Thank you for trusting Ada with your customers’ personal data. We take this responsibility very seriously and make every effort to be transparent and careful when handling this data on your behalf.
Ada uses industry standard technologies and services to secure your data from unauthorized access, disclosure, inappropriate use, and loss of access. We ensure that the security policies of all our subprocessors are documented and up-to-date with industry compliance standards where required (PCI, GDPR, etc).
Security at Ada is overseen by our Data Protection Officer and carried out by our entire team.
If you would like to report a vulnerability, please contact email@example.com with a proof of concept, list of tools used, and the output of the tools.
If a security disclosure is received, we will work quickly to reproduce each vulnerability to verify its status before taking the steps needed to remedy.
Compliance and Certification
Ada’s payment and card information in the Ada Dashboard is handled by Stripe, which has been audited by an independent PCI Qualified Security Assessor and is certified as a PCI Level 1 Service Provider.
Ada has been certified as a PCI Level 1 Merchant by an independant PCI Qualified Security Assessor. Ada will provide a copy of an Attestation of Compliance (AOC) upon request to Enterprise customers.
Ada is compliant with the GDPR. If you have customers who reside in the European Union and use Ada then we recommend that you sign a Data Processing Agreement (DPA) with Ada. This document is a contractual agreement that recognizes Ada as being GDPR compliant and makes your organization GDPR-compliant when it comes to using Ada as a sub-processor. We provide our DPA as a self-serve document here.
In an effort to provide the best security for all our customers when it comes to personal information, Ada treats all data as if it is bound by GDPR regulation.
Any person (including EU residents) wishing to submit a personal data request to Ada may do so by sending an email to firstname.lastname@example.org explaining their data request.
If you wish to contact our Data Protection Officer (DPO) for any concerns around Personal Data collection or usage, they can be reached at email@example.com.
Infrastructure and Network Security
Ada infrastructure is hosted on Amazon Web Services (AWS). The AWS data centers are equipped with multiple levels of physical access barriers, that include:
- Outer Perimeter Fencing that is crash-rated for vehicles
- Electronic Access Cards
- Video Surveillance
- Internal Trip-Lights
For more information on AWS Security features, you can refer to this whitepaper. Ada employees do not have physical access to AWS data centers, servers, network equipment, or storage.
The location of the AWS servers where we run our infrastructure depends on where your bot is deployed. Unless specified, all bots are deployed in N. Virginia, USA.
We are not able to provide the exact physical address of the data centre as Amazon has historically been quite reticent in publishing location information of their facilities for security reasons.
We currently run Ubuntu 18.04 on all our servers and use a combination of automated and manual inspection to determine if new vulnerabilities are introduced in the software packages on our systems. Our Infrastructure team ingests security bulletins and prioritizes remediation according to our internal Security Vulnerability Identification documentation.
Logical Access Control
Ada has full control over all its infrastructure on AWS, and only authorized Infrastructure Team members at Ada have access to configure infrastructure when needed in order to add new functionality, or respond to incidents. All access required for control of infrastructure has mandated two-factor (2FA) authentication. The levels of authorization for infrastructure components is mandated by the principle of least privilege.
Ada undergoes grey box penetration testing conducted by an independant third-party agency on an annual basis. For grey box penetration testing, Ada will provide the agency with an overview of application architecture and information about system endpoints.
Information about any security vulnerabilities successfully exploited through penetration testing is used to set mitigation and remediation priorities.
Amazon Web Services undergoes third-party independent audits and can provide verification of compliance controls for its infrastructure. This includes, but is not limited to, ISO 270001, SOC 2, and PCI.
It’s important to know when suspicious activity is occurring on Ada's infrastructure. We employ Intrusion Detection and Prevention systems (IDS/IPS) on each host under our control. This notifies us on common alert channels whenever suspicious activity may occur. Our infrastructure team will check each alert, investigate the activity, and then respond accordingly.
Business Continuity and Disaster Recovery
Every part of the Ada service uses properly-provisioned, redundant servers (e.g., multiple load balancers, web servers, replica databases) in the case of failure. All our deploys are zero-downtime deploys using Kubernetes, and we implement gradual rollout and rollback of services in the case of deployment errors.
