Credits are managed at the organisation or company level and can be used by authorised users according to their account access. This makes credits a shared organisational resource rather than a separate balance for every individual user.
Usage tracking helps organisations understand how credits are consumed across modules, projects, and users. This is useful when several reviewers, analysts, or project teams are working inside the same organisation account.
Yes. SYMPRO AI supports connected workflows so that review material can move between supported modules instead of being repeatedly re-entered. For example, AccuSearch can produce search results and citation records that are used in AccuScreener. Screening outcomes and uploaded full texts can then support downstream work such as risk-of-bias assessment, structured extraction, abbreviation review, and report drafting.
The important detail is that movement between modules is governed by workflow state and user permissions. A downstream tool should receive the right studies, documents, and context for its purpose.
Yes. Organisations can purchase additional AI credits.
If available credits run out, additional credits can be added. Existing project data and completed outputs remain available as well.
Yes. Supported workflows retain traceable records for project activity and user decisions. This helps teams understand what changed, who performed the action, when it happened, and which workflow state was used.
Examples of tracked activity
- Project setup changes and workflow configuration.
- Reviewer assignments and screening decisions.
- Conflict outcomes and exclusion reasons.
- AI runs, processing statuses, and generated outputs.
- Uploaded files, linked documents, exports, and update-cycle records where supported.
Audit value: traceability helps teams explain how a review moved from search to screening, appraisal, extraction, and reporting.
SYMPRO AI uses organisation-level separation and role-based access controls. Users should only access projects, modules, and administrative functions that belong to their organisation and match their assigned role.
What this means in practice
- Organisation separation: project data is scoped to the relevant company, university, or organisation account.
- Role-based controls: admin, editor, reviewer, analyst, and other permissions restrict what a user can view or change.
- Project assignment: reviewers and analysts generally work only on assigned projects or tasks.
- Administrative access: project setup, user management, publishing, and platform-level controls are limited to authorised users.
This matters because review projects can contain unpublished protocols, licensed documents, confidential client work, and internal decisions.
AI credits are consumed when a workflow runs an AI-assisted operation. The exact credit use depends on the module, action, file count, article count, field count, or processing scope. The product shows available credit information so users can monitor usage before and during a trial or organisation plan.
Examples of credit-consuming actions include AI search generation, AI screening or full-text processing, risk-of-bias processing, AccuSynthesis extraction, AQUA processing, AccuScripter structure generation, and AccuScripter writing generation.
AccuSearch credit use is typically tied to search or AI search-generation actions. AccuScreener credit use can be tied to article-level AI screening or file-level full-text processing.
RoB Master and AQUA credit use is typically file-based. AccuSynthesis credit use can depend on files, extraction fields, or the configured extraction scope. AccuScripter credit use can depend on file count and generated structure or writing operations.
Manual review, ordinary navigation, editing already generated text, and downloading existing exports generally do not consume credits unless they trigger a new AI run.
Yes. AccuSearch can use a research question and structured review inputs such as PICOS to generate a draft search strategy. It helps translate review intent into searchable concepts, synonyms, Boolean logic, and database-oriented query strings.
How users should work with the output
- Enter the review question and structured review details.
- Generate the draft strategy.
- Review concepts, synonyms, filters, and field logic.
- Test hit counts and inspect sample records.
- Edit or regenerate the strategy before final use.
Yes. AccuSearch can help generate keywords, synonyms, Boolean terms, and controlled vocabulary terms such as MeSH where appropriate. This helps convert a plain-language review question into a structured search strategy.
What users should review
- Whether generated synonyms are clinically relevant.
- Whether generated terms are appropriately scoped, current, and database-ready.
- Whether Boolean logic matches the intended concept structure.
- Whether field tags and controlled vocabulary terms are suitable for the database.
The best workflow is to use AccuSearch for drafting and organisation, then test, inspect, and refine the strategy before final use.
This helps make the search process more transparent. Instead of starting from an empty query box, users can see the term groups, understand how concepts are combined, and make deliberate changes before retrieving results.
Yes. Users can edit and regenerate AI-generated search strategies. This is important because search development is normally iterative: teams test early strings, inspect hit counts, adjust terms, refine filters, and improve database syntax.
When regeneration is useful
- PICOS fields have been revised.
