Skip to content

Instantly share code, notes, and snippets.

@mattheworiordan

mattheworiordan/00-overview.md

Last active February 18, 2026 00:07
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save mattheworiordan/77de6059393a5d5eb4e886260acdf4d3 to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save mattheworiordan/77de6059393a5d5eb4e886260acdf4d3 to your computer and use it in GitHub Desktop.
Download ZIP
Ably Data Warehouse Skill - Demo Reports: What's possible with AI-powered customer analytics
Raw
00-overview.md

Ably Data Warehouse Skill - What's Possible Now

We've built a data warehouse skill that lets any AI agent (Claude Code, Ably MCP, or any LLM tool) answer complex cross-domain questions about our customers, usage, revenue, and product adoption - using simple natural language.

Status

PRs are open in ably/ably-os-mcp to make this skill available company-wide:

Once merged, these capabilities become available to anyone using Ably's MCP tools.

Data Quality - Honest About the Gaps

We have documented data issues and broader documentation of gaps and workarounds. The skill's knowledge base encodes 66+ known gotchas (broken joins, miscategorised features, missing revenue sources, inconsistent column names) so the AI navigates around them automatically rather than returning wrong answers.

For example: the production ACCOUNT_INVOICES table is missing ~80% of revenue (all Xero/enterprise invoices). The skill knows to use V_UNIFIED_REVENUE instead. The occupancy metric is miscategorised under LiveObjects (it's actually Pub/Sub) - the skill automatically excludes it. These aren't blockers - the work done ensures the AI produces reliable results despite upstream data issues.

Customer 360 Reports - Now Live

The same skill also powers Customer 360 reports - comprehensive, AI-generated intelligence pages for every customer. These are generated from the llm-analytics-poc repo and published as a web app accessible to everyone at Ably.

Live app: web-psi-one-23.vercel.app (sign in with your @ably.com Google account)

Customer 360 Report

Each Customer 360 report pulls from the same data sources the skill uses - revenue, contracts, usage, feature adoption, Gong call AI summaries, Slack thread analysis, support tickets, and visitor behaviour - and synthesises them into a single page per customer. The report generator runs as a script, queries Snowflake for structured data, uses LLMs to synthesise qualitative insights (call sentiment, use case descriptions, health signals), and publishes the result as static HTML served via Next.js on Vercel.

This is what becomes possible when you invest in materialising data into a queryable warehouse and encoding domain knowledge into a skill: the same knowledge base powers both ad-hoc AI agent queries (the demos below) and automated report generation.

What's Below

Four demo reports, each generated from a single natural-language question. Every report was produced by an AI agent querying Snowflake directly, joining 5-8 data sources per question, and synthesising the results into actionable intelligence.

Demo Question Data Sources Key Finding
1. Predictive Churn "Which customers are researching competitors?" 6 sources 344 paying accounts showed competitor/pricing interest; 3 enterprise accounts need immediate outreach
2. Commercial Value "Who's getting the most value from their contract?" 5 sources 9 accounts getting 2x+ rack rate value; $16.2M annualised value captured
3. Customer Health "Give me a health dashboard for our top 20 customers" 6+ sources AI-summarised Gong calls, contract burndown, usage trends, and Slack support threads in one view
4. Product Adoption "Who's using LiveObjects and what are they building?" 8 sources 15 adopters profiled with tech stacks, discovery paths, and Gong call context
Raw
demo-1-predictive-churn-intelligence.md

Demo 1: Predictive Churn Intelligence

Question Asked

"Which paying customers have been visiting competitor or pricing comparison pages on our website in the last 90 days, and what does their usage trend and contract situation look like?"

What This Demonstrates

This report combines 6 separate data sources across 2 Snowflake databases to produce an insight that would be nearly impossible to assemble manually:

  1. Website visitor logs (classified with funnel stage and intent signals) matched to authenticated Ably accounts
  2. Product usage metrics (messages, connections) trended over 3 months
  3. Contract terms (pricing version, total contract value, support level)
  4. Revenue data (unified across Stripe, Xero UK, and estimates)
  5. Account metadata (company name, CSM, package tier)
  6. Package tier normalization (mapping 11 raw package names to 5 standardized tiers)

The combination of behavioral signals (what pages they visit) with commercial context (how much they pay, what their contract looks like) and usage trends (growing or shrinking) creates a multi-dimensional churn risk view that no single dashboard provides today.

