Griaule
Biometrics company’s success story combines incubation,
qualified professionals and financing from public agencies

While studying electrical engineering at the Federal University of Goiás (Universidade Federal de Goiás, UFG), Iron Calil Daher associated with José Alberto Fernandes Canedo — who then studied computing engineering — to open a company in Goiânia, in 1999. The young men, recalls Daher, wanted to make “something that would be at the same time different, stimulating and easily exportable.” They decided to develop a software for biometrics, that is, for identification through digitals, face, hands, iris, voice, signature. They founded Griaule and together created the algorithm necessary for identifying digitals. However, according to Daher there was a problem for expanding the company: in Goiás, Griaule was removed from the information technology (IT) cluster – and thus removed from business opportunities.

The chance to come to Campinas — which is an important IT cluster — came in 2002, with the first edict calling companies to the then recently created Incubator of Knowledge-Based Companies of the State University of Campinas (Incubadora de Empresas de Base Tecnológica da Unicamp, Incamp). The partners opened a new Griaule and got one of the places available. In the incubator they came in contact with what Daher describes as “crucial” for the company’s growth: the State of São Paulo Research Foundation (Fundação de Amparo à Pesquisa do Estado de São Paulo, Fapesp). With funds from Fapesp through the Technological Innovation in Small Businesses Program (Programa Inovação Tecnológica em Pequenas Empresas, Pipe) and scholarships from the National Council for Scientific and Technological Development (Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq), they were able to buy equipment, pay consultants and hire personnel to improve their old algorithm. Today, Griaule’s system of identification of digitals is used in the States of Goiás and Tocantins, and is being implemented in the States of Espírito Santo and Mato Grosso. Abroad, the main clients are the United States and Mexico. Recently Costa Rica bought the technology for the emission of passports.

Out of Incamp since July of 2005, Griaule’s total sales have increased six-fold compared to 2004. It earned national exposure when it was featured in an article about innovating enterprises published this year by Veja, Brazil’s largest weekly news magazine in circulation. Daher does not reveal the company’s total sales – all he says is that he wishes they would be higher. But he says that half of them come from the products that have been developed with the support of fostering agencies. The combination of funds from the Fapesp and the CNPq was so successful that Griaule is going to repeat it with its new project of facial recognition: the former has already approved a Pipe phase 1 for the study of viability and the latter will grant new scholarships. Holders of scholarships have participation in the company’s results.

The future components of Griaule’s software for facial recognition must comply with the pattern of the International Civil Aviation Organization (ICAO). This shouldn’t be a problem for the company: it already has eight certifications from the Federal Bureau of Investigation (FBI). In addition, in 2003 its technology for the recognition of digitals was considered the eighth best in the world in Fingerprint Vendor Technology Evaluation (FpVTE) test for large scale comparisons — one billion comparisons —, carried out by the United State’s National Institute of Standards and Technology (Nist). Griaule was the only company from the Southern Hemisphere to take part in the test. “Companies such as NEC and Motorola were ahead of us, but we left behind some big ones, such as Raytheon”, says Daher proudly. “Without money from the agencies probably we wouldn’t even have participated.” The goal for the next test, to be carried out in Italy later this year, is be number three.

In the next years the partners want to turn Griaule into “a big international player in biometrics.” Quoting figures from the International Biometric Group, Daher says that this market has been growing more than 30 percent a year. The participation of exports in Griaule’ sales is significant: 80 percent. But in the partners evaluation, in order to grow it’s time again to go after the cluster: later this year Daher is moving o San Jose, California, where the company has opened a branch. According to him, it is there, in the Silicon Valley, that Griaule’s clients are concentrated: large integrators that buy software components to mount their solutions directed to the final customer — from governments that need to have the civil and criminal records of its citizens to small establishments that use digitals to identify their costumers and employees. The choice of being a supplier to integrating companies is a business model: "It’s the industry, the manufacturer that typically carry out research and development – and normally the manufacturer doesn’t sell to the final customer,” he explains.

In the interview below, given on June 27 to Mônica Teixeira and Rachel Bueno at Griaule’s headquarters in Campinas, Daher describes the business, tells the history of his company, details the plans for the future, and assures: research and development activities will continue to be carried out in Brazil. Today, 16 of the 20 people who work at Griaule deal with R&D, including eight PhDs, two masters and one Master’s candidate. With the exception of two doctors who are involved in the development of a digitals reader, one of them a chemist and the other a physicist, the professionals of the R&D team are all from the computing area — and only one has not come from Unicamp. Here in Brazil, emphasizes Daher, Griaule finds brains as good as abroad, has a strong link with the University – where it can recruit its professionals – and, while still considered a small business, is entitled to financing from the Fapesp.

