Seeding artificial intelligence for quality detection, cost reduction in agriculture sector

Agri-start up firm Agdhi relies on artificial intelligence (AI)-based technology to check quality of seeds. This technological intervention can detect seed defects within fraction of second as against normal procedure that takes around 3-6 months for seed producers to know the efficiency levels. Such technology also gets rid of manual inspection, which is resource intensive and subjective in nature. In an exclusive interview with

Bizz Buzz, the Bengaluru-based firm's founder Nikhil Das said that the company is currently in the process of raising funds from investors to scale up its marketing spend, which will augment its operations. The company, which till date remains bootstrapped, is also eyeing to serve the needs of small seed producing firms through its technology solutions. Currently, the seed industry is pegged at $50 billion market. Agdhi is trying to cash in from the needs of seed firms to save R&D costs that accounts for more than a quarter of their total cost

Agtech is the emerging space in the startup ecosystem with many players joining the fray. What is the unique selling proposition of 'Agdhi'?

There are many startups, which are coming up in the agtech space. But the main differentiator is our expertise in the hardware part. In agriculture, every operation has to be on the field and there are limitations to what can be done virtually. So, virtual software applications have a limited role to play in the agriculture space. Proper hardware solutions are required to resolve the problems of agriculture sector. Our hardware-enabled solution helps the seed industry detect the defective seeds through artificial intelligence-powered vision technology.

Currently, the seed industry needs three to six months to examine the efficiency of the seeds. For this, seed companies provide good quality seeds to growers and examine the quality of the seeds after the growth of plant. But the industry to accept the seeds, these seeds have to be examined through various parameters like genetic purity, size, and germination ability among others. So, the seeds to be accepted, it has to go through various steps, which becomes a time-consuming process.

Our solution offers a vision-based approach in which any genetic parameter can be examined at the seed-level itself. SeedVision, which is our solution, enables an objective and automated seed screening method that saves a lot of time and cost for the users.

What kind of value addition will happen for the seed industry, which often faces the problem of spurious seeds?

It brings in a lot of value addition for the seed industry. They can save on the land cost (for growing the crop for judging the efficiency) and resource cost. This also brings down overall cost of seeds, which helps farmers. So, the whole ecosystem gets enriched from this solution.

Which are the companies that have adopted your AI-based vision technology solution?

We are currently working with many domestic and international companies. I will not be able to name those firms. But there are a lot of advanced level activities that are going on. Not only big seed companies, our solution will benefit the smaller seed firms the most. This is because this will reduce time for trading cycle for those companies. Small companies will also see the expansion of the market by using our solution. Currently, small companies can't invest much on producing quality new seeds due to resource constraints. Our hardware infrastructure will enable these small companies to expand their market by selection of quality seeds through our solution. Currently, seed companies have to spend around 25-30 per cent of their total cost on the research and development (R&D). It also takes around four-five years to release a new seed to the market. That cycle can be brought down to 2-3 years, which will bring in massive gain for companies.

How much cost saving a seed company can get through your AI-based vision technology solution?

This machine costs around Rs 45 lakh and the payback for usual paddy seed producer will be around 9-10 months. So, in two seasons, the producer can get their money back. The processing time for the machine is 8-12 seeds per second.

Do you patent for this new technology?

Yes, we have applied for the patents, which is pending.

As the solution is hardware-based, this will be capital intensive as your company has to manufacture the units. Are you adequately funded as of now or planning to raise funds?

The seed industry is valued at more than $50 billion and there are many players operating in this space. So, the opportunity is huge in this industry. To tap the opportunity, we are planning to raise some capital in the near future and the process is on.

As we also require working capital for tapping the overseas market. So, the current funding round is mainly for investing more in marketing apart from meeting our working capital requirements. We are looking at raising around $1 million of funding in this round.

Have you raised funds before or will this be the first round of funding for the company?

We are bootstrapped till now and have not raised any money from investors. As promoters, we have invested around Rs 5.5 crore till now. We are not looking at much equity dilution at this point of time.

Do you have own manufacturing plant for producing the machines? What are your plans for ramping up the capacity?

Yes, we have a manufacturing unit in Bengaluru. Right now, we have three systems available with us. We are not looking at further investment in the manufacturing units as current capacity can take care of the new demand. As I have pointed out, majority of the capital will go towards marketing and working capital expenses. We are also tying up with some of the international partners to make our detection of seed qualities more accurate. As we get more data and feed those data into our AI platform, our algorithm will be richer with time.