Metaforms

What is Metaforms AI?

Metaforms AI is an AI delivery platform provided by WorkHack India Private Limited for market research firms, sample suppliers, and survey programming teams. It enables agents to participate in questionnaire programming, Tracker revisions, data processing, and quality checks, while still allowing for manual reviews, version tracking, and team-specific rules.

Metaforms is neither a general-purpose search tool nor merely a form generator for respondents. Its main purpose is to standardize the repetitive tasks involved in research projects, from entering questionnaire data, managing survey codes, conducting verifications to delivering the results; it is suitable for research teams that already have professional processes and hosting platforms in place.

Main functions

• AI-driven survey programming: Upload the questionnaire and related files, select a template and level of complexity, and the AI will generate or modify the survey XML in batches. Programmers can examine the differences section by section, without having to accept the entire questionnaire code at once.
• XML Editor: On the left side, you can edit the survey structure; on the right side, you can view the original questionnaire. You can use Generate with AI or Edit with AI to work on specific questions. Each time you save changes, a history is created, allowing you to view or restore previous versions.
• AI Prompt Assistant: It suggests the next prompts based on the uploaded questionnaires and the generated XML files; users can edit these prompts or accept them. The optional Auto-Accept function allows for the continuous generation of sections, but for complex scenarios it is better to disable auto-acceptance and review each item individually.
• Think Hard Mode: When faced with nested loops, MaxDiff, Conjoint, cross-question validation, JS/CSS, or hidden logic, this mode prompts the agent to plan first before generating the output. It may improve the integrity of the structure, but it cannot replace test data or platform previews.
• Specializations: Depending on the project, various functionalities such as regular question types, verification, loops, quotas, sample sources, images, data searching, and scripts can be enabled. The longer the questionnaire is, the fewer specializations that can be selected at the same time; therefore, it is necessary to create and edit the questionnaire in stages.
• AI Rules: The results are generated based on instructions from teams or clients, code examples, explanations, and XML schema constraints. These rules can be set to apply to all projects as a requirement, or they can be selected on a case-by-case basis; in addition, permissions, usage insights, and version history are retained.
• Templates and Shell: Basic XML files, Quota Sheets, dynamic placeholders, and AI Rule folders are combined to create reusable templates. New projects start from these standard templates, which helps to reduce differences in the output produced by different programmers.
• Tracker version handling: It compares the old and new versions of supported tracking studies, identifies the affected issues and logic, and safeguards question numbers, labels, stored values, data layout, quotas, and history. Customers do not need to redo the entire project after making revisions.
• Data validation: Checks are generated using Python or SPSS based on the variables in the questionnaire; after execution, the records that failed are listed. Users can view the code, add checks manually, use AI to fix syntax issues, and review for logical errors.
• SPSS Macros: Allow for the reuse of common validation functions within data processing templates, and enable the creation of consistent scripts by combining them with AI Rules. Macros are currently part of SPSS’s validation framework; it cannot be assumed that Python supports them as well.
• Platform integration: API keys can be used to create or retrieve Forsta Decipher, IntelliSurvey, and Confirmit projects; custom integrations can also be set up for internal survey platforms. The scope of connections and the permissions related to credentials should be controlled by administrators.
• User roles and groups: Owners, Admins, and Editors have different permissions regarding user management, projects, templates, and data retention; they can be authorized to investigate projects on an individual basis or as part of a group.

Input, processing, and output

Investigate programming workflows

1. Configure the API key for the survey hosting platform in the administrator settings, and test the connection using the minimum required permissions.
2. Create an Ad-hoc Study or Tracker batch, and select a new project or import an existing platform project.
3. Select a template suitable for the customer and the platform, then enter the project name, directory, and storage path.
4. Upload the original questionnaire, quotas, materials, and other supporting files, and check the version and naming.
5. Select Basic or Advanced depending on the complexity of the questionnaire, and enable the required Specializations and AI Rule folders.
6. In the XML editor, generate questions in batches, and compare the questionnaires, XML files, and their differences one by one.
7. For simple repetitive blocks, Prompt Assistant can be used; for complex branches, loops, and scripts, Think Hard should be employed along with manual review.
8. Run it in the preview environment to ensure normal operation, and check the boundary and conflict data, as well as the transition rules, quotas, storage values, sample sources, and termination conditions.
9. After approval, it is synchronized to the hosting platform, with records of versions, modifications, testing, and delivery retained.

Data processing workflow

1. Create a Data Processing project, and select the Python or SPSS validation tool along with the corresponding template.
2. Upload the questionnaires and variable data, and use AI Rules and macros to define the team’s validation criteria.
3. Generate checks based on the question, read the human-readable instructions, and view the actual code.
4. Upload SAV data or prepare the execution environment to run individual question tests as well as full-project validations.
5. View failure records to distinguish between actual data issues, questionnaire exceptions, mapping errors, and script errors.
6. Use Edit with AI or make manual modifications for verification, run it again, and record the reasons for the corrections.
7. The final script, failure reports, and review records are exported and delivered for signature by the person in charge of data processing.