Why doesn’t Griaule sell its softwares to the final customer?
That would involve the creation of a much stronger commercial layer. Our strength is research and development of technology. We develop software components for companies that integrate the components to a product. That way all we have to find are integrators. We sell all over the world without moving from here – it’s the integrators that serve the final customer and knock from door to door; they have to go to him to implement the systems and solve their problems.

Which are Griaule’s main products?
Our two big markets are the corporate’s and the public security’s. For the corporate market we have the Desktop Login, which replaces the password with digital impression when making the login in computers; the Desktop Identity, for sales points and presence control; and the Rex 3, an equipment that opens doors to registered individuals. For government customers we have the Afis, which makes the recognition of digitals in a large scale environment – that is, where there are millions of records.

What is the difference between products directed to the corporate market and those that are directed to government customers?
In a public security system, the primary concern is to find the digital impression – and sometimes what you have is just a fragment, an impression in the crime scene, which makes localization more difficult. That would never happen in a corporate system, in which the primary concern is the speed. Another difference is the size of the databank. All States in Brazil have millions of digitals – the number of residents times ten, because each person has ten digitals. For that reason we have developed the Speed cluster, which is a technology for processing the databank on the side – with a set of dozens of computers that accelerate the search. The corporate solution is typically limited to the thousands of digitals – of employees, of visitors.

Which are the applications of the public security systems?
Civil and criminal identification, digitalization of collections, document emissions, border and prison control and welfare benefit control. The first 20,000 ballot boxes with digital impression readers were delivered to Brazil’s Superior Electoral Court this year by a company called Procomp, which used Griaule technology. In the next elections [scheduled for October], the ballot boxes will already have the readers, but they won’t be used because the voters’ digitals have not been registered yet. The Court is going to register everyone in order to prevent frauds: currently it is possible for a person to get several voting titles and vote several times.

Why did you come to Campinas?
We were approved for the incubation at Unicamp and that’s what brought us here. In technology, and in practically any sector of the economic activity, you must be in the cluster. In agribusiness, the States of Goiás and Mato Grosso are important clusters. In computing the cluster is the State of São Paulo: Campinas, São Paulo, São José dos Campos and São Carlos. The competitiveness of those who are not in it is seriously jeopardized. If we had stayed in Goiás we wouldn’t have grown as much as we did. For us incubation was crucial.

Why?
Most important was the fact that we were in a top university – in computing, Unicamp is practically the best in Brazil –, in contact with top students and top professors. The environment of ideas was very good at Incamp. In the area of information technology, in order to compete globally you need the best brains. You also need money. In São Paulo there’s the Fapesp. There’s no other State in Brazil with a research fostering agency as active – some don’t even have a research fostering agency.

When you applied for the Incamp, did you have the Fapesp in mind?
No, our main goal was to be in the cluster. We only learned of the Fapesp when we got here, and it was crucial for our development. In 2003 we had a project approved through the Pipe, and later another one. We have finished both. Recently the Fapesp approved through the Pipe a project for facial recognition, but it’s a phase 1, that is, a study of viability, because we still don’t have the product. We have another project for the first phase, whose objective is the development of a Brazilian-made reader of digitals – all readers available here are imported. We’re waiting for the answer. This project was not approved the first time we submitted it. From the CNPq we have three RHAE [Recursos Humanos para Atividades Estratégicas, or Human Resources for Strategic Activities] scholarships, which total some R$ 114,000 [approximately US$ 52,000]. They last two years and will expire in December. From the Finep [Financiadora de Estudos e Projetos, or Brazilian Innovation Agency/Research and Projects Financing] we received R$ 250,000 [about US$ 114,000] for a project approved in 2003. The project was carried out in 2204 and 2005 in partnership with Unicamp.

Was the financing from the Pipe the first funding Griaule received?
We got almost simultaneously the funds from the Pipe and the Finep, but for different objectives. The objective of the 2003 Pipe, for which we received R$ 250,000, was to improve the quality of recognition of our algorithm and to improve the set of computers that process activities on the side. The project included the installation of a pilot in the State of São Paulo’s Secretary of Public Security. We were unable to do it, and the pilot went to the State of Mato Grosso. Mato Grosso is an important customer because it has correctly digitalized its more than one million identification cards. It’s going to be out main pilot.

How important for Griaule was Pipe’s first financing?
The company would not be able to make the same investment – especially for two years, because the benefits take time to come back. The Pipe and the RHAE complemented each other perfectly: the first was for purchasing equipment and paying consultants, and the second for hiring researchers with scholarships. At first we didn’t apply for scholarships from the Pipe, but we ended up applying for them. The result was so good we’re going to repeat that format with the project of facial recognition, which has been approved by the Fapesp and also by the CNPq for the scholarships.