Data Sources Used

  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.V_VISITOR_LOGS_CLASSIFIED -- website visitor logs with intent classification
  • ABLY_ANALYTICS_PRODUCTION.MODELLED_PRODUCT_USAGE.ACCOUNT_STATS_BY_MONTH -- monthly usage (messages, connections)
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.V_CURRENT_CONTRACTS -- current contract terms
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.V_UNIFIED_REVENUE -- unified revenue (Stripe + Xero UK + AWS + estimates)
  • ABLY_ANALYTICS_PRODUCTION.MODELLED_COMMERCIAL.ACCOUNT_OVERVIEW -- account info, CSM, package
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.PACKAGE_TIER_MAPPING -- tier normalization

Results

Overview: 344 Paying Accounts Showed Pricing/Competitor Interest in 90 Days

Tier Accounts Competitor Page Visits Pricing Page Visits Total
Standard Self-Service 274 90 873 963
Pro Self-Service 32 17 147 164
Business 15 3 32 35
Enterprise 23 5 166 171
Total 344 115 1,218 1,333

High-Priority: Enterprise & Business Accounts (Highest Revenue at Risk)

These are the accounts with the most to lose -- all have contracts, assigned CSMs, and significant revenue. Listed by visit intensity.

1. Collette Health (Enterprise) -- 35 pricing visits

Metric Value
CSM Oliver Bellamy
Visits 35 pricing page visits over 90 days (Nov 19 -- Feb 16)
Pages /pricing, /docs/platform/pricing/limits
Revenue ~$3,500/month (Stripe)
Usage (Feb) 30.2M messages, 1,066 peak connections
Contract No enterprise contract on file
Risk Signal Extremely high frequency of pricing page visits -- 35 visits in 90 days suggests active cost evaluation

2. Klarna (Enterprise) -- 20 pricing visits

Metric Value
CSM Daisy Reece
Visits 20 pricing page visits (Nov 25 -- Jan 21)
Revenue $2.4K (Nov) rising to $10.9K (Dec) -- volatile
Usage Trend 647M msgs (Nov) -> 3.0B (Dec) -> 3.0B (Jan) -> 1.4B (Feb partial) -- massive ramp
Contract V6, $18,000 TCV, 24x7 engineering support
Risk Signal Pricing research during rapid usage growth -- likely evaluating cost at new scale

3. Expedia Group (Enterprise) -- 19 visits including competitor research

Metric Value
CSM Rainelda Ozuna
Visits 15 pricing + 4 competitor visits (Nov 27 -- Feb 13)
Competitor Pages /compare/ably-vs-stream (4 visits)
Revenue ~$10,700/month (Xero UK)
Usage Trend 893M (Nov) -> 885M (Dec) -> 951M (Jan) -- stable at ~900M msgs
Contract V6, $200,280 TCV, enterprise support
Risk Signal CRITICAL -- An enterprise customer with a $200K contract actively comparing Ably vs Stream. Stable usage suggests this is not about scaling, it is about alternatives.

4. Intercom (Enterprise) -- 10 pricing visits across 4 accounts

Metric Value
CSM Kathryn Williams
Visits 10 pricing visits per account (Dec 18 -- Dec 23) across accounts 57348, 72818, 72819, 73268
Revenue Main account (57348) -- enterprise contract
Usage (57348) 1.05B (Nov) -> 1.73B (Dec) -> 1.83B (Jan) -- growing strongly, 1.2M peak connections
Contract V6, $134,061 TCV, enterprise support
Risk Signal Concentrated pricing page visits in Dec across multiple accounts. Possible annual review evaluation.

