Summary

• Strawberry 1.0 adds redesigned AI companions, nine predefined specialist configurations and the Companion Plaza interface.
• Browser automation, application integrations, scheduled routines, meeting transcription and team collaboration were introduced during earlier releases.
• Enterprise deployment requires scrutiny of permissions, data handling, workflow reliability and usage-based costs.

Stockholm-based Strawberry has released version 1.0 of its AI-enabled browser, introducing redesigned assistants and predefined specialist configurations as it seeks to establish browser-based automation as an alternative to navigating individual workplace applications manually.

The release, dated 26 October, allows users to customise the appearance of their AI companions and select from nine specialist starting configurations, including assistants intended for sales and recruitment activities. Each comes with an initial set of skills and routines that users can adapt to their working practices.

Strawberry has also introduced Companion Plaza, an animated interface where assistants appear as characters within the browser. It is an unusual design choice for workplace software, although the feature concerns how users interact with assistants rather than a new ability to complete business processes.

The underlying automation platform has developed throughout the preceding year. Application integrations, scheduled routines, meeting recording, document creation and shared team assistants were introduced before version 1.0, meaning the latest release represents a product milestone rather than the first availability of those capabilities.

Browser access changes the automation model

Employees often perform routine work by moving information between websites and applications that were designed for separate purposes. A sales team might collect details from a prospecting service, enter information into a customer relationship management system and prepare a message in an email application.

Conventional automation can connect these systems through dedicated integrations, but configuring those workflows may require technical expertise, administrative permissions or specialist software.

Strawberry’s approach places an AI assistant inside the environment where much of that work already takes place. Its companions can use connected applications, interact with web pages and carry out multistep tasks that would otherwise involve repeated manual actions.

The browser added support for more than 100 applications during May 2026, including services used in finance, recruitment, marketing and customer management. The same update introduced wider compatibility with Chrome extensions and improved the ability of assistants to break larger tasks into smaller activities.

Earlier releases had already introduced routines that run at scheduled times or respond to events in connected applications. Those features allow selected work to begin without a user issuing a separate instruction every time.

For example, an assistant might prepare a morning briefing from messages and records, or draft a response when a particular type of communication arrives. The usefulness of the result depends on whether the assistant has access to the appropriate information and understands the limits of the task.

Automating a process through a browser also introduces vulnerabilities associated with changing interfaces. A form may be redesigned, a field renamed or a website’s authentication procedure altered without the automation provider controlling those changes.

Where a task depends on recognising elements within a page, such changes can cause an assistant to fail or take an unintended action. Reliability must therefore be measured across repeated executions, not inferred from a successful demonstration.

Collaboration creates additional permission requirements

Strawberry introduced team features during 2025, allowing employees to share configured assistants and workflows with colleagues. Its later releases extended shared files and working environments, making reusable automation part of its wider proposition.

That collaboration can help organisations avoid recreating the same administrative process for individual employees. However, shared instructions do not necessarily imply that every person using an assistant has authority to access the same information or perform identical actions.

An assistant used for recruitment may encounter candidate records, internal assessments and contact details, while one operating within a sales workflow could process commercially sensitive information or customer correspondence.

The organisation must establish which applications an assistant can use, what data it may retrieve and whether particular actions require approval. Permissions should reflect the person using the tool and the business process involved, rather than the technical capabilities available to the assistant.

Strawberry describes controls allowing connected actions to be configured according to whether approval is required or activity is permitted. Nevertheless, the effectiveness of those controls depends on how the organisation configures them and whether they are consistently applied across integrations.

Tasks involving outgoing messages, payments or modifications to authoritative records require particular care because errors can affect people and systems outside the browser itself.

External web content can create another risk. An assistant that reads instructions or documents from a third-party site must distinguish information relevant to the task from material attempting to influence its behaviour.

Conventional browser security remains necessary, but AI introduces a further requirement to control how information retrieved from websites is interpreted and acted upon.

Data storage claims need careful examination

Strawberry describes protections for passwords and browser information, while its more recent product history confirms that encrypted conversations can be synchronised through a Strawberry account and accessed from another computer.

That distinction complicates broad statements that all user information remains stored locally. Different data categories may be handled differently, particularly where synchronisation, connected applications and external AI services are involved.

Organisations evaluating the product need precise information about which records remain on the device, which are transmitted to company infrastructure and which are processed by external model providers.

They must also consider retention policies, administrative access, recovery arrangements and the contractual treatment of information supplied through business applications.

The same concerns apply to meeting transcription. Strawberry supports recording and summarisation of calls conducted through widely used meeting applications, providing a record of discussions that can be connected to subsequent work.

However, recording meetings introduces consent, transparency and information-management obligations that depend on the participants, organisational policies and applicable law.

Usage economics extend beyond credits

Strawberry offers a credit-based charging system in which ordinary browsing is separated from activity performed by AI assistants. Plans provide different usage allowances, while team customers can arrange shared credits and administrative controls.

The cost of an automated workflow can vary according to how many steps an assistant performs, which models are used and whether a task needs to be repeated after an error.

This creates a different purchasing calculation from a conventional fixed-price productivity application. Organisations need to estimate not only expected AI consumption but also the time spent configuring workflows, supervising execution and correcting unreliable results.

A browser agent may complete an administrative task quickly while still requiring a person to check its output before it is used. Such review is appropriate for consequential activities, but it affects the extent to which automation produces genuine time savings.

Strawberry has presented examples of successful workplace use, although independently verified benchmarks covering completion rates, error frequency and costs across different organisations are not yet available.

The product also operates within a competitive market in which browsers, enterprise applications and AI providers are adding assistant capabilities to existing software. Organisations may already have access to some forms of automation through their established tools.

Strawberry’s proposition is that the browser can serve as a common workspace across applications rather than requiring each process to be automated separately. The advantage will depend on how reliably those applications can be controlled and whether organisations consider the additional software layer worth its cost.

Version 1.0 introduces a more distinctive interface and a simpler route to specialist assistants, building on capabilities released during the beta period. Its longer-term commercial prospects will rest on whether the underlying automation remains dependable when used repeatedly across changing business systems.