1
Connect your data.
An SPSS, CSV, Excel, Word or PDF file, or your data lake.
Proactive insight discovery · IIEX West, San Francisco
Find what will move the needle.
Surface the insights that could have the biggest growth impact, and bring them to the business while there's still time to act.
Sophisticated data science. Made remarkably simple.
- What drives growthDiscover what influences your most important metrics.
- Who to winUncover meaningful segments and what makes them different.
Bring a past study.See what might have been missed.
IIEX WESTSan Francisco · 2026
Your everyday workflow
Explore the way you always have. Without the setup.
Prefer to work through the study yourself? Everything is ready the moment your data lands.
2
Our statistics engine does the groundwork.
In seconds, every question is summarised into charts and tables, and every pair of questions is tested for correlation.
3
Ask, and see the answer.
The AI works from results already calculated and drives the interface for you, building the chart and applying filters, so your answer comes in words and as a visual you can use.
An evidence engine for modern insight teams
The evidence leads. AI connects the dots. You decide what matters.
One statistical engine investigates the whole study. The method adapts to each question and data type, so you get to the findings without setting up tests one by one.
Drivers are associations to investigate, not proof of cause.
Differences
Know which gaps are real
Every audience, market and wave gap tested. Spot differences worth paying attention to.
Drivers
Know what to build on and what to fix first
See which factors are most strongly linked to the result you want to improve, across several models.
Segments
Segments you can take to the business
Our AI agent tries different ways of grouping people and brings you groups that are distinct and easy to describe.
Patterns
Find patterns you didn't think to look for
Links across ratings, categories and multi-select questions you didn't plan to test.
Themes
Hear the why
Open ends coded into themes and ranked alongside the numbers.
Opinions, preferences, attitudes, segments and themes, pulled into one ranked report. Ranked Impact Analysis. Multi-Type Adaptive Discovery: established statistical and machine-learning methods, selected and combined to fit the question and the data.
Confidence in an AI world
You shouldn't have to choose between embracing AI and trusting your work.
Insight teams don't have to work backwards from an AI answer to spot-check the evidence. AI can help you move faster. But checking generated code, tracing calculations and keeping track of every analytical decision can slow you down again.
AddMaple gives your team a consistent statistical engine, visible filters and inspectable results. Explore further, compare audiences and follow the evidence without rebuilding and rechecking the analysis at every step.
Move at AI speed. Keep your confidence in the numbers.
The engine does the calculating. AI explains the results; it doesn't compute them.
Bases, tests and model diagnostics are there to inspect, so anyone can see what a finding rests on.
Numerical results are reproducible for the same data, methods and settings. The AI's wording may vary.
One statistical engine. All your AI tools.
Explore in AddMaple, or from Claude, ChatGPT or Cursor. One engine runs every calculation, so results are consistent, reproducible and shown in full. No AI-written code to audit.
Quant and qual, together
The numbers show what. Open ends show why.
Open ends, interview transcripts and videos are coded into themes and ranked with the survey data, with every quote and clip linked back to the evidence.
A team capability
More people exploring. One shared standard of evidence.
Make discovery a team capability, not a specialist bottleneck. Brand, product and category colleagues can explore a study themselves. The AI never does the maths: our statistical engine runs every calculation, table and test, so each question gets the same answer, whatever the experience or phrasing.
Your tools. Your data. The same rigorous evidence.
You shouldn't have to choose between the AI tools you love and the statistical rigor your work demands. Connect SPSS, CSV, Excel, Word or PDF files, or your data lake, and use your own AI. AddMaple stays the calculation engine and the evidence source.
Find what others miss. Show why it matters. Inspire action.
What's new
What's new in Mosaic
A new engine, new ways to build tables, segmentations and stories, and AI that operates the same analysis you do.
The foundation
NewA rebuilt analysis engine
Rebuilt around how research is actually structured, so your study stays a study.
- Multi-selects, Likert scales, batteries, weights and waves survive import
- Every variable charted the moment your data lands
- Long code lists open as Top 15 + Other, so galleries stay legible
Prepare
NewCalculated columns, rebuilt
Build the measures a study needs in the same project, with the definition kept beside the data.
- NPS, CSAT, CES, UMUX-2 and UMUX-4 scores from your scale questions
- SQL-like formulas with CASE, window aggregates and automatic dependency tracking
- Likert top-box and net scores, custom bins and combined multi-selects
Tables
NewData Table Studio
Interactive banner tables and crosstabs, built inside the product.
- Bases, column %, indices, significance letters and z-score shading
- Holm-adjusted p-values, with crosstabs ranked by statistical strength
- Crosstab-level AI to ask what stands out in the table you built
Quant + qual
NewThemes behave like banner variables
AI-coded open ends sit in the same ranked table view as your survey questions.
- Cross, rank and shade coded themes like any other variable
- Use themes in drivers and clustering
- Paste a discussion guide and get one text column per interview question
Dimensions
NewPCA, built into the project
Find the dimensions behind a battery without leaving the study.
- Loadings, variance explained, a component map and an interpretation view
- Optional Varimax rotation for a simpler, more readable structure
- Carry the structure into segmentation and the story
Drivers
NewKey driver analysis with automatic routing
Target-appropriate models on mixed survey data, with the evidence in view.
- Random Forest plus Elastic Net for numeric, binary, ordinal and multiclass outcomes
- Numbers, scales, categories and multi-selects together, with related columns shortlisted first
- Signed, class-specific associations when one global ranking would mislead
Segmentation
NewNew clustering, compared before you commit
Run several approaches on numeric, categorical or mixed survey data and keep the one you trust.
- K-Means, K-Medoids, HDBSCAN and Latent Class Analysis
- Silhouette, PERMANOVA and cluster quality, plus AIC, BIC and entropy for latent classes
- Profile each segment, then save it as a project column for banners, drivers and filters
Deliverables
NewExcel tab books, authored by the engine
Formatted workbooks with a table of contents, brand styling and the full statistical context.
- Categorical and numeric tables, summary batteries and nets
- Optional mean, std dev and median rows, with significance shading
- Draft, validate and publish a tab spec through MCP
Brand tracking
NewMental availability and CEP
Group a battery as a Category entry points family and analyse it as one thing.
- Mental penetration, market share, network size, share of mind and advantage
- CEP × brand matrix with a focal brand and audience or wave splits
- Save results to Insights and Story Dashboards
AI
NewAsk Maple and MCP
AI that operates AddMaple, from inside the product or from the tools you already use.
- Connect Cursor, Claude Desktop, Claude Code and ChatGPT with one sign-in
- Same project permissions as the browser
- Promote agent-built charts into Insights and Story Dashboard pages
Delivery
NewTeam Story Dashboards
Page-based stories built from live charts, not pasted screenshots.
- Live stat blocks that refresh with dashboard filters
- Multi-source dashboards with filter scopes at dashboard, page or block level
- Immutable releases, password-protected links and explorable experiences
Charts
NewPivot dashboards and brand-ready charts
Keep several live pivots in view and make the output look like the client deck.
- Pin multiple pivots on one board, with an optional banner table
- Logo-style category markers on line, dot and bubble charts
- Switch counts, percentages and chart family without rebuilding filters
What we kept seeing
Great research rarely fits neatly into one tool.
We've seen the same problem again and again.
A study might start with survey data, move into crosstabs and significance testing, jump out to Excel for tables, into another package for modelling, somewhere else for qualitative coding, and finally into PowerPoint or a dashboard to tell the story.
Every handoff loses something.
The filters change.
The weighting gets recreated.
Qualitative evidence becomes detached from the numbers.
Statistical context disappears into a slide.
And when someone asks a follow-up question, the researcher often has to go back to the beginning.
We thought there should be a better way. So we went back to first principles.
A new generation
We rebuilt AddMaple from the ground up.
Mosaic is a complete new generation of AddMaple, designed around what we've learned from years of working with researchers and analysing thousands of real-world datasets.
We didn't start with the old product and add more features.
We rebuilt the underlying analysis engine, the data model and the research experience around a simple idea:
All the pieces of your research should work together.
Your survey structure should survive import.
Your weights, filters and segments should mean the same thing everywhere.
Your qualitative themes should be analysable alongside your quantitative data.
Your statistical tests should be visible and inspectable.
Your AI should run real analysis rather than guess at answers.
And the work you publish should remain connected to the evidence behind it.
That is why we call it Mosaic.
Mosaic
Every signal. One picture.
Mosaic brings together the parts of research that have traditionally lived in separate tools:
- Explore
- Crosstab
- Test
- Model
- Segment
- Code
- Ask
- Explain
- Publish
All working on the same underlying research.
So you can start with a dataset and immediately understand what's in it. Discover the relationships that matter. Move into deeper statistical or qualitative analysis when the question demands it. Ask AI to investigate alongside you. And turn the findings into something your team or client can continue to explore.
Without constantly exporting, rebuilding and starting again.
One system
One research system, built around the way research actually works
For years, research software has been organised around individual tasks.
- A tool for tabulation.
- A tool for stats.
- A tool for qual.
- A tool for modelling.
- A tool for dashboards.
And increasingly, a separate AI sitting above all of them.
Mosaic takes a different approach.
It starts with the research itself.
One project. One underlying analytical system. Many ways to interrogate it.
- Human or AI.
- Quant or qual.
- Simple exploration or advanced modelling.
- Analysis or communication.
Different pieces of evidence, kept connected.
Mosaic in depth
New to AddMaple, or would rather read than watch? Seven deep dives, one study.
Each page introduces one part of the product from scratch — what it is, what problem it solves, and how to read what it shows you. Every image is the real product on one illustrative concept test.
Start here · the field guide
What AddMaple is, in one study
New to the product? What it is, who it is for, and how a study moves through it — clean, tabulate, explain, group, connect, publish.
Read →
Table Studio
Crosstabs for the whole study, tested and ordered
What a banner table is, how to read the shading and the letters, and the Excel workbook explained page by page.
Read →
Key drivers
Which answers relate most to the result?
What key driver analysis is, when to use it, and how to read the result screen part by part. Relationships, not causes.
Read →
Audience groups
Who are you talking to?
What segmentation is, how AddMaple chooses the method, what BIC and entropy mean, and how to compare two runs.
Read →
Clean data
From a raw export to a study you can analyse
What the summary dashboard is, what ‘flagged’ means, how fixes are proposed as a batch — and why the source is never touched.
Read →
Ask Maple
A research agent inside the analysis
What the agent can see and do, the panel explained part by part, and one investigation from a question to a draft story.
Read →
Why AddMaple? · comparison
Why not just ChatGPT, Claude or Cursor?
An honest comparison: numbers computed, not written; evidence you can click; methods that fit survey data — and an MCP server so your assistant can use the same engine.
Read →
Under the hood
The Multi-Type Adaptive Discovery Framework
Ask more of a mixed study: find relationships you didn't plan to test, see what still explains an outcome, and compare groups — with the method chosen for the data.
Reports
Three ways to turn a study into an answer
A Topline Report for what the study found. A Strategic Report for what it means. Ranked Impact Analysis for what matters most — including what nobody thought to ask.
What did the study find?
Topline Report
A clear first read: headline results, the audience differences that matter, and the evidence behind each finding.
Read this report →
What does it mean, and what should we do?
Strategic Report
The evidence pulled into themes, tensions and implications — written for a decision, not a questionnaire.
Read this report →Across the study, what matters most?
Ranked Impact Analysis
A systematic scan of the study, ranked so the strongest differences, relationships and patterns rise first.
Read this report →
Story Dashboards
See what connected research looks like
Haven, AI Values, Women’s Freedom and PISA — four stories with the narrative still attached to the evidence. The live dashboards are coming soon.
AddMaple Mosaic
Bring a dataset. Find the key impact drivers.
The value of an insight isn't just what it reveals. It's what happens because of it. Book a walkthrough and bring a past study. We'll run it through our statistics engine and show you which areas could have the biggest impact on the business.