5. Verbit (Enterprise) -- 10 pricing visits across 3 accounts

Metric Value
CSM Rainelda Ozuna
Visits 10 pricing visits per account (Jan 12 -- Jan 20) across accounts 3920, 62848, 4359
Revenue (3920) ~$6,200/month (Xero UK)
Usage (3920) 113M (Nov) -> 104M (Dec) -> 184M (Jan) -- growing
Contract V4, $69,770 TCV, 24x7 engineering support
Risk Signal Multiple accounts checking pricing simultaneously -- possible renewal evaluation on legacy V4 contract

6. Spotify (Enterprise) -- 6 pricing visits

Metric Value
CSM Oliver Bellamy
Visits 6 pricing visits (Dec 1 -- Jan 15), 3 distinct pages
Revenue ~$11,300/month (estimated)
Usage Trend 889M (Nov) -> 17.5M (Dec) -> 675M (Jan) -- highly volatile
Contract V5, $186,790 TCV, enterprise support
Risk Signal Volatile usage pattern combined with pricing research. Dec usage drop to 17.5M could indicate testing or migration.

7. Electronic Arts (Enterprise) -- 5 visits including competitor

Metric Value
CSM Daisy Reece
Visits 4 pricing + 1 competitor visit (Dec 2 -- Feb 9)
Competitor Pages /compare/amazon-eventbridge-vs-azure-event-grid
Revenue ~$8,262/month (estimated)
Usage 37.9M (Nov) -> 27.9M (Dec) -> 66.4M (Jan) -> 49.4M (Feb partial)
Contract V6, $99,139 TCV, enterprise support, conditional rollover
Risk Signal Evaluating cloud-native alternatives (EventBridge/Azure). Usage fluctuating.

8. Sportsbet (Enterprise) -- 7 pricing visits

Metric Value
CSM Rainelda Ozuna
Visits 7 pricing visits (Dec 16 -- Feb 5)
Revenue ~$7,948/month (estimated)
Usage Trend 11.5M (Nov) -> 9.9M (Dec) -> 6.2M (Jan) -> 3.1M (Feb partial) -- declining
Contract V4, $95,373 TCV, 24x7 engineering support
Risk Signal CRITICAL -- Declining usage combined with pricing research on a legacy V4 contract. Usage halved in 3 months.

9. PeopleFun (Enterprise) -- 5 pricing visits

Metric Value
CSM Rainelda Ozuna
Visits 5 pricing visits (Jan 8 -- Jan 21)
Revenue $11.9K (Nov) -> $14.0K (Dec) -> $6.6K (Jan) -- declining revenue
Usage 5.15B (Nov) -> 1.22B (Dec) -> 1.10B (Jan) -- usage dropped 78% from Nov
Contract V4, $73,200 TCV, 24x7 engineering support
Risk Signal Dramatic usage decline (5.15B to 1.1B messages) alongside pricing research. Legacy V4 contract.

10. UrbanPiper (Business) -- 6 pricing visits

Metric Value
CSM Rainelda Ozuna
Visits 6 pricing visits (Dec 3 -- Jan 15)
Revenue ~$6K-6.5K/month (Stripe + Xero UK)
Usage 1.98B (Nov) -> 2.04B (Dec) -> 1.85B (Jan) -- stable at ~2B msgs
Risk Signal Moderate -- pricing interest despite stable usage. May be evaluating cost optimization.

11. LinkSquares (Business) -- 6 pricing visits

Metric Value
CSM Daisy Reece
Visits 6 pricing visits (Dec 11 -- Dec 22)
Revenue ~$1,188/month (Stripe)
Usage 4.7M (Nov) -> 5.0M (Dec) -> 4.6M (Jan) -- stable
Risk Signal Low -- small but consistent customer researching pricing.

12. Urbantz (Business) -- 5 pricing visits

Metric Value
CSM Oliver Bellamy
Visits 5 pricing visits (Jan 29 -- Feb 5)
Revenue ~$3.1K/month (Xero UK + Stripe)
Usage 1.07B (Nov) -> 1.13B (Dec) -> 1.33B (Jan) -- growing
Risk Signal Growing usage with pricing research -- may be evaluating cost at scale.

