Enterprise AI and data science

Speed up product developments, bring down costs, increase net new revenue and profit, convert more consumers…

We’ve delivered 12 AI products

…deployed 59 AI models, parsers, and algorithms for 15 clients in five countries

Our clients come from diverse industries

…including manufacturing, telecom, banks and financial services, consumer goods, meat products, health-tech, mining, education, economic research center, et al

Bangladesh

USA

Australia

UAE

Myanmar

Our clients are using AI to solve high-value problems

…which the industries have been trying to solve for years

Biponon / বিপণন

Tracking point-of-sale merchandise at 800,000 outlets

The country’s leading mobile financial service provider, bKash, has been using our trade marketing AI product, Biponon, since August 2018.

Using four object detection models, bKash tracks the executions and retentions of POS materials, monitors brand campaigns, manages inventory, accurately calculates field force salary—and maintains overall market superiority in this competitive landscape.

Using AI, bKash has gained a staggering 320% increase in productivity—serving 585 outlets by a merchandiser, up from 139.

Fordo / ফর্দ

Precision marketing to convert millions of consumers

Unilever has been using our precision marketing AI product, Fordo, since July 2019.

Built with repeat purchase data of 2.3 million unique consumers, Fordo recommends personalized product bundling and offers.

Unilever team has achieved successful conversion rates of 6% and above (often as high as 13%), up from 1% before using AI.

Shobdo / শব্দ

Speech recognition of sales calls at 200,000 outlets

A global consumer goods company used our speech recognition AI, Shobdo, for a year from December 2020 to analyze 5,500 sales representatives’ outlet calls at 196,670 unique outlets in real time.

The team was able to increase the effectiveness of their brand message deliveries from 15% to 75% using AI.

Jontro / যন্ত্র

Reducing manufacturing costs by optimizing spare parts replacements

A leading FMCG company has been piloting our manufacturing AI product, Jontro, since August 2022 to predict the optimum replacement time of production equipment.

Jontro detects microvariances in temperature, noise, and vibration of machines — and then predicts the remaining lifetime of equipment. It can generate substantial savings by helping manufacturing teams to implement predictive maintenance.

Jolchhap / জলছাপ

Preventing counterfeit products by embedding secret codes on packaging

A large manufacturing company has been piloting our another manufacturing AI product, Jolchhap, since March 2023.

Jolchhap uses image steganography to embed a secret message on product packaging, which consumers can scan to verify its authenticity. It is impossible to replicate the code from images and cannot be reprinted by copying or any other means.

For teams who are working to prevent illicit products, Jolchhap can be a powerful tool.

Borno / বর্ণ

Supercharge document processing by reading handwritten texts

Telenor, Ooredoo, and Wave Money have been using our handwritten text recognition (HTR) product, Borno, to complete KYC of more than 2.5 million users since May 2021.

Unilever team has also used this AI to glean critical market insights from 4.7 million handwritten shop receipts.

Chotur / চতুর

Fraud detection of identity documents

Telenor, Ooredoo, and Wave Money have also been using our fraud detection tool, Chotur, since October 2020 to verify passports, driving licences, and national identity cards of their subscribers.

Dharapat / ধারাপাত

CIB report and bank statement analyzers

IDLC Finance has been using CIB report module of our financial statement analyzer AI, Dharapat, since February 2022. The team has also been using the bank statements module since July 2023.

IDLC Finance has shortened the statement processing time from three working days to less than 30 minutes.

Prime Bank has signed up for the CIB report analyzer—deployment is currently in progress.

Pothik / পথিক

Efficient logistics management

A global dairy company used our route optimization AI, Pothik, to travel 953 km less per week (20% saving on fuel and time) in October 2020.

How can AI help your business?

  • Some notable experts group current AI capabilities into three categories: Optimize, accelerate/enhance, and transform.

    Optimize – ‘Make your internal efforts more efficient’

    Accelerate/enhance – ‘Improve your current product or service’ (for examples, Canva’s Magic Studio uses AI to ‘design, write, expand, morph’ images and videos and produce all these in the ‘brand voice’, localizes the content; and Khan Academy’ Khanmigo applies AI to ‘tutor students without giving answers’ in a Socratic fashion)

    Transform – Completely transform your business model, such as creating ‘new products or service lines’ (for example, Salesforce Einstein integrates AI ‘to generate sales materials, respond to customers in real-time, analyze pipeline and customer data’)

    We recommend reading academics and industry experts’ insights to form an informed mindset.

    You can also consider applying structured frameworks to evaluate/assess your own organization, function, and team to develop a coherent AI strategy.

  • One good place is Google’s showcase of Industry Solutions grouped by retail, consumer packaged goods, manufacturing, supply chain and logistics et al.

    With the rapid progress of AI, there are plenty of use cases for almost every industry. The question is more about identifying the right ones to pursue for your organization.

    Because, end of the day, it should be determined based on ROI, your business readiness, including the organizational culture, stability, and legal and regulatory compliance, among other critical factors.

    Talk with an expert to learn about AI’s inevitable impact on your business and stay in the driving seat by designing your AI strategy before being forced to follow suit.

  • We have published high-level details of AI projects for business leaders to help everyone understand the steps, timeline, and cost of AI projects.

    The document highlights all the key concepts your project leads need to consider and manage well.

    For instance, like any strategic project, it starts with defining a high-value business problem. (1 week)

    The next step would be completing a data infrastructure and data engineering analysis. (3 days)

    After that, the team would move to mapping the business problem with AI domains. Determining which AI capabilities would be required to achieve each component of the business solution. (1 day)

    Following this, the team would proceed to determine the solution development approach: check if existing AI models serve the purpose, or if fine-tuning an existing one with new data will do the job, or if we need to build custom modules on top of the existing one, or would we instead need to build one from scratch. (1 week)

    The document then continues to outline the remaining steps all the way to implementation. Hope you find it useful.

    And we’re happy to talk about our learning and share our experiences for worthy goals. Just give us a call.

How can we help you in your AI journey?

We’ll take care to understand your goals, priorities, and challenges. Then, we will work with you to define the use cases.

Next, we will work out suitable approaches and prepare the timeline and budget.

After that, we will build and deploy the AI and support you in completing the business readiness activities.

We’ll start by helping your team understand the nature of AI capabilities.

We will then work with your team to complete an evaluation of your business readiness and understand what types of AI capabilities we should target.

We will complete five more steps to help you define a coherent, practical, and robust AI strategy to give you a competitive edge.

Full spectrum of data science service—building data pipeline, data processing and cleaning, data exploration, data modeling et al.

But—important—we will all these so you can discover powerful insights and develop critical data capabilities.

For instance, you will be able to set individual functional KPIs based on key metrics that matter.