The researcher who has a scholarship is not an employee. How does Griaule deal with that?
We are implementing now a policy of participation in the results. For us, the scholarship is just a way of making the work of the researcher viable. The researcher participates actively in the company. Even trainees participate – and holders of scholarships are essential. We don’t want anyone leaving when the scholarship expires – if it’s the case, we pay the researcher ourselves. The knowledge the researcher has is very important, it took him a long time to acquire it. We have never had a turnover of scholarship holders, to this day no one has left the company.

What proves that Griaule is a success?
Our exports, which account for 80 percent of our sales. They prove that we are globally competitive. Our main clients are the United States and Mexico. We had a peak in sales for Costa Rica, which bought our technology for the emission of passports through the French integrator Oberthur, which makes passports for 80 countries. In 2003, our technology ranked eight among the best in the world in the Fingerprint Vendor Technology Evaluation (FpVTE) large scale test — one billion comparisons of digital impressions —, carried out by the Nist [National Institute of Standards and Technology], of the United States. Griaule was the only participant from the Southern Hemisphere. Companies such as NEC and Motorola were ahead of us, but we left behind some big ones, such as Raytheon. Without money from the agencies probably we wouldn’t even have participated. In October we’re going to take part in another test, in Italy, called FVC 2006 [Fingerprint Verification Competition]. Thanks to the advances we’ve made through the Pipe and the RHAE we believe we’ll be among the first three.

What are Griaule’s total sales?
Our total sales are quite reasonable, but they’re low compared to what our clients expect – they don’t reach R$ 100 million (US$ 46 million) a year. When we left the incubator we already had one product and total sales were very small. I can say that between 2004 and 2005 total sales increased three-fold, and from 2005 to 2006 they doubled.

How much of this jump in total sales was gotten thanks to Pipe?
The Pipe and the RHAE helped us develop a much better product. We improved the quality of the product we already had, and that led to an increase in sales. Today about 50 percent of our sales come from the products we’ve developed with the Pipe and the RHAE scholarships. Without those funds, if the company had invested all the necessary resources needed for the research we would have neglected the commercial part and Griaule would no longer exist. The product was already being developed, but without the resources from the Fapesp and the CNPq we would not have gotten the results we got.

Does Griaule have international certifications?
We have eight WSQ [Wavelet Scalar Quantization] certifications from the FBI and a certification from Microsoft for our set of computers that process activities on the side: Certified for Windows 2003 Server — that is, our product runs well on Microsoft’s Windows 2003. In Brazil, Griaule is the leading local company in biometrics technology and the only one that has its own technologies. Last year we ranked third in the Southeast in the Fapesp Award for Technological Innovation [Prêmio Finep de Inovação Tecnológica].

What did Griaule lose when it graduated from the incubator?
The incubator offers a number of services that are missed once you’re out of it, such as consulting services – outside the incubator, consulting services by good professionals are very expensive. But we haven’t lost access to the financial resources of the research fostering agencies nor cut the ties with Unicamp, which continue to be strong and are most important.

What’s Griaule policy regarding intellectual property?
We register softwares and want to start patenting – Inova [Agência de Inovação da Unicamp, or Unicamp’s Innovation Agency] may help us with patents. We have five registered softwares, all of them large and complex.

Do the PhDs who work at Griaule publish papers?
No, because they don’t have the time, but we want them to publish. Papers are important for the relationship with the university and to get financing. No paper will reveal the secrets of our business. On the contrary, they give a general idea of what we do and this brings credibility to our researcher and to our company. Our international competitors, who have a larger structure, publish papers and have patents in the area of information technology.

What’s Griaule goal?
To become a large international player in biometrics in the next few years. Contrary to our competitors, which are more active in specific areas, we should expand our line of products to cover other types of biometrics – we started with the face, the next one should be the voice. The critical factor for that to happen is financing – risk capital or financing through debt. There’s no question that what pushes a company’s development is capital. Good brains with good ideas are just not enough. The ideas need to be implemented, and then the products must be sold – and selling is more expensive than developing.

Will Griaule choose risk capital or financing through debt? If it chooses risk capital, will it accept to have an investor inside?
We’re going after risk capital, in Brazil or abroad. We are already talking with a Brazilian investor; if we close the deal he is going to bring in capital and will put people to work in the company. That’s going to be good because we’ll have well-qualified people in the management area, which I don’t have today. These people also bring good networking, because they have strong ties with large companies.

Being in Brazil is a limiting factor for Griaule?
In terms of business, yes. If you want to do business globally in this area you have to be in California, in Silicon Valley — that’s where the cluster is. All the large integrators, or at least most of them – which are my clients – are located in Silicon Valley. That’s why we opened a branch in San Jose, where I’m going to in December. In terms of research, being in Brazil is not limiting, because our brains are as good as the European’s, the Americans’ or anywhere in the world — and here there’s less competition for them. We’ll continue to carry out research and development here. In addition to good brains, we have ties with the university that we don’t have abroad.