This are my running notes from visiting Accenture’s Technology Labs in Sophia Antipolis, as part of a Master of Management program called "Strategic Business Development and Innovation" for the Norwegian School of Management.
Accenture’s Technology Labs is a relatively small organization: 200 researchers, 180000 employees in Accenture. There are four tech labs: Silicon Valley, Chicago (the largest), Sophia Antipolis, Bangalore, they should be able to do everything, but in practice there is specialization. The four main activities of the tech labs are technology visioning, research, development of specific platforms, and innovation workshops (with clients, press, consultants etc.) The themes pursued are mobility and sensors; analytics and insight; human interaction & performance; Systems Integration (architecture, development methods); and infrastructure (virtualization, cloud computing).
Kelly Dempski: Power Shift: Accenture Technology vision
The visioning used to be far-thinking, visionary etc., now have a much more immediate focus, want to look at things that you can implement today, make it much more "grounded in reality"
Eight critical trends:
- 1: Cloud computing and SaaS: Hardware cloud (amazon.com, IBM, Google (now the third largest producer of servers in the world)), desktop cloud (Google, Zimbra, MS Office Live Workspace), SaaS cloud (Netsuite, CrownPeak, salesforce.com), and services cloud (Google Checkout, Amazon web services, eBay, Yahoo)
- examples: Flextronics has changed over their HR applications to an SaaS model. AMD emulates chips on software for testing purposes, now contract with Sun to do that in the cloud. New York Times had 4Tb of articles that they wanted to translate to PDF: Translated it all twice (because there was a bug the first time), someone went on Amazon with their credit card, uploaded 4Tb, processed it (24h), there was a bug, had to do it again, 48h, total cost $250 on someone’s credit card.
- data location (where is the data)
- privacy and security
- 2: Systems – regular and lite
- SOA as the integration paradigm (regular), mashups (lite)
- traditional back-end apps vs. end-user apps
- small number of apps maintained by CIOs vs. large number of User and user-group created applications (long tail)
- REST is a light architectural approach for interoperability & data extraction
- Mashups (JackMe (trading platform tools), Serena, Duet (SAP and Microsoft), IBM) becoming more important in the enterprise arena
- Widgets and gadgets are light-weight desktop UIs that continually update some data
- 3: Enterprise intelligence at scale
- combination of internet-scale computing, petabytes of data, and new algorithms
- almost all the large systems vendors have partnered with or acquired some analytics oriented software company (such as Microsoft acquiring FAST)
- rampant use of data: evolution through access, reporting, external & internal, unstructured etc.
- Trends 1-2-3 together: The new CIO
- hardware and software procured from the cloud
- business units, end-users create their own lightweight apps
- The new CIO:
- "Data Fort Commander" – ensure security, privacy, integrity of corporate data and manage back-end apps
- "Chief Intelligence Officer" – provide data analysis services & insights to business units
- 4: Continuous access
- mobile device "first class" IT object
- No concept of enterprise desktop/laptop
- location-based services
- 5: Social computing
- amplify and support the value of the community
- three major directions: Platformization, inter-operability, identity management
- 6: User-generated content
- community-based CRM (users making videos about how to run certain kinds of software or build something from IKEA)
- new forms of entertainment
- revenue erosion of traditional media companies
- this has marketing implications: You can measure the sentiment out there in the user community. You switch from advertising to engaging.
- 7: Industrialization of software development
- converging trends will increase integration: Predictive metrics, model-driven development, domain-specific languages, service-oriented architecture, agile-development & Forever Beta.
- 8: Green computing
- global warming, energy prices, consumer pressure, compliance and valuation
- switch out energy-intensive processes for information-intensive processes: Electronic collaboration; Warehousing, supply chain & logistics optimization; Smart factories, plants, buildings & homes; and new businesses such as carbon auditing and trading
Cyrille Bataller: Biometric Identity Management
Biometric identification is coming, driven by increasing demand and technological progress. Biometric identification is defined as "automated recognition of individuals based on their physiological and/or behavioral characteristics. Physiological can be face, iris, fingerprint; behavioral can be signature, voice, or walk. Involves a tradeoff, as with all security systems, between the level of security and the convenience of the system. Fingerprint is most used (38%), face is the most natural, iris the most accurate. Many others: Finger/hand vein, gait, ear shape, electricity, heat signature, hand geometry and so on…
Balance between FMR (false (positive identification) m rate) and FNMR, called equal error rate. Iris has an EER of .002%, 10 fingerprints .01%, fingerprint .4%, signature 3%, face recognition 6%, voice 8%. Many parameters in addition to this.