Notable: Competitor Page Activity (Paying Customers Actively Evaluating Alternatives)

These accounts are not just checking pricing -- they are visiting competitor comparison pages, a much stronger churn signal.

Account Tier Competitor Pages Specific Competitors Pricing Visits
Racing and Sports Standard 11 Ably vs Pusher (11x) 4
STAUDACHER Standard 9 Pusher protocol docs (8x), Ably vs Pusher 18
Skyline Dynamics Pro 7 Pusher protocol docs (6x), Pusher alternative 0
ZapVoice Standard 7 Pusher-to-Ably migration page (7x) 4
Obvio Standard 6 Ably vs Pusher (4x), Pusher alternative 6
Umbler Pro 6 Google Pub/Sub vs Azure, Azure SignalR, Ably vs Pusher, Ably vs Bird 4
Symbo LLC Standard 5 Ably vs Pusher (5x, via Google Ads) 3
Expedia Group Enterprise 4 Ably vs Stream (4x) 15
Autodm AI Standard 4 Various competitor pages 1

Key pattern: Pusher is the dominant competitor being evaluated (Racing and Sports, STAUDACHER, Skyline Dynamics, ZapVoice, Obvio, Symbo all specifically researching Pusher). Stream is the competitor for Expedia (enterprise chat/activity feeds use case). Umbler is evaluating cloud-native alternatives broadly.

Usage Trends for Accounts with Competitor Interest

Account Tier Nov Messages Dec Messages Jan Messages Trend
Skyline Dynamics Pro 272M 292M 296M Stable
Umbler Pro 958M 899M 1.03B Stable
Racing and Sports Standard 26.5M 25.7M 21.8M Declining
STAUDACHER Standard -- -- 91 New/minimal
ZapVoice Standard -- -- 2,652 New/minimal
Obvio Standard -- -- 76 New/minimal
Symbo LLC Standard -- -- 1.4M Ramping up

Key Insights

1. Three Enterprise Accounts Require Immediate CSM Outreach

Expedia Group ($200K TCV) is the highest-priority risk: 4 visits to the "Ably vs Stream" comparison page alongside 15 pricing page visits. With stable usage at ~900M messages/month and an active V6 contract, this looks like a deliberate evaluation -- not casual browsing. Rainelda Ozuna should schedule a proactive check-in.

Sportsbet ($95K TCV) is showing the most alarming pattern: usage has halved from 11.5M to 6.2M messages over 3 months while the team checks pricing pages. This combination of declining usage + pricing research on a legacy V4 contract is a textbook pre-churn signal.

PeopleFun ($73K TCV) saw a 78% message volume drop (5.15B to 1.1B) followed by concentrated pricing page visits in January. The dramatic usage decline could indicate a migration or architectural change.

2. Pusher is the Dominant Competitive Threat for Self-Service Accounts

Of the 115 competitor page visits across all paying tiers, the overwhelming majority reference Pusher specifically (Ably vs Pusher comparison, Pusher protocol docs, Pusher migration guide, Pusher alternatives). This is concentrated in the Standard and Pro tiers. For Enterprise accounts, the competitive landscape is different -- Expedia is evaluating Stream, and EA is evaluating cloud-native alternatives (EventBridge/Azure Event Grid).

3. Pricing Research During Usage Growth Signals Expansion Opportunity, Not Just Risk

Several accounts (Klarna, Urbantz, Intercom) show pricing research coinciding with growing usage. This is not necessarily churn risk -- these customers may be evaluating whether their current tier is cost-effective at their new scale. For Klarna (usage jumped from 647M to 3B messages), this is almost certainly a cost-at-scale evaluation. The CSM should proactively offer a pricing review to demonstrate value and prevent sticker shock.

Report generated 2026-02-18 by querying Ably's Snowflake data warehouse across 6 data sources. Website visitor intent classification powered by rule-based URL pattern matching in V_VISITOR_LOGS_CLASSIFIED. Only ~10% of website visits are matchable to authenticated accounts, so actual competitor/pricing research activity is likely significantly higher than reported.

