The Hindu Business Line : `Go for product development'

After more than three decades as Professor of ComputerScience at Stanford University and research scientist at MIT AI (artificial intelligence) lab, and having supervised over 40 Ph.D.s, here is Dr Thomas O. Binford, in India, along with his wife Ione.

The couple lives in Bangalore, not as retirees enjoying the mild weather of the Silicon Plateau, but as co-founders of Read-Ink Corporation. Tom is the Chairman and CTO of the company, and Ione, the CEO, responsible for setting the company strategies, building its overall business and marketing.

Ione holds an MBA, an MA and BA in psychology. Her experience includes 20 years at Hewlett-Packard in various management positions in marketing, business development, engineering and outsourcing. Her career also includes stints as CEO for ClickAccounts.com, EVP for Symmetric Security Technologies, and EVP of Marketing for ANTs Software Inc.

Read-Ink (www.read-ink.com) is a small, privately held company, based in Silicon Valley, and with an R&D centre in Bangalore. The company specialises in document analysis and handwriting recognition technologies.

"Twenty young engineers, mostly from the Indian Institute of Technology (IIT), India's premier technology school, peer into computer monitors in the no-frills office of Read-Ink Technologies, a start-up company housed in a small building in the bustling Indiranagar neighbourhood of this city," began a March 2006 story by Saritha Rai in The New York Times. The Binfords had arrived in Bangalore `four years ago with five suitcases', narrated Rai.

Here are the Binfords interacting with eWorld, even while battling with erratic power supply in the IT city.

What was the trigger for your decision to move your business into India?

We are self-funded and we wanted to stretch our funds. India had talented engineers. India was familiar to us and a friendly country.

About your products.

Read-Ink automates recognition of data from handwriting and machine print to convert the data into machine-readable form for enterprise systems and consumer mobile devices. We are near introducing products based on a revolutionary handwriting recognition engine with near-human accuracy, aiming at high 90's character recognition.

Near-human accuracy is required for business adoption and required for consumer acceptance for mobile devices. Humans are smart. They read well. Near-human accuracy is a very challenging standard. That is Read-Ink's goal. Read-Ink's recognition engine recognises cursive writing, handprint and machine print at a near-human level of accuracy. Our products are going to be multi-lingual. The first release will be for English. We intend to customise products for Indian languages, also.

What are the applications?

Users write on mobile devices. Enterprises have large volumes of paper forms. Governments have legacy paper records. Paper is going away slowly. Mobile devices are increasing rapidly. Users have not accepted handwriting input rapidly for mobile devices such as Tablet PCs. Studies show that user acceptance requires near-human accuracy. Correction is a pain. Correction is slow.

Enterprises have been slow to adopt automation for forms that are handwritten, typewritten and poor quality machine print. Low accuracy recognition increases human costs for review and correcting inputs. Automation becomes cost-effective for near-human accuracy in recognition.

How big is the market going to be for your products?

Customers use Smart Phones, Digital Pens, Tablet PCs, and PDAs. Read-Ink's accuracy exceeds the threshold for user acceptance for these mobile devices. Mobile device markets include both enterprise and consumer market segments. Some enterprise segments are the forms for hospitals, insurance, mortgage, field workers, lab notebooks and census. Examples of the consumer market segments are note-taking on Tablet PCs, PDAs, Digital Pens, and for text entry (email) on SmartPhones. There are 3 million Tablet PCs; there will be 17.2 million PDAs (by 2009), SmartPhones 123 million units. Digital Pens are a rapidly emerging market.

Enterprises process large volumes of input scanned from paper forms containing hand print, cursive handwritten and poor quality machine printed data. Enterprises and BPOs process medical claims, forms, and invoices, financial documents including mortgages, insurance, applications forms, and faxes. Governments are now converting handwritten paper surveys and legacy records in machine-readable formats. We hope that we can work with the Indian Government on conversion projects, e.g. for e-governance. Digital library projects have started massive digitisation and OCR conversion projects. The need for handwriting recognition in the medical and legal profession is growing rapidly. We aim to develop automation for mid-scale enterprises, also.

On the economics of using handwriting recognition.

With near-human accuracy, we expect our products will save a factor of 10 in cost of processing forms. Cheaper than China! Read-Ink will target large and medium enterprises and BPOs initially. We plan to offer an end-to-end solution for market niches, e.g. invoice processing and medical claims, by interfacing data capture from semi-structured and unstructured forms with enterprise workflow and ERP systems. Read-Ink is poised to dominate the market for products and services that vendors have promised for years but have failed to deliver because of the lack of accuracy and flexibility in machine recognition of handwriting and printed materials.

On what's good about today's young in India... and what's not-so-good.

Good: There is a good pool of young technical engineers in India. They are motivated and want to learn. For Read-Ink since we are working in a niche area we have to develop the small extremely technical pool to work in our area. We are very happy and proud of our technical team in India. It is more difficult to find, recruit and keep in India compared to Silicon Valley where startups like ours can attract the experienced and top graduates from the best universities.

Not so good: Personnel have high attrition rates and low expertise. They have short term focus on their financial packages instead of on their value to the company; focus on credentials instead of acquiring solid capabilities and experience. Hence the companies can invest less on their growth also.

Where does the future of software development lie?

The future in software development is highly technical complex systems that are extremely simple for the average non-technical user. Next generation products require experienced and knowledgeable software experts in the latest computer sciences, algorithm technologies. To compete in product development in addition to services India needs software experts with solid and longer term experience in real product development. Opportunities are much greater in product development.

How does a typical day as the CTO/CEO of your company in Bangalore go?

Tom, the CTO: Exercise, shower, breakfast and meeting and working side by side with technical team, from design to debugging. We have an excellent team of very bright and technically competent software experts. These days we are busy making changes and testing for accuracy. On our last mile there is a lot of hard work to get everything right.

Ione, the CEO: Managing, hiring, budget and keeping the operations going. A good part of the time spent also in marketing, relationship building, position and budget planning for product introduction. We are starting also to discuss and look for potential customers for large joint projects in the forms automation and processing. I run also a small accounting/BPO (ClickAccounts) for US customers to complement our self-funding. It is not a 9 a.m.-6 p.m. job.

Are there frontiers in your field that you are still pushing ahead? What are they?

Yes, to get near human recognition rates we have to work on a mix of machine learning, artificial intelligence, and computer vision using the latest technologies and developing and pushing the limits on the technology front.

Do you think India's policies need changes to encourage more of entrepreneurial activity?

I think India has done an excellent job is the services and application development area. For products it will need to develop longer-term view on return on investments and requires angel, government and VC (venture capital) funding with more risk available for product development.

MuraliDe@gmail.com