Ada keeps continuous backups of our production databases using the MongoDB Atlas Fully Managed Backup Service. These backups are typically just a few seconds behind the operational system¹, allowing us to restore easily to any time in the last 24 hours in the case of data corruption or loss.
Ada stores all infrastructure as code and as such is able to bring up complete copies of production and staging environments quickly (currently < 12 hours and always improving!). In the event of a complete region-wide outage, the Ada Infrastructure Team will bring up a duplicate environment in a different AWS region.
Data into System
Ada provides an embeddable web window for use on our clients’ websites for users to interact with a client's personal chatbot. This chat window will send data back to Ada's APIs over TLS 1.2 or greater. The chat window assets use a subresource integrity (SRI) check to ensure that the files fetched from our CDN are cryptographically verified to prevent Man-In-The-Middle attacks.
Data through System
Data is sent from end-user chat platforms to the Ada backend via TLS 1.2. All data is AES-256 encrypted at rest.
Ada's latest SSL Labs Report can be found here.
Data out of System
Ada maintains intelligent network firewall rules at the infrastructure level that limit the surface for data extraction. We scrutinize our preferred partners and integrations to ensure that they comply with necessary security regulations (GDPR, PCI, etc), before transferring data for processing.
Data Security and Privacy
All data in Ada servers is automatically encrypted at rest using AWS EBS Encryption using Ada’s master encryption key stored in AWS Key Management Service. All volumes are encrypted in AWS using the industry-standard AES-256 algorithm.
Ada only ever sends data over TLS 1.2 or greater, and never downgrades connections to insecure early TLS methods like SSLv3 or TLS 1.0.
Data may be retained after termination of service according to specification within our main customer contract. If data is kept after termination of service for machine learning training purposes Ada will scrub all personally identifiable information (PII) from customer data. This includes, but is not limited to, usernames, emails, phone numbers, credit cards, IPs.
Ada currently supports redaction of personal information. If you would like this enabled for your bot, please contact your customer success representative.
In addition to password login, two-factor authentication (2FA) provides an added layer of security to Ada via a time-based one-time password algorithm (TOTP). We encourage 2FA as an important step towards securing data access from intruders.
Ada supports 2FA for all user accounts. 2FA can be enabled for a user in the Profile section of the Ada dashboard.
In the settings page, we include an Activity section where dashboard Owners and Administrators can view the editing history of Agents. This is listed chronologically so you'll have insight into the organization's most recent activity within the Ada dashboard.
Secure Application Development
Ada practices continuous delivery, which means all code changes are committed, tested, shipped, and iterated on in rapid sequence. A continuous delivery methodology, complemented by pull request reviews, continuous integration (CI), automated security scanning, and automated error tracking, significantly decreases the likelihood of a security issue and improves the mean response time to security vulnerabilities. Internally, Ada enforces at least one authorized reviewer for all code changes, and deployments to our production environment are gated under condition that all code is reviewed.
Ada uses the NIST CyberSecurity Framework (CSF) to guide and manage our cybersecurity-related risks. The NIST CSF is a policy framework that was developed by the U.S National Institute of Standards and Technology to help private sector organizations asses and improve their ability to prevent, detect, and respond to cyber attacks.
Ada enforces at least one authorized reviewer for all code changes, and deployments to our production environment are gated under condition that all code is reviewed. All code changes must go through a series of automated security scans before being deployed to production.
Ada maintains internal copies of security documentation, which are updated on an ongoing basis and reviewed annually for gaps:
- Information Security Policy
- Data Policy
- Risk Management Framework
- Incident Response Plan
- Security Vulnerability Identification
Ada conducts a mandatory background check and reference check for all employees prior to joining our team.
Ada enforces a mandatory security training program for all new and existing Ada developers that must be completed annually. This security training covers the OWASP Top 10 in specific programming languages that the developer uses.
In the event of a data breach, Ada defers to GDPR regulations, which maintains that customers shall be notified within 72 hours of a data breach, where feasible.
Ada maintains a live report of operational uptime and issues on our status page. Anyone can subscribe to updates via email from the status page.