Turn unstructured communications into structured data that drives action.

Turn unstructured communications into structured data that drives action.
Turn unstructured communications into structured data that drives action.

The World’s Leading Cognitive Intent Engine

Made for enterprises with high volumes of communications that need this data operationally available and actionable at scale.

Used by these companies


"re:infer's AI platform lets us understand what our customers tell us – at scale. Ultimately it helps us make better product decisions.”

John Veichmanis

John Veichmanis

CMO, Farfetch

The road to intelligent automation starts here

We designed the most powerful cognitive intent platform to provide transformational analytics and automation opportunities. Discover the potential of advanced machine learning.

Front-office Client Analytics

Front-office Client Analytics

The only way to understand your customers is to listen. Take the first step towards driving sales, reducing churn and improving your product.

Advanced Operations Analytics

Advanced Operations Analytics

Most manual processes are transacted in a conversation. Uncover the hidden inefficiencies in your organisation by monitoring what’s being said.

Automated MI Capture

Automated MI Capture

Transcribing management information manually is unreliable, inconsistent, slow and expensive. Automate MI collection with AI.

End-to-end Automation

End-to-end Automation

Interpreting requests and collecting data manually introduces latency and it's unreliable. Automating interpretation unlocks massive cost savings.

Optimise in Real-time

Optimise in Real-time

Every exception to a product or service is surfaced in communications data. Measure and optimise change initiatives in real-time.

Front-office Client Analytics

Front-office Client Analytics

The only way to understand your customers is to listen. Take the first step towards driving sales, reducing churn and improving your product.

Advanced Operations Analytics

Advanced Operations Analytics

Most manual processes are transacted in a conversation. Uncover the hidden inefficiencies in your organisation by monitoring what’s being said.

Automated MI Capture

Automated MI Capture

Transcribing management information manually is unreliable, inconsistent, slow and expensive. Automate MI collection with AI.

End-to-end Automation

End-to-end Automation

Interpreting requests and collecting data manually introduces latency and it's unreliable. Automating interpretation unlocks massive cost savings.

Optimise in Real-time

Optimise in Real-time

Every exception to a product or service is surfaced in communications data. Measure and optimise change initiatives in real-time.

How re:infer is helping enterprises

Tier 1 Investment Bank

Group Operations


Challenge

COO’s office tasked with delivering a 50% reduction in operational costs in the next 5 years.

Our Role

re:infer’s cognitive intent platform identified quality-of-service issues, information security risks, STP breaks, manual and redundant processes, fulfillment pathways and cross-organisational opportunities for digital transformation: automation, self-service and remediation.

Impact


200 Million

Conversations Analysed

$135M

Projected Annual Savings

3,000,000

FTE hours per annum

100s

Facilitated end-to-end automated processes

How it works

Step 1

Import multi-channel communications data

We connect to your existing communications channels, be that emails, calls, chats or via your CRM platform and extract email threads, tickets and phone conversations to be analysed. Our technology works on any type of human communications data, including automatic speech-to-text transcription of phone calls as part of the integration.
Step 2

Automatically discover and classify intents

Our proprietary models have been trained on terabytes of text where using novel representation learning approaches researched at the University College London AI research lab. This neural model is then fine-tuned on each of our client’s own data to automate the discovery of intents, requests, entities and sentiment, creating a unique model tailored to every client.
Step 3

Active learning with minimal data

re:infer combines design thinking with breakthrough machine learning in a modern web interface. A user is guided to annotate auto-discovered intents in a continuous active learning loop that maximises information gain for the supervised models at every step. We are democratising AI so that non-technical users can create ML models at the click of a mouse requiring only tens of examples to achieve high accuracy.
Step 4

Optimise and drive automation

re:infer provides granular insights into processes, workflows, customer issues and automation opportunities via an advanced analytics dashboard—comprehensive management information, cohort comparison and sentiment analysis. With re:infer’s structured, machine-readable annotation of every verbatim, you can enable end-to-end automation by connecting re:infer to other parts of your IT real-estate, such as downstream Robotic Process Automation (RPA) installations.
How it works
How it works

re:infer is based on Breakthrough Research

Enterprise Architecture

Reliability at Scale

re:infer's machine learning platform was built from day one to scale horizontally to billions of conversations while maintaining millisecond latency.The data storage is replicated and durable, exposing a strongly-consistent (CP) view of your data, which ensures correctness even in the face of node failures and network partitions.Even in the unlikely event of catastrophic failures affecting a majority of nodes, the real-time prediction services continue to be available.

Enterprise-Grade Security

We use best defense-in-depth practices throughout our systems, guided by the principle of least-privlege.All data stored with re:infer is encrypted in flight and at rest.Fine grained role-based access controls which integrate with your organisation's existing identity management solution (e.g. LDAP).All authentication is done using strong key derivation functions, two-factor authentication and account lockouts.

Flexible Deployment

re:infer works best as a cloud service, with immediate support, automatic security patches and continuous feature deliveryFor businesses whose requirements mean that a SaaS deployment is not an option, re:infer can be deployed in a 'Private Cloud' or on-premises.Our engineers will work closely with your IT team to ensure a smooth deployment regardless of the option you choose.

Enterprise Architecture
Enterprise-Architecture

In the news

The hottest machine learning startups in UKThe hottest machine learning startups in UK

The hottest machine learning startups in UK

The year 2017 saw a host of tech giants acquire AI startups, as the likes of Google...

Read Article
At your service: the robots saving e-commerceAt your service: the robots saving e-commerce

At your service: the robots saving e-commerce

In customer relations, call centres are expensive but computers have lacked the human touch – until now...

Read Article
AI startup re:infer gets $3.5m to develop cognitive…AI startup re:infer gets $3.5m to develop cognitive…

AI startup re:infer gets $3.5m to develop cognitive…

re:infer, a London-based artificial intelligence startup delivering cognitive autom...

Read Article