Breaking
Accenture acquires Avanseus AI technology - Feb 2026 Cognitive Assistant predicts network faults 7-30 days in advance Founded 2015 in Singapore - R&D in Bengaluru Patented unsupervised learning from as little as 3 months of alarm data ~46 employees - ~US$6.2M raised across seed, bridge and Series A Available on Red Hat & VMware marketplaces Accenture acquires Avanseus AI technology - Feb 2026 Cognitive Assistant predicts network faults 7-30 days in advance Founded 2015 in Singapore - R&D in Bengaluru Patented unsupervised learning from as little as 3 months of alarm data ~46 employees - ~US$6.2M raised across seed, bridge and Series A Available on Red Hat & VMware marketplaces
Company Profile / Artificial Intelligence

Avanseus

The Singapore deep-tech company that taught machines to see a network fault coming - before the customer ever does.

AIOpsPredictive MaintenanceTelecom AIFault PredictionAutonomous Networks
Avanseus logo
AVANSEUS - founded 2015, Singapore, with R&D in Bengaluru. Acquired by Accenture in February 2026. Wordmark: company logo.
Founded
2015
HQ
Singapore
Team
~46
Total Raised
~$6.2M
Stage
Acquired
Sector
Enterprise AI
The Story

Reading the pattern before the outage

Every large network hums with warning signs. Alarms flare, degrade, and clear thousands of times a day, and most of them mean nothing. Buried inside that noise is the one signal that matters: the fault that is about to take a cell tower, a backbone link, or an enterprise IoT deployment offline. Avanseus was built to find that signal.

Founded in Singapore in 2015, Avanseus Holdings set out with an unusually specific ambition - not to build another dashboard, but to predict network faults and performance degradations before they happen. Its flagship product, the Cognitive Assistant for Networks, forecasts failures 7 to 30 days in advance and recommends the likely root cause based on historical fault patterns.

The technical wager underneath the company was that you do not need years of labeled data to do this well. Avanseus developed a patented, neural-network-based unsupervised learning algorithm that can start predicting from as little as three to six months of historical alarms - and, in the process, tackled a stubborn problem in AI known as long temporal dependency.

That combination - fault prediction on minimal data, delivered non-intrusively across everything from legacy wireline to 5G - is what set Avanseus apart. In February 2026, Accenture acquired the company's AI technology to accelerate its telecom clients toward autonomous networks. For a roughly 46-person deep-tech team, it was a quiet, defensible kind of win.

"Accenture's acquisition of our AI solution marks an important next chapter for the technology we have built."

Bhargab Mitra - Co-founder & CEO, Avanseus

What It Does

From raw alarms to a maintenance schedule

The Cognitive Assistant for Networks watches historical fault data, learns each network's own patterns, and turns that into an advance warning - plus a ranked view of which faults actually matter.

// Fault prediction window

Today
+7 days: early alert
+30 days: forecast horizon
predicted fault ▶

Configurable 7-30 day advance warning, built from as little as 3-6 months of historical alarm data.

Data needed
3-6 months
Advance warning
7-30 days
Tech coverage
Legacy wireline → 5G
Products & Services

The cognitive network stack

Flagship / 2019

Cognitive Assistant for Networks

Cloud-native AI/ML platform that predicts faults and degradations 7-30 days ahead and recommends root causes. Non-intrusive, spanning mobile, fixed, IP/MPLS backbone, transmission and enterprise IoT.

Core IP / patented

Universal Prediction Algorithm

A patented neural-network-based unsupervised learning engine that forecasts from as little as 3-6 months of alarms and addresses long temporal dependency in AI.

Models

Anomaly Detection & Optimization

Models for anomaly detection, cross-domain correlation, impact analysis and optimization to lift operational efficiency in complex networks.

Actioning

Impact-Based Actioning

Fault impact estimation, impact scoring and prioritization so operations teams act on the faults that matter most - not the ones that shout loudest.

Who It Serves & Why It's Different

Built for the ops team, not the demo

Who uses it

  • Telecom service providers running mobile, fixed and backbone networks
  • Industrial and IoT (IIoT) enterprises with complex asset fleets
  • Network operations centers seeking to shift from reactive to proactive maintenance
  • A lean, high-value B2B customer base served by a ~46-person team

What sets it apart

  • Predicts from as little as 3-6 months of data - a deliberate cold-start design
  • Non-intrusive: no rip-and-replace integration into live network gear
  • Unsupervised learning, so it works where clean fault labels don't exist
  • Technology-agnostic across generations, from legacy wireline to 5G
  • Prioritizes by impact, not just probability

Where it fits: Avanseus sits in the AIOps and network-analytics market alongside players such as Anodot, Moogsoft, BigPanda and Nokia's AI tooling - but with a sharper focus on advance fault prediction as a building block for autonomous networks. That focus is precisely what drew Accenture, which folded the technology into its cognitive network platform.

Milestones & Funding

A capital-efficient climb

2015

Founded in Singapore

Bhargab Mitra and co-founders launch Avanseus, backed by a US$2.5M seed round from SEEDS Capital.

2017

Patented prediction algorithm

The universal prediction algorithm - unsupervised, neural-network-based - takes shape.

2019

US$1.3M bridge round

Convertible-note financing led by TNB Aura, with SEEDS Capital participating, as Cognitive Assistant for Networks gains traction.

2020

Series A

A roughly US$2.4M Series A with TNB Aura and SEEDS Capital to scale the platform.

2021

Marketplace availability

Listed on Red Hat and VMware marketplaces and certified on Red Hat OpenShift.

2026

Acquired by Accenture

Accenture acquires the AI technology to accelerate autonomous-network journeys for telecom clients.

The People & The Details

Who built it

Founders & leadership

  • Bhargab Mitra - Co-founder & CEO
  • Giuseppe Donagemma - Co-founder & Chairman
  • Chiranjib Bhandary - Co-founder
  • Mei Lan Ng - Co-founder
  • Rajendra Panda - Co-founder

Fun facts

  • Can begin predicting faults with as little as three months of alarm history
  • HQ in Singapore, but the core R&D engine runs in Bengaluru, India
  • Predictions span technologies from legacy wireline to 5G
  • Raised only ~US$6.2M total, yet built patented AI that drew an Accenture exit
FAQ

Common questions

What does Avanseus do?

Avanseus builds AI-based predictive maintenance software that forecasts faults and performance degradations in telecom and industrial IoT networks, and recommends likely root causes, so operators can fix issues before they cause outages.

What is Avanseus's main product?

Its flagship product is the Cognitive Assistant for Networks (CAN), a cloud-native AI/ML platform that predicts network faults 7 to 30 days in advance using a patented unsupervised learning algorithm.

Who founded Avanseus and where is it based?

Avanseus was founded in 2015 and is headquartered in Singapore, with an R&D center in Bengaluru, India. Bhargab Mitra is co-founder and CEO; other co-founders include Giuseppe Donagemma (Chairman), Chiranjib Bhandary, Mei Lan Ng and Rajendra Panda.

Did Accenture acquire Avanseus?

Yes. In February 2026, Accenture announced it acquired Avanseus's advanced AI technology to strengthen its cognitive network platform and help communications companies accelerate their journeys toward autonomous networks.

How much funding did Avanseus raise?

Avanseus raised roughly US$6.2M in total, including a US$2.5M seed, a US$1.3M convertible-note bridge in 2019 led by TNB Aura with SEEDS Capital, and a Series A of around US$2.4M.

Share & Explore

Spread the signal