With Search Copilot from Read AI, millions of people can easily and instantly turn on enterprise-quality workplace search—for free. We’ve eliminated historic barriers to entry, including high fees and untold hours of set-up time. Read AI supports individuals and teams at no cost and we are delivering a product that can be implemented in 20 minutes or less.
The standout differences in our product versus others does not stop with cost and implementation. Read AI is committed to new and different levels of permissioning and controls, rooted in our founding principles of transparency-first and privacy that stem from decades of experience in the digital analytics and advertising world. As an organization, we believe that individuals should have the power to decide who can access their data. By giving individuals’ freedom, flexibility, and control over data sharing, it better protects them and the companies they work for.
There are two options when it comes to enterprise search. One is built on an antiquated permissions model that’s been around since floppy disks, which gives outsize control to leaders with only a birds’ eye view. AI requires a more modern approach—one that champions user-by-user controls. With Red, your information is yours.
Just like you know when to forward an entire email thread and when not to, and you decide whether to share an Instagram post with your close friends or entire network (or frankly, no one at all), that is the same principle at work with Read AI, and workplace search. It is never anyone else’s decision to share, reveal, or otherwise open up your data to others. That is entirely up to you.
It’s aligned to this mental model that we deliver Search Copilot, empowering millions of customers to apply the same easy sharing actions to the workplace, ensuring a superior user experience from Day 1.
Enterprise search delivers the type of productivity gains that organizations will need to be successful in the agentic AI era and beyond. It is an example of the type of AI application that truly solves real workplace problems—such as, ‘I can’t remember a detail,’ ‘I can’t find that piece of information,’ ‘I want an update on the latest,’ etc.—and yet, it needs to be adopted by teams for organizations to be able to best utilize it and see the results.
While our competitors enable an IT team to make decisions about what information gets filtered into the database, we know that this can create too much transparency and can give employees unintentional access to sensitive information. At Read AI, we implement the following:
We apply a user-by-user data permissioning model. When pulling from integrated services like email, the data is surfaced only from your knowledge base to start. This means that no one in your company can accidentally find an email from your inbox when performing their own search.
However, the most productive individuals and teams will see the immediate value that comes from shared information—but sharing happens piece-by-piece, item-by-item. There is no carte-blanche access. When an item is shared with you, only then will the information within the meeting, document, email, or message, be scannable and appear in your results. To say it differently, Search Copilot only answers questions with information derived from the raw content you have permission to see.
We don’t train against your data, hard stop. You can let us, but it requires that you opt in. Currently, only around 10 to 15% currently opt in, and that’s entirely fine with us.
Even for those who take part in that program, we don’t store any content—ever. In fact, there isn’t even an option to contribute your actual content to the system or the product. Opting into data sharing simply allows Read AI to test things like new features, gather performance data, and track how well the product is working overall. It helps us improve the service, but is not a requirement nor is it an expectation.
Our powerful integrations have limits by design. As an example, when you give Read AIaccess to your Google Drive, you will still be able to remove certain documents from Search, and you will only make documents available to others with explicit sharing of each one.
We built many features within Read AI alongside customers in the financial sector. For that reason, there are additional tools in place that are unique to our platform and serve highly regulated industries specifically. Yet other industries and individuals also benefit. As an example, we offer the ability for teams and individuals to eliminate the storage of high-fidelity recordings or full transcripts; when this is in place, these artifacts are used for meeting summaries and action items, then immediately deleted. They can never be scanned within Search Copilot. This feature is valuable for businesses with strict data retention policies that still want to maintain essential meeting details for future reference.
There’s a flip side to this control. While companies can set restrictions—such as eliminating transcription—they also cannot automatically store everything employees produce. Any employee can decide what is stored, and if they would like transcription turned off. Having these two levels of governance enables transparency and ensures that employees can effectively collaborate without compromising privacy or exposing confidential information.
Meetings operate with similar permissioning. When multiple people are in a meeting with Read AI, everyone who’s part of the meeting automatically gets access to the shared report, which is added to their database—even if they weren’t the one who invited Read AI to the call. But meeting owners are always in control and can add or revoke viewer privileges at any time. When they make the report private, it stays private, and when it is shared with specific people, even if they weren’t in the meeting or aren’t part of your company, it remains open.
Owners can also revoke access to reports at any time, whether it’s before or after the meeting. And if owners initially set the report to be public, they can still make it private later. These changes are applied and carried through to Search Copilot, and information is available or not, in accordance with the owners’ decisions. The Read AI internal authorization service runs half a billion checks daily, confirming that any time a user wants to access a piece of data, they actually have permission to do so.
In principle and in alignment with the way our latest expectations around AI search tools, Search Copilot always reveals the sources behind its results using citations. This transparency ensures accuracy, eliminates hallucinations, and makes it clear to the user what data is being used within search.
Because we believe in the importance of quickly sharing information, we made sure that Search Copilot responses could be shared without exposing sensitive materials. While you'll get extensive feedback on the sources that tell you how the AI came to its response, none of that goes along when it's shared. Instead, viewers can only see the citations.
We believe that Search Copilot can fundamentally change the current operating norms in place for hybrid and remote work. Starting now, remote employees can share documents, meeting reports, and other outputs with their colleagues. When that happens, their work—whether that’s progress on projects, customer calls, or any other artifact—becomes part of their teammates’ knowledge base, and it will be surfaced as often and with the same level of detail as the work of their in-person collaborators, immediately leveling the playing field.
With this subtle but powerful change, work can truly be reviewed and judged on merit, as information is no longer locked up in silo. Instead, their contributions are summoned by the Search Copilot tool, and elevated alongside its true merit.
When an employee prompts Search Copilot, they’ll be able to glean answers from meeting reports and documents they didn’t attend or participate in, assuming these documents have been explicitly shared with them. The draw of this is that employees don’t need to sit through every meeting or crawl through every document in the company knowledge base in order to learn what they need to know about the organization. That also means less meeting burnout and dramatic improvements in employee productivity.
Search Copilot empowers a privacy- and productivity-first approach to remote work, to hybrid work—to all work. It’s enterprise search for everybody.
Because we believe in the importance of quickly sharing information, we made sure that Search Copilot responses could be shared without exposing sensitive materials.
© 2024 Read AI, Inc.
%2C%20Color%3DOriginal.png)