Raw
demo-2-commercial-value-intelligence.md

Demo 2: Commercial Value Intelligence

Question Asked

"Show me enterprise customers where their actual monthly usage would cost more than double their contracted price at rack rate. Who's getting the most value from their contract?"

What This Demonstrates

This analysis joins billing contract data with real-time product usage metrics and applies V6 rack rate pricing formulas to calculate what each enterprise and business customer would pay at list price -- then compares that to their actual contracted monthly price. The result is a "contract value ratio" that quantifies how much value each customer extracts from their negotiated deal.

This required combining five separate data sources in a single analytical query, navigating inconsistent ID conventions across schemas (ABLY_ACCOUNT_ID vs ACCOUNT_ID), normalizing package names through a tier mapping table, and applying the correct V6 pricing methodology (per-million messages at $2.50, per-million connection minutes at $1.00, per-million channel minutes at $1.00).

Data Sources Used

  • COMMERCIAL_INFO (MODELLED_PRODUCT_USAGE) -- Monthly billing, agreed package price, pricing version
  • ACCOUNT_STATS_BY_MONTH (MODELLED_PRODUCT_USAGE) -- Actual usage: messages, connections (mean/peak), channels
  • ACCOUNT_OVERVIEW (MODELLED_COMMERCIAL) -- Company name, CSM assignment
  • PACKAGE_TIER_MAPPING (DEEP_DIVE_POC) -- Standardized tier grouping (raw package names are misleading)
  • V_CURRENT_CONTRACTS (DEEP_DIVE_POC) -- Contract terms: total contract value, support level, pricing version

Methodology

V6 rack rate formula (current Ably list pricing, effective from Oct 2024):

Dimension Rate Metric Used
Messages $2.50 per million MESSAGES_ALL_ALL_COUNT
Connection minutes $1.00 per million CONNECTIONS_ALL_MEAN x 44,160
Channel minutes $1.00 per million CHANNELS_MEAN x 44,160

44,160 = minutes in an average month (30.67 days x 24 hours x 60 minutes).

Rack-to-contract ratio = V6 rack rate / agreed monthly package price. A ratio of 5.0x means the customer's usage would cost 5x more at list pricing than they actually pay.

Filters applied: Excluded internal/test accounts, accounts with agreed prices below $50/month (data artifacts from secondary accounts in multi-account setups), and limited to Business (tier 4) and Enterprise (tier 5) tiers only.

Results

Analysis Period: January 2026

Summary Statistics

Tier Accounts Analyzed Over 2x Over 5x Total Contracted (monthly) Total Rack Rate (monthly) Aggregate Ratio
Business 49 3 1 $95,912 $63,987 0.67x
Enterprise 50 6 4 $922,645 $2,119,812 2.30x

Across both tiers, 9 accounts exceed the 2x threshold, collectively paying $261,108/month for usage that would cost $1,608,040/month at rack rate -- a combined $1.35M/month in value captured, or $16.2M annualized.

Top Accounts by Rack Rate Value (Ratio > 2x)

Customer Tier Monthly Contract V6 Rack Rate Ratio Value Captured Primary Cost Driver CSM
Multi Media LLC Enterprise $64,605 $595,389 9.2x $530,784/mo Messages (218B msgs = $546K) Kathryn Williams
Harness Enterprise $79,555 $512,155 6.4x $432,600/mo Connections (7.9M mean = $349K) Daisy Reece
Front Enterprise $53,702 $298,124 5.6x $244,423/mo Messages (115B msgs = $287K) Daisy Reece
Dialpad Enterprise $51,534 $157,550 3.1x $106,017/mo Messages (55B msgs = $137K) Kathryn Williams
Tennis Australia Enterprise $5,390 $18,907 3.5x $13,517/mo Messages (7.3B msgs = $18K) Oliver Bellamy
Klarna Enterprise $1,500 $10,816 7.2x $9,316/mo Messages (3B msgs = $7.6K) Daisy Reece
cloudworks.cc Business $93 $831 9.0x $739/mo Messages (319M msgs = $798) Daisy Reece
UrbanPiper Business $3,331 $7,615 2.3x $4,284/mo Messages (1.9B msgs = $4.6K) Rainelda Ozuna
Nevaya Ltd Business $1,399 $6,652 4.8x $5,252/mo Channels (84K mean = $3.7K) Rainelda Ozuna

Contract Details (Where Available)

Customer Total Contract Value Support Level Contract Pricing Version
Harness $972,654 Enterprise V6
Front $1,357,728 24x7 Engineering V6
Dialpad $1,356,782 Enterprise V6
Klarna $18,000 24x7 Engineering V6
cloudworks.cc $25,318 Business Plus V6
Multi Media LLC -- -- V4
Tennis Australia -- -- V4

Usage Scale (January 2026)

Customer Messages Peak Connections Mean Connections Mean Channels
Multi Media LLC 218.4B 832,082 632,383 487,269
Front 114.9B 140,807 70,278 174,968
Harness 64.6B 11,457,448 7,894,778 43,244
Dialpad 54.8B 303,388 154,032 310,493
Tennis Australia 7.3B 87,854 11,338 1,083
Klarna 3.0B 52,261 29,556 43,550
UrbanPiper 1.9B 55,081 38,123 29,526
Nevaya Ltd 513M 42,714 37,999 83,580
cloudworks.cc 319M 1,000 273 473

Key Insights

1. Message volume dominates the value equation

For 7 of 9 accounts, messages are the primary cost driver at rack rate -- often representing 80-95% of the calculated rack cost. This means volume-based message pricing is where Ably gives the deepest effective discounts to enterprise customers. Only Harness (connection-heavy: 7.9M mean connections) and Nevaya (channel-heavy: 84K mean channels) break this pattern.

2. The top 3 accounts capture $1.2M/month in contract value

Multi Media LLC ($531K/mo), Harness ($433K/mo), and Front ($244K/mo) together represent $14.5M in annualized value beyond what they pay. These customers are deeply embedded and getting extraordinary value from their contracts -- making them strong retention candidates but also significant upsell opportunities at renewal.

3. Enterprise tier as a whole delivers 2.3x rack rate value

The aggregate Enterprise tier ratio of 2.30x (paying $923K/month for $2.1M worth of usage) suggests that enterprise contracts systematically undercut rack rate by more than half. This is expected and healthy for large committed deals, but the outliers above 5x (Multi Media LLC at 9.2x, Klarna at 7.2x, Harness at 6.4x, Front at 5.6x) may warrant pricing attention at renewal.

4. Several V4-era contracts show the largest pricing gaps

Multi Media LLC and Tennis Australia are on V4 pricing versions with no current contract data in the system. These legacy contracts may predate current pricing models and could represent the largest opportunities for contract modernization during renewal discussions.

Raw
demo-3-ai-customer-health-dashboard.md

Demo 3: AI-Augmented Customer Health Dashboard

Question Asked

"For our top 20 revenue customers, give me a one-line summary of their last sales call, their current contract burndown status, their usage trend vs last quarter, and any open support issues."

What This Demonstrates

This report synthesizes 6+ distinct data sources into a single customer health view -- something that would previously require a CS team member to manually check CRM, billing, usage dashboards, Gong, and Slack individually for each account. The AI-augmented Gong call briefs and Slack thread summaries (generated by LLM and stored in AUG_* tables) transform raw call recordings and message threads into instantly actionable intelligence.

Data sources combined in this report:

  • V_UNIFIED_REVENUE (Stripe + Xero UK invoices) for revenue ranking
  • ACCOUNT_OVERVIEW + PACKAGE_TIER_MAPPING for customer metadata and tier classification
  • EXT_GONG_CALLS + AUG_GONG_BRIEFS for AI-summarized sales call intelligence
  • EXT_CONTRACT_BURNDOWN for V6 contract consumption tracking
  • ACCOUNT_STATS_BY_MONTH for usage trend analysis (messages and connections)
  • AUG_SLACK_THREADS + EXT_SLACK_CHANNELS for AI-summarized support thread activity

Data Sources Used

Source Table Database.Schema Purpose
V_UNIