Survey analysis for consultancies

How D.A.V.E. works



From consultation responses to a client briefing
More than 4,000 people responded to Highland Council’s visitor levy consultation. GoLLM worked with Diffley Partnership to analyse over 100,000 answers and bring the findings into the report that informed councillors’ deliberations.
Read the Highland Council case studyAs featured in:


From raw feedback to evidence you can act on
Analysis designed for real decisions
Blended intelligence
Quantitative + qualitative analysis
Advanced data science
Evaluation and decision support
Multi-wave trend insights

Works with your existing data
Survey platforms
Spreadsheets
Written sources
Audio sources
Use cases
See D.A.V.E. in action
FAQs
Transparency is important to how we approach AI-assisted analysis. D.A.V.E. provides visibility into qualitative categorisations and supporting source responses so users can review how broader themes and findings relate back to the underlying evidence.
We are continuing to strengthen the platform's auditability and provenance features as part of the product roadmap.
Yes. Our service is designed to be a partnership, not just a platform.
Our in-house research insights team is available to work alongside D.A.V.E., providing a critical layer of expert human review and strategic oversight. This ensures that every finding is not only data-backed but also validated and enriched with years of professional experience.
User privacy is paramount. D.A.V.E. is designed with a selective data handling feature, which allows you to process only the specific, isolated questions you need insights from. This means you can easily exclude any demographic or identifying fields from the data you upload, which significantly mitigates re-identification risks and ensures you can get valuable insights from your responses while keeping sensitive data private.
No. We have a strict policy that no partner or client data is ever used for AI model training. When D.A.V.E. utilises third-party APIs (such as OpenAI or Anthropic), it is done under enterprise-grade terms that explicitly prohibit any data retention or training on the data you provide. Your data remains yours, always.
We offer multiple layers of data protection. Through masking and preprocessing, you have the option to redact or anonymise any sensitive data locally before it ever leaves your environment. Even within open-ended, free-text fields, our platform can apply automated filters to remove personally identifiable information like names, locations, or other sensitive details, ensuring your data is protected from start to finish.
D.A.V.E. applies consistent analytical methods across the data, helping reduce some of the inconsistency that can arise in purely manual analysis. Outputs can be reviewed against the underlying responses, and expert human review can be added where needed.
No analytical method is completely free from assumptions or bias, so we focus on making the evidence and methodology as clear as possible rather than claiming that AI removes bias altogether.
D.A.V.E. is designed to interpret sentiment, themes and context across qualitative feedback, but nuanced language can still benefit from human judgement. GoLLM's analyst team is available to provide additional review, interpretation and validation where that level of support is useful.
D.A.V.E. is designed to work with a wide range of unstructured and structured data sources. This includes open-ended survey responses, call transcripts, customer service chat logs, employee reviews, social media comments, and more. Our platform is flexible and can be customised to analyse data from almost any text or audio-based source.
D.A.V.E. turns analysis into clear findings, executive summaries, implications and recommended actions so teams can focus on interpretation and decision-making rather than manually working through large volumes of data.