- Outcome categories have been added or narrowed.
- Date, language, article-type, or study-design filters need to change.
- The first query is too broad, too narrow, or syntactically unsuitable.
Yes. AccuSearch can export generated or edited search strategies, search steps, hit counts, and retrieved article details. Exported citation information can include title, abstract, authors, journal, publication year, DOI or other identifiers, keywords, MeSH terms, and other available metadata.
Why exports are useful
- They document the search process for audit and reporting.
- They allow search results to move into screening workflows.
- They help teams preserve hit counts and retrieved records from each run.
- They support comparison between original and update searches in living-review workflows.
AccuSearch supports PubMed-oriented search and retrieval workflows. It can generate, run, and retrieve PubMed search results and article details where the query and retrieval limits allow it.
The platform can also help prepare database-oriented search strings for other evidence sources such as Embase, Cochrane. Integration of Clinical Trials and OpenAlex is under development as well.
AccuSearch works best when users provide a clear review title or question, population, intervention or exposure, comparator where relevant, outcomes, and any important study-design or publication filters.
Useful details to provide
- Review question: the exact question the search should support.
- Population: condition, disease area, patient group, age, setting, or geography.
- Intervention or exposure: drug, device, procedure, programme, diagnostic test, risk factor, or policy.
- Comparator: placebo, standard care, alternative treatment, no treatment, or not applicable.
- Outcomes: efficacy, safety, HRQoL, economic, epidemiology, diagnostic, or custom outcomes.
- Limits: dates, article types, study designs, languages, humans or animals, and text availability.
Yes. Mini-DET supports focused summary extraction for configured fields in screening workflows. It can help extract concise information from records or documents so reviewers can see key details while making screening decisions.
Mini-DET should be configured around the review's needs. For example, a team might ask for population, intervention, comparator, outcome, design, or other short fields that help screening and triage. The value comes from making extraction focused rather than asking for broad unstructured summaries.
Mini-DET outputs are designed to be checked alongside the source material so reviewers can use them confidently during screening and reporting workflows.
Yes, where AI Screener mode is configured. In this mode, AI can act as an independent reviewer stream alongside a human reviewer. The AI output can then be compared with human screening decisions in a way that resembles double screening.
The team configures the AI run, provides the criteria, reviews disagreements, and decides how conflicts are resolved. AI decisions become part of the reviewable workflow record, with final inclusion and exclusion decisions kept under human governance.
AI as a second reviewer can be useful for scale and consistency, but teams should pilot the setup, confirm the protocol fit, and document how AI-generated decisions were used.
Yes. Screening workflows can be configured around the criteria needed at each stage. Title and abstract screening may use broader inclusion and exclusion logic because only citation information is available. Full-text screening can use more detailed criteria because the complete document can be reviewed.
For example, a title and abstract reviewer may decide whether a record appears to involve the right population and intervention. A full-text reviewer may then check study design, outcome reporting, follow-up duration, duplicate publication status, or whether the full text contains usable data.
Yes. References can be added after screening begins, depending on the upload route, linked workflow, and project setup. This is useful when a team receives late search results, hand-searching records, citation-chasing records, or update-cycle records during a living review.
Newly added references can enter the relevant screening stage with their source documented. Teams can identify whether they belong to the initial search, a supplementary source, or a later update cycle.
For living or update workflows, cycle tracking is especially useful because it helps users see which records were introduced after the original search.
Yes. Reviewers can record reasons for exclusion, especially during full-text review where PRISMA-related reporting often requires clear exclusion categories. These reasons can be reviewed, resolved, and exported for use in reporting or audit files.
Good exclusion-reason practice matters. Teams should define exclusion categories before full-text review, avoid overlapping categories where possible, and ensure reviewers use categories consistently. If categories change later, the team should decide whether existing decisions need cleanup.
Exports can support internal QA, sponsor review, manuscript preparation, or PRISMA flow documentation, but final reporting should still be checked by the review team.
Some protocol or screening settings can be updated after screening begins, but this should be treated as a controlled workflow change rather than a casual edit. Changes to eligibility criteria, decision categories, reviewer assignment, AI settings, or full-text configuration can affect how old and new records are interpreted.
