Radware today announced it would purchase Bengaluru, India-based bot management firm ShieldSquare. The deal is expected to close in the first quarter of 2019. Financial terms were not disclosed.

ShieldSquare, founded in 2014, is a security firm that specializes in bot management. It delivers an API-based service that can detect and eliminate “bad bots” from websites, mobile applications, and APIs. The technology detects bots and relies on a cloud engine to classify visitor activity as either human, search engine crawler, or bad bots to secure the application.

Bot management — according to Michael Groskop, Radware’s vice president of product management portfolio — refers to the detection of malicious bot activity and the practice of mitigating bot attacks by differentiating legitimate flows from bots. It also must allow differentiation between good and bad bots. He noted that as bots grow more sophisticated, bot management must include “advanced behavioral detection and intent analysis,” delivered with machine learning, which ShieldSquare provides.

ShieldSquare technology will be offered as a new product line called Radware Bot Manager. The product line will integrate with Radware’s additional attack migration products, particularly its web application firewall (WAF) cloud services.

Groskop said it selected ShieldSquare for its machine learning capabilities, the way it expands its existing security services, and the “strong technology synergy” between the two firms. Additionally, he said that ShieldSquare’s technology “sets a stronger path for us in the context of API and mobile application protection.”

The new product line of bot management technology will boast a number of tools. This includes an out-of-path, inline-capable bot detection engine, deep behavior analysis that identifies attack intent and stops automated attacks, and anti-bot feeds to allow both proactive and preventive bot management. The technology will also rely on an extensive database of bot fingerprints to identify and detect threats to applications. The fingerprints were derived from threat intelligence gathered from over 80,000 internet properties spread across 70 countries, said Groskop.