The State of the A

1
The State of the AI Market The publishers
perspective?

• Mark FurneauxBusiness Development Director,
  Wize Nordic AB
• mark.furneaux_at_wizenordic.com
• 13 November 2009

2
Outline

• Types of AI
• Value of value added
• Is anything dying?
• What does the future hold
• What else is changing?
• Conclusion

3
Types of AI

• Scenarios
• Index header records from e-journals
• As a) but include non-electronic and/or
  non-serial content
• Index electronic and hardcopy content using
  editorial skills to include hard to process
  content
• As c) but include foreign language content
  without English summaries
• As c) but include hard to find material, local
  grey literature, with local input centres
• Index at a deeper level within a paper (e.g
  tables and figures), make objects interactive
• Battle lines are drawn between some AI producers

4
Value of value added

• Print and CDROM AI products are not dead
• Heavy investment in value added by AI publishers
  
• Value added for librarian Scopus analytical
  tools, helps those who buy the products
• Value added for end-user author profiles,
  citation searching, Derwent patent families,
  advanced search options, social tagging to allow
  comments, ratings etc
• New products as alternate revenue sources
  RefWorks, CSA Illustrata, Summon

5
Is anything dying?

• High price, easy to process, low added value
  products should die Google Scholar will do.
• Low price, easy to process, low added value may
  have a role (eg BL Inside, J-Gate)
• High price, high value added comprehensive
  databases- Scopus and Web of Knowledge big
  brands, end-user penetration, high value added,
  high price tagIs there room for Scopus and WoK?
• Specialised sources niche AI publishers, often
  non-commercial or side-line, close to their
  market, high value, high AI content, no
  economies of scale bought up or consumed
• Content changing AI producers change what they
  include because of free web (eg TWI patents)
• Key measure of AI success is usage, ProQuest
  reports exponential
• increase in usage driven by value added extras
  and alerts

6
What does the future hold?

• AI innovation has been disappointing most
  databases like hardcopy. Innovation in delivery,
  value added services, AI record mainly cost
  reduction
• Biggest weakness of AI and full-text more and
  more content retrieved, less and less time to
  read
• On the fly review creation from AI search
  results and full-text articles
• On the fly translation of non-English content
  into English and of English search results and
  full-text into non-English. Searching in
  non-English
• Searching often pretty simple still
• Summon will indexed combination of library
  catalogue, institutional repositories, free web,
  plus a few paid for sources be good enough (95?)
  to replace need for Scopus, WoK,
• METADEX etc?
• Is a single interface needed?

7
What else is changing

• Publishing specialisations are becoming
  blurredPublishers have own platforms,
  aggregators publish, agents aggregate, primary
  publishers produce AI, AI publishers aggregates
  - disintermediation
• Blurring of division between AI, full-text and
  e-books indexed summary still one of most
  effective ways to retrieve full-text accurately
• Some products combine AI and full-text CABI
  has abstracts database with hard to find
  full-text
• ProQuest acquisition of Dialog impact of a
  common platform
• AI services trying to become one stop shop,
  almost a federated search (Thomson adds CABI,
  PsycINFO to WoS), also Elsevier, EBSCO, ProQuest
  etc
• Horses for courses - complex searches (eg
  patents) need specialist sources, native
  interfaces with advanced search features, depth
  of coverage, simple searches (often by
  undergraduates) need easy to use tools, common
  interface, less than comprehensive results.
  Research based universities have greater need of
  AI.

8
What else is changing (cont)

• Corporate need for AI may differ from academic
  ROI, time, comprehensiveness, specialist use
  specialist tools
• Searchers want answers, dont care where they
  came from or who paid for them
• Many AI databases now indexed in Google Scholar
  and can be accessed from subscribing institutions
  via that route drives usage
• Most searching is Google-like
• Platforms becoming portable, customisable widgets
  allow access from departmental web pages, course
  pages etc
• Any search of a library resource is better than
  Google.
• Quality and authority less important will
  Wikipedia do?
• Increasingly hard to get new sales of AI
  products
• Third World needs may be different

9
Use of AI
10
Conclusion

• Evolution and change
• Some species of AI will die out
• A few innovations (mutations) will flourish and
  dominate
• A more crowded information universe demands
  better tools to prosper
• Need to see innovative AI products, not just
  value added enhancements of same old products