If a change is made after screening has started, earlier decisions may not automatically represent the new criteria. Depending on the project, users may need to rerun AI support, reassign records, re-screen affected studies, or document why previous decisions remain valid.
Yes. AccuSynthesis can extract relevant data from tables, figures, and supplementary files, provided the content is readable and accessible within the project.
This includes information from sources such as results tables, subgroup analyses, safety tables, appendices, and supplementary materials. When the required data fields are defined in the extraction template, AccuSynthesis can capture the relevant information and organize it into a structured format for review.
Extracted data can also be reviewed and exported.
Yes. AccuSynthesis supports template creation, customisation, and reuse. Users can build templates manually, import structured templates, convert suitable workbooks into template drafts, and use AI assistance to generate template.
Reusable templates are valuable because evidence extraction often repeats similar structures across reviews. A team may use recurring fields for study characteristics, intervention arms, outcomes, timepoints, subgroups, safety events, economic variables, or diagnostic accuracy measures.
Templates should still be reviewed before extraction begins. The review team should confirm that each field has the right label, description, value type, allowed options, grouping, and row identity so that exported data matches the intended analysis.
Yes. AccuSynthesis is built around researcher review. Users can inspect extracted values, review the source context, identify not-reported or uncertain values, revise extracted data, and decide what is ready to export.
This review step is central to the product. AI can help populate a structured template while the research team confirms values, study arms, subgroups, and not-reported values in a consistent evidence workflow.
This combination of AI-assisted extraction and researcher review helps teams maintain consistency across complex evidence while keeping researchers in control of the data used for analysis and reporting.
Yes. AccuSynthesis can represent studies with multiple treatment groups, subgroups, timepoints, outcome domains, and repeated results. It does this through template structure, row identity, dimensions, and measures.
For example, a template can distinguish treatment arm, comparator, subgroup, visit, outcome, statistic, and value. This helps prevent values from being flattened into a confusing single row or mixed across arms.
The team should design templates carefully for multi-arm studies. Clear row identity and field descriptions reduce ambiguity and make the final export more useful for evidence tables, quantitative synthesis, or narrative reporting.
AccuSynthesis extracts the information defined in the project's extraction template. This can include study characteristics, population details, interventions, comparators, treatment groups, outcomes, timepoints, safety results, economic results, diagnostic results, epidemiology values, qualitative findings, and custom fields. You can also upload your own template.
The template is the key control layer. It defines sheets, sections, groups, fields, dimensions, measures, value types, allowed options, and instructions. That means extraction is not a generic summary of the article. It is a structured workflow guided by the review team's intended evidence table.
AccuSynthesis is designed to make extraction more structured and reviewable, not to remove the need for scientific verification.
AccuSynthesis supports common research document formats, including PDF, DOC, DOCX, and TXT.
These formats can be uploaded directly to the platform and used for structured data extraction. AccuSynthesis can process relevant information from the document content, including text, tables, and figures, and organize the extracted data according to the project’s extraction template.
AQUA extracts abbreviation-related information from uploaded documents. It separates abbreviation findings by document area, including text, tables, and figure-related content. It can also surface relevant candidates that may need user confirmation.
This is useful when a research document contains many short forms, acronyms, technical terms, or table-specific abbreviations. Instead of asking users to manually scan the entire document, AQUA presents a structured review surface so users can inspect detected abbreviations and possible candidates per file.
Text abbreviations are abbreviations and definitions detected in the main body text. Table abbreviations are abbreviations and definitions detected from table content.
Figure abbreviations are abbreviations and definitions associated with figures or figure-related content. Relevant candidates are possible abbreviations, short forms, or terms that may need confirmation by the user.
AQUA supports standalone projects and workflow-linked document use. In the current upload workflow, users can upload PDF, DOC, and DOCX files, process multiple documents, review each document separately, and export results.
AQUA supports PDF, DOC, and DOCX input files. It can be used through standalone AQUA projects or with workflow-linked documents.
Multiple documents are supported, with per-document processing and review. The current upload workflow supports files up to 50 MB each. AQUA results can be exported as DOCX and Excel files.
Readable document content helps AQUA present cleaner abbreviation results.
AccuScope supports scoping review workflows by helping organise and prioritise articles against the review's objectives. It can rank or score records using configured criteria and available article information so users can focus attention on the most relevant material first.