Suitable for users and scenarios

• Market research firms: Improve capabilities in programming multi-client surveys and data processing, while maintaining standards for each client.
• Sample and panel suppliers: Integrate the research questionnaires, quotas, and sample sources into the existing survey management processes.
• Investigate programmers: reduce the need to write XML repeatedly, and focus time on complex logic, platform constraints, and testing.
• Data processing team: Generates Python or SPSS scripts to verify the data and to view failed records in a centralized manner.
• Tracker team: Handles batch revisions and safeguards question numbers, tags, storage values, history, and data layout.
• It is not suitable for individuals who only want to create simple questionnaires, nor can it replace research design, sampling, statistical analysis, or the person in charge of ensuring the quality of human elements involved.

Platforms and modules

Prices and billing

Metaforms adopts a pricing model based on projects and modules, rather than on seats or tokens; members can be added to a project as needed, and proxies can be used according to work requirements. The public page does not provide specific annual fees, prices per project, a list of modules, costs for exceeding the designated number of projects, or trial prices.

The rules regarding refunds, early termination, unused items, project rollover, and price increases upon renewal have not yet been made public. When purchasing services, companies should include in the order details such as the modules, platform, definitions of the projects, delivery support, training, data migration, acceptance criteria, and terms for exiting the service.

API, SDK, and open-source status

• Metaforms uses the API keys from the survey platforms provided by clients to connect with Decipher, IntelliSurvey, and Confirmit; it can also create integrations for internal platforms.
• This connectivity capability is not equivalent to a public developer API; there is no confirmation regarding unified endpoints, SDKs, Webhooks, rate limits, or separate API pricing.
• No official GitHub repository for the complete product or open-source license has been found; the programming platform and AI agents should be considered proprietary enterprise software.
• It depends on the customer contract, ownership of the templates, and the rules of third-party platforms to determine whether XML, Python, and SPSS codes created or exported by users can be reused in other projects.

Privacy, security, and research compliance

• Metaforms handles account names and email addresses; it may also collect identifiable information, questionnaires, responses, sample data, and health information from customers for research purposes.
• The privacy policy outlines the commitments related to HIPAA, GDPR, and India’s IT Act, and it provides options for accessing, correcting, deleting, restricting, objecting to, and porting data. When dealing with protected health information, it is still necessary to sign the applicable agreements and verify the specific controls in place.
• The product page states that it meets the ISO/IEC 27001 and SOC 2 Type II standards, has undergone annual independent penetration testing, and is compliant with HIPAA and GDPR. Purchasers should request the current certificates, reports on the scope of testing, bridging letters, and a list of subcontractors.
• Logs may include IP addresses, browsers, service providers, timestamps, pages visited, and clicks; websites also use cookies. The privacy policy mentions third-party advertising technologies as well, and companies should verify whether such technologies are also used in their own environment.
• User roles should be configured with a separation of responsibilities, so that only the Owner has access to data retention controls. Before sharing projects, templates, and rules, it is necessary to check the customer isolation and editing permissions.
• The Terms of Service page still reflects the definitions set forth on the WorkHack website in 2023, which is significantly out of date compared to the current Metaforms product, as well as its billing and module offerings. Final approvals, responsibilities, refunds, and data processing are governed by the signed contract and its most recent appendices.
• The analysis of data from respondents must comply with the principles of informed consent, purpose limitation, minimization, retention, deletion, and cross-border data transfer; AI-generated code cannot resolve compliance issues on its own.

Advantages and usage limitations

• Advantages: The agent is designed around a professional investigation delivery process, allowing for the reuse of XML, platform rules, customer templates, and validation code.
• Advantages: Batch generation, differential review, version history, and anomaly detection facilitate maintaining manual control.
• Advantages: Pricing is based on projects rather than seats or tokens, making it suitable for collaborative work and frequent use by agents.
• Limitation: The capacity for Specialization varies depending on the length of the questionnaire; complex questionnaires may need to be created in stages.
• Limitation: Auto-Accept speeds up the occurrence of duplicate blocks, but it may also cause errors to spread to subsequent issues.
• Limitations: Python and SPSS scripts may contain syntactic or logical errors; they must be run using actual test data and signed manually.
• Limitations: The standard price, the unified API, refunds, the standard SLA, specific data retention policies, and platform coverage are not disclosed.

Summary

Metaforms AI is suitable for professional research teams that need to carry out large-scale survey programming, make revisions to trackers, and verify data. When evaluating it, it is necessary to use actual questionnaires and historical data to assess the accuracy of XML processing, the ability to detect anomalies, the efficiency of platform synchronization, and the costs associated with manual review; in addition, project specifications, security certifications, and contractual rights must also be verified.