Securimetrix has something called HIIDE, a mobile unit that does a number of biometrics, used in Iran. Voice is very interesting because it can be done over the phone, interesting for call centers, banks etc. Multimodal important, because it is hard to spoof.
Airports is a good example of what you can do with proper identification: You can move 99.9% of the check-in away from the airport. Bag drop can also be almost fully automated. Portugal is the leader in the EU, have automated passport control with facial recognition (scan, use electronic passport etc.). Most people are not concerned very much with privacy given some assurance and convenience. Likely to see lost of automated border clearance for the masses, but also registered travelers that go through even quicker and are interoperable across many airports. One common misunderstanding is that automated identity checking is moving away from 100% accuracy, but human passport/security control is an error-ridden process and mostly automated processes are more accurate.
Antoine Caner: Next Generation Branch
This is a showcase exhibit of best practice banking technology and processes. This showroom has about 40 companies (banks, mostly) visits per year.
Most banks have a multi-channel strategy, have returned from a strategy of getting rid of branches but want to redefine it. Rather than doing low-value transactions, the branches are seen as a mesh network for business development.
Key principles behind the branch of the future:
- generating and taking advantage of the traffic
- flexibility throughout the day
- adaptation to client’s value
- sell & service oriented
- modular space according
- entertaining and attractive
- focused on customer experience
- turning the branch windows into an interactive display (realty, for instance)
- Bluetooth-enabled push information
- swipe card at entrance to let branch know you are there, let your account manager know, apply Amazon-like features
- digital displays for marketing
- avatar-based teller services
- biometric-based ATMs to allow for more advanced transactions, as well as more opportunistic sales applications
- do both identification and authentication
- digital pen user interface for capturing data from forms
- RFID-based or NFC (Near Field Communication) in brochures, swipe and get info on screen
- "interactive wall" for interaction with clients in information seeking mode
- visual tracking of movement in the branch
- modular office that can change shape during the day, reconfigurable furniture
What impressed me was not the individual applications per se – though they were impressive – but way everything had been put together, with a back-office application that can be used by the branch manager to track how this whole customer interface (i.e., the whole bank branch) works.
Alexandre Naressi: Emerging Web Technologies
Alexandre leads the rich Internet applications community of interest within Accenture. He started off giving some background on Web 2.0 and used Flickr as an example of a Web 2.0 application, where a company use user-generated content and tagging to get network effects on their side. Important here is not only the user interface but also having APIs that allow anyone to create applications and to have your content or services embedded into other platforms. Dimpls is another example. More than one billion people have Internet access, 50% of the world has broadband access, which allows for richer applications. Customers’ behavior is changing – it is now a "read-write" web. It has also gotten so much cheaper to launch something: Excite cost $3m, JotSpot $200k, Digg cost $200.
Rich Internet Application and Social Software represent low-hanging fruit in this scenario. RIA allows the functionality of a fat client in a browser interface, with very rich and capable components for programmmers to play around with.
Two families of technologies: Jacascript/Ajax (doesn’t require a plugin, advocated by Google), and three different plugin-based platforms: Silverlight (Microsoft), Flash/Flex from Adobe, and JavaFX from Sun. All of them have offline clients that can be downloaded as well. A good example is Searchme.com, which gives a better user interface – Accenture has developed something similar for their internal enterprisesearch.
Social Software: Accenture has its own internal version of Facebook. Youtube is also a possible corporate platform where people can contribute screencasts of all kinds of interesting demos and prototypes.
Kirsti Kierulf: Nordic Innovation Model for Accenture and Microsoft
Accenture and Microsoft collaborating (own a company, Avanade, together), and have set up an Innovation lab in Oslo called the Accenture Innovation Lab on Microsoft Enterprise Search. Three agendas: Network services, enterprise search (iAD), and service innovation. Running a number of innovation processes internally. This happens on a Nordic level, so collaboration is with academic institutions and companies all over.
Have made a number of tools to support innovation methodologies: InnovateIT, InnovoteIT, and InnomindIT (mind maps), as well as a method for making quick prototypes of systems and concepts for testing and experimentation: 6 weeks from idea to test.
Current innovation models are not working for long-term, risky projects. Closed models do not work – hence, looser, more informal and open innovation models with shorter innovation cycles. Pull people in, share costs throughout the network, Try to avoid the funnel which closes down projects with no clear business case and NIH. Try to park ideas rather than kill them.
Important: Ask for advice, stay in the question, maintain relationships, don’t spend time on legalities and financials.