s not quite perfected yet.

In addition, below is a series about how AI can revolutionize Waste Management, including Covanta Environmental Solutions, WasteExpo’s “Rise of the Robots” and cleaning up nuclear waste, Republic Services Inc., Casella Waste Systems, and companies wanting to be the “Uber of Trash."

 

Final Thoughts

At Bitvore, we look at a lot of different topics.  Building an off the cuff concept of an “AI project,” when combined with our years of precision intelligence news data and our reference metadata, allows us to see industry trends, do deep dives on specific companies, and put our fingers on the pulse of an industry.  Artificial intelligence and machine learning are clearly here to stay.  The technology is resulting in a major impact across all the industries in the world, however it’s not being adopted uniformly.  Where it is being adopted, some are doing so with below average alacrity when compared to adoption in other industries.  As the movers and shakers either adopt the technologies more deeply or bounce out, we're going to keep our eye on adoption.

Try Bitvore's Interactive Data Tool

Greg Bolcer, CDO Bitvore

Greg Bolcer, CDO Bitvore

Greg is a serial entrepreneur who has founded three angel and VC-funded companies. He's been involved at an early stage or as an advisor to at least half a dozen more. Greg has a PhD and BS in Information and Computer Sciences from UC Irvine and a MS from USC.