This is useful in scoping reviews because the goal is often to map a broad evidence area rather than answer a narrow intervention-effect question. Ranking can help users understand clusters of relevant literature, identify likely priority records, and manage large result sets.
Yes. AccuScripter can generate evidence tables and support PRISMA-related reportcontent using information available within the project.
It can help create structured evidence tables, study summaries, report sections, and content describing the review methodology and study selection process. AccuScripter can also use screening, extraction, and other project data to develop relevant methods and results content.
By bringing project evidence into a structured reporting workflow, AccuScripter helps researchers prepare clear, consistent, and evidence-linked content for systematic reviews and other research reports.
Yes. AccuScripter is designed around staged review rather than one-click final writing. Users can review proposed report structures, inspect how evidence is linked to sections, confirm table placement, review generated writing, and edit content before it is approved.
This staged process matters because evidence reports often require project-specific style, scope, interpretation, and compliance with protocol or reporting standards. Users may need to adjust section headings, move evidence between sections, rewrite interpretations, or correct table content.
The final report should represent the review team's approved interpretation. AccuScripter can speed up organisation, drafting and report creation.
Living SLR is designed for repeated search cycles and ongoing monitoring. Scheduling support depends on the configured workflow and account setup. Where enabled, scheduled runs can help retrieve updated search results and support controlled sync into screening.
A responsible living-review workflow should define the database source, search strategy, refresh frequency, sync behaviour, reviewer responsibilities, and how new records are screened. Users should also know whether updates are manual, scheduled, or partially automated.
Yes, the Living SLR workflow is intended to preserve search-run and sync-run provenance. Where cycle tracking is enabled, users can see which search run, refresh cycle, or sync introduced a study into the review workflow.
This helps teams distinguish original records from later update records. It also supports reporting questions such as which studies came from the baseline search, which came from a scheduled refresh, which were already present, and which were newly synced into screening.
Cycle visibility is important for auditability because living reviews evolve over time. Review teams should be able to explain when evidence entered the project and how it was handled after entry.
Yes. RoB Master supports per-file assessment selections. This matters because one evidence review can include different study designs that need different appraisal methods.
For example, a project may include randomised trials, observational studies, economic evaluations, or systematic reviews. Those evidence types can be assessed with the appropriate tool selected by the user. RoB Master can group files by the user's selected type or scale for processing.
Yes. RoB Master can generate downloadable assessment outputs, including workbook and document files. For supported assessment outputs, the result package can also include judgement summaries and generated visualisations such as summary plots or traffic-light plots.
Outputs are packaged for download so teams can review them outside the platform, share them for QA, or use them in reporting workflows. The exact contents can depend on the selected assessment tool and processing result.
RoB Master reviews uploaded study files and recommends an assessment type or scale for each file. The recommendation is intended to help users identify which risk-of-bias or critical-appraisal framework is likely to fit the study design and evidence type.
Users can review the recommendation, change the selected type or scale, and save their selections before processing. This is important because the AI may not know project-specific protocol requirements, sponsor preferences, or methodological decisions made outside the document itself.
The recommended workflow is: upload files, review AI recommendations, correct type or scale selections where needed, save selections, then process the assessment outputs.
RoB Master supports multi-file upload and background processing. Files should contain readable study content because recommendations and assessment outputs depend on the text that can be extracted from the document.
RoB Master supports PDF, TXT, DOC, and DOCX files. A RoB Master project can include up to 30 uploaded files, with a current limit of up to 20 MB per file.
After upload, the AI recommendation can be reviewed and changed for each file before processing. Processing runs in the background, with status updates and result downloads available after completion.
For best results, users should choose supported file types, keep files within the size limit, and upload readable study documents that contain the information needed for appraisal.
RoB Master supports structured risk-of-bias and critical-appraisal workflows with AI-assisted tool recommendations. Configured assessment options include tools such as ROBINS-I, Cochrane-style risk-of-bias options, NICE, Newcastle-Ottawa, JBI, Drummond's scale, Downs and Black scale, and AMSTAR 2.0 where available in the deployment.
The exact framework list may depend on account configuration and implementation status. Teams working under a protocol, regulatory submission, HTA process, or sponsor-specific method should confirm the required framework before processing files.