How DialpadGPT keeps your data secure: A quick overview

The release of DialpadGPT adds generative AI capabilities alongside existing Dialpad Ai features like industry-leading voice transcription, Ai CSAT, and agent assists. By building and running its own custom language model, Dialpad can provide powerful GenAi features—while keeping all user data within our secure environment. This gives our customers new insights and tools along with the same security and privacy they have come to expect from Dialpad.

From Ai Scorecards that help speed up the QA process for supervisors…

To Ai Recaps that flag action items and summarize customer calls in natural language…

The Dialpad Ai features that we’ve already seen (with many more coming in the next few months) are just a small preview into the true potential of generative AI.

But what about the many legitimate concerns that businesses have about using generative AI tools?

No third-party AI vendors

Because DialpadGPT is run in-house, we can offer customers the ability to avoid the risks third-party associated with plug-in Ai vendors. This allows us to assert all of Dialpad’s existing assurances like SOC 2, ISO 27001, 270017, and 270018, and Business Associate Agreements (BAAs) for our GenAi products. This means our GenAi capabilities are available even to customers with heightened privacy and compliance requirements, such as those in the healthcare industry.

You’re in control of your data

Ai can power great business conversations, but when does it help most? By default, Dialpad Ai tools are turned off for newly created individual User accounts, and are turned on for group accounts—meaning those associated with Departments, Contact Centers, and Coaching Teams—to give them a head start on gathering conversational data for training and quality assurance.

We offer customizable automated recording and transcript notifications, which are also configurable at the business unit level.

The Ai Training section of Dialpad’s Privacy Policy explains in straightforward language how customer data processed through Dialpad’s Ai products is used to improve the services—and how any customer, from a large company to an individual user, can opt out of this usage.

GenAi training data? Anonymized.

As an early leader in AI-powered transcription, Dialpad has been used by businesses to generate transcripts for work calls and customer conversations for years. This process has helped refine Dialpad’s Automated Speech Recognition (ASR) and natural language process (NLP) models, providing a superb foundation on which Dialpad can offer its unique GenAi features.

To create DialpadGPT, Dialpad has anonymized those transcripts and removed personal information, which results in a custom AI model that prioritizes data privacy but is also optimized for the structure and vocabulary of business conversations.

Guided by Principles

Dialpad’s Ai Ethics Principles are the guidelines that we use to ensure we’re building Dialpad Ai products that are part of a world we want to live in.

Fairness and inclusiveness

No matter how accurate an AI model is, it’s based on data collected from the real world, by humans. This means that sometimes unconscious biases exist in AI models themselves, and also in the data that is input into the system, which may be reflected in the output.

The principles of fairness and inclusiveness mean that AI should work consistently and fairly across languages and cultures. At Dialpad, we do our best to design models with this tension in mind — with a goal of providing a service that works equitably and inclusively for every user.

Some of the ways we do this:

• Taking cultural differences into account when evaluating call sentiment. We understand that different cultures react and express themselves differently in speech, from apologies to swearing.

• Continuously expanding the set of dialects and languages that our speech recognition supports.

• Recognizing that many factors of identity are relevant to communications, such as gender, sex, dialect, ability, culture, and language.

Security and safety

Because our AI features and models are only as good as the real world data they are based on, the security and safety of customer data is fundamental to ethical data use. Dialpad’s information security program is therefore a critical part of its ethical AI principles, to ensure that both inputs and outputs are secure.

• Securely encrypting data in transit and at rest.

• Integrating with access management providers like Okta, Azure, Google Workspace, OneLogin to ensure your users are set up quickly and securely

• Undergoing annual 3rd party audits of our security and privacy program, including SOC2 and ISO 27001, 27017, and 27018.

Privacy and control

Whereas Security & Safety involve the steps Dialpad takes to ensure that your data is protected and available, Privacy involves the steps we undertake to respect the basic premise that your data remains yours, such as:

• Incorporating Privacy-by-Design principles into our development process, from data minimization to access limitation and customizable retention policies as a basic feature available to all users.

• Supporting data versioning tools ensure that your data is retained with Dialpad only as long as you want. Regardless of whether you share your data for Ai training, when you opt out of data sharing or delete your data via a retention policy, it is automatically deleted from all datasets.

Dialpad specifically designs its models to prevent inadvertent disclosure of sensitive information.