As much of the file as fits in one prompt
Every record, enriched by a parallel pipeline
Dimension Labs/Analytics agent for customer language 1Bn+ interactions analyzed
Dimension Labs, the analytics agent for customer language. Mapping the universe of customer language.
Every conversation, review and survey holds a reason your numbers moved. Dimension Labs reads all of them, turns each one into dozens of measurable dimensions, blends them with your business data, and proves what drives your metrics.
What you’re looking atClick, or scroll to travel ↓
Every point is one conversation: a support ticket, a call, a chat, a review, a survey answer. Billions of them, and on their own, noise.
Each conversation is read in full and measured on dozens of axes: intent, effort, churn risk, whether a promise was kept.
The result is rows of typed data, joined to revenue and outcomes, ready to query, trend and test for cause.
Trusted by the NBA, Travelers, Mars, Concentrix and more
[00] The product
The product. An analyst that answers in charts, reports and decks.
Ask a question about your customers the way you'd ask a colleague. Watch it read, enrich, join and test, then hand you something you can send to leadership.
[00.1] The agent at workRead · Enrich · Blend · Prove
The agent at work Live reading · 04 Prove 80 accounts · 782 messages · 8 channels
Know why your numbers move.
Dimension Labs reads every conversation, review and survey, extracts dozens of signals from each one, blends them with your revenue and operational data, and uses causal statistics to prove what drives your metrics. If you can type a question, you can run it.
› What were customers saying before they left? Which signals actually predicted churn?
- 01Read
- 02Enrich
- 03Blend
- 04Prove
Language · unstructuredDimensions · the structure we addBusiness data · structured
Working…
Finding 91% of accounts that wrote “we’re evaluating alternatives” left. 14% of the rest did.
From $299 a month. No demo, no sales call, no contract on monthly plans.
One conversation, measured on many axes 78 Dimensions in the library
Dimensions. Language has more axes than sentiment.
Sentiment scores a conversation on one axis: positive or negative. A single support ticket carries dozens more. What the customer wanted. How hard it was. Whether a promise was kept. Whether a competitor came up. Dimension Labs measures each one as its own axis on every record, so you can compare, trend and test them like any other metric.
[01–06] The systemWhat sits under every answer
Why it works. A model can write a chart. It can’t build the data underneath it.
Dimension Labs is the layer that turns raw customer language into structured, joined, tested data, and keeps doing it every day. Six parts, one system.
01
Blend. What customers said, joined to what they did.
- 1.1Every channel, normalized
- 1.2Joined to revenue, renewal, orders
- 1.3Outside data when it explains the outcome
Enrich. Dozens of signals on every record.
- 2.1Dimensions built for your industry
- 2.250 to 100 per record for some customers
- 2.3Typed fields, not summaries
Pipeline. At scale, every day.
- 3.1Parallel enrichment of every record
- 3.2The same schema on every run
- 3.3Writes back to your warehouse
Prove. Causal statistics, not guesses.
- 4.1Effect sizes with confidence intervals
- 4.2Controls for what confounds the answer
- 4.3Refutation tests on every finding
Deliver. Years of client work, packaged as templates.
- 5.1Templates by use case and vertical
- 5.2Reports, decks and live monitors on day one
- 5.3Scheduled on fresh data
Research. Methods you can check.
- 6.15 published papers
- 6.21Bn+ interactions analyzed
- 6.378 Dimensions in the library
[01] BlendStructured + unstructured
Blend. What customers said, joined to what they did.
Tickets, calls, chats, surveys and reviews sit next to revenue, renewals, orders and usage, row by row. Add the outside world too: weather, seasonality and market events. It isn’t a complete view if half of it is missing.
- 1.1Every channel normalized and comparable, in every language
- 1.2Joined at the account, customer, location or order level
- 1.3Outside data added only when it can explain the outcome
[02] EnrichOne record, by industry
Enrich. Dozens of signals from every record, from day one.
Each conversation is read in full and written back as typed fields we call Dimensions: intent, effort, churn risk, promise kept, competitor mentioned and more. Some customers run 50 to 100 on every record. You don’t design them from scratch: we start you with the set that works for your industry and your question.
Browse all 78 Dimensions › Raw text
We're running a 60-day parallel evaluation of another vendor before we sign the renewal. Honestly, we were promised SSO in Q2 and it still isn't here, and this ticket has been open for three weeks.
Dimensions written to the record
- Competitor Mentionedyes · evaluation underway
- Churn & Cancellation Riskhigh
- Lifecycle Stagerenewal
- Urgency & Deadline60 days
- Expectation Gaproadmap promise missed
- Feature RequestSSO
- Resolution Statusopen · 3 weeks
Each colour links a phrase in the raw text to the Dimensions extracted from it. Hover either side.
[03] PipelineDaily enrichment
Pipeline. Enrichment at scale, every day.
A chat window can summarize a few thousand rows once. Our pipeline enriches millions of conversations in parallel, validates every field, and runs again on new data each day with the same schema. Results write back to your own warehouse.
Language · 37 sources
ZendeskIntercomSalesforceLiveChatDriftSlackQualtricsMedalliaSurveyMonkeyQuestionProGenesysAmazon ConnectAvaya Cloud8x8TwilioSMSAdaKore.aiYellow.aiCognigyRasaMicrosoft Bot FrameworkAmazon LexAmazon AlexaOracle Digital AssistantBotcopyCapacityFacebook MessengerKikX (Twitter)OpenAIEmailCall recordingsApp store reviewsGoogle reviewsOpen-ended surveysCommunity forums
Business + world data
SnowflakeDatabricksBigQueryAmazon S3Salesforce CRMCSV + ExcelREST APIRevenueOrdersRenewalsUsage eventsLoyaltyStore + location dataWeatherSeasonalityMarket events
Dimension Labs pipeline · Runs daily
01
Ingest + normalizeEvery channel and language into one schema
02
EnrichDozens of Dimensions written onto every record
03
Map + governValidated for drift, versioned, the same fields every run
04
Analyze + proveJoined to outcomes and tested for cause
Where it goes
Dimension Labs agentAsk, report, schedule
Your warehouseSnowflake, Databricks, BigQuery
BI toolsTableau, Power BI, Hex
Claude and ChatGPTvia MCP · coming soon
[04] Prove8,504 rated public reviews, Orlando
Prove. Dashboards rank what’s loud. Causal analysis finds what’s costly.
We rebuilt a hotel group’s guest dashboard from public reviews, then asked a different question: not what guests mention, but what actually moves a one-star stay. Front desk and check-in, the dashboard’s number-two priority, causes almost nothing. Staff attitude, mentioned in under 4% of reviews, is the biggest driver of all.
Backdoor adjustmentEffects with confidence intervalsEffects by segmentPlacebo + refutation tests
[05] DeliverOutput from the agent
Deliver. Years of client work, packaged into templates for every use case and vertical.
We’ve delivered this work for insurers, sports leagues, contact centers and consumer brands. Everything we learned is packaged into templates: the right Dimensions, the right analysis and the right report for each use case and industry. Your first answer already looks like the work of a senior analyst.
Reports · Live monitors · Apps · Decks · Charts See every artifact ›
Template 5.1Driver analysisWhat moves NPS, CSAT or ratings, ranked by effect
Template 5.2Churn signalsWhat customers said before they left, measured against who renewed
Template 5.3Location comparisonWhy one site outperforms another
Template 5.4Bot and agent auditWhere automation fails, and what it costs
Template 5.5Root causeWhy a number moved this week
Template 5.6Leadership readoutA report and a deck, checked figure by figure
[06] ResearchPublished methods
Research. Years of work on language as data.
Our methods are published, so you can check them. The research covers how to score experience from text, why sentiment misses most of what matters, and how much signal a piece of language can actually carry.
5Published papers
1Bn+Interactions analyzed
78Dimensions in the library
Signal in the noise
Every enterprise has a dashboard of what happened. None of them have a model of why.
[07] ComparisonAI assistants vs Dimension Labs
“Can’t I just upload this to an AI assistant?” You can. Here’s where it stops.
Export the tickets, drop them into a general AI assistant, ask what’s driving churn, and you’ll get a fluent answer. It reads what fits in one prompt, summarizes it differently every time, and can’t test the answer against your revenue. Dimension Labs is the system underneath: every record structured, joined to your data and proven.
A general AI assistant, file uploadedDimension Labs
A summary that changes every time you ask
Typed Dimensions on every record, the same schema every run
By how often they're mentioned
By measured effect on the metric, with significance tests
Whatever you paste in
Joined to revenue, renewals, orders, and outside data like weather
Starts from zero in every chat
Runs daily, so this month compares cleanly to last year
A blank prompt
Templates and report formats built for your use case and vertical
[08] EcosystemIn development
Coming soon. Your Dimensions, inside Claude and ChatGPT.
We’re building MCP servers for the major AI assistants. Your team can ask in the assistant they already use and get answers built on the same enrichment, the same governed data and the same causal tests. Or use our agent. The layer underneath is the same.
Which customer signals actually predicted churn last year? Use our Dimension Labs data.
tool call
dimension_labs.find_drivers{ "outcome": "churned", "population": "enterprise accounts",
"controls": ["tier", "tenure", "ticket_volume"] } returned · governed Dimensions, tested against the outcome
- alt_vendor_evaluation6.5×p<0.001
- promise_not_delivered4.7×p<0.001
- repeat_issue2.6×significant
Accounts that wrote “we’re evaluating alternatives” were 6.5 times more likely to leave. It’s the earliest warning in your data: for most of the accounts that left, support heard it before any survey did.
[09] CustomersTravelers · Mars · Stake.com
“Dimension Labs gives us that 50,000-foot view we needed to figure out what's going on out there.”
$3.13Min savings delivered to Travelers
600%ROI within 18 months
4Mcall transcripts joined to renewal records at Travelers
1Bn+interactions analyzed
“Structured, reusable conversational data is foundational to our future strategy.”
“We can move from question to insight fast enough that it actually changes what teams do that week.”
“Teams don't wait for reports anymore. They explore insights themselves and use them to set priorities.”
[10] Who it’s forQuestions you can ask today
Who it’s for. Everyone who was told the data team would get back to them.
›Why is NPS dropping, and for whom?›Which ticket themes predict cancellation?›Which accounts are at risk right now, and why?›What separates top agents from the rest?›Did the change we shipped last month work?›Which locations are underperforming, and what's different about them?›What did customers say before refunds spiked?›Which complaints are loud but cost nothing?›Where does the bot fail, and what does it cost?
Enterprise grade security
Built to handle Fortune 500 customer data.
With the controls enterprise security teams ask for, on every plan.
Security at Dimension Labs- SOC 2 Type IIAudited controls over security, availability and confidentiality.
- ISO 27001Certified information security management.
- GDPR and CCPACompliant, with data processing agreements for customers.
- PII redactionPersonal data redacted at ingest, before enrichment.
- Single or multi-tenantCloud hosted, with single-tenant and on-premise options for enterprise.
- SSO and rolesSingle sign-on and role-based permissions, plus full data deletion on request.
Dimension Labs/Start with one file From $299 a month
Stop reporting. Start understanding.
Start with one file. Ask why a number moved. Plans from $299 a month, no demo and no contract on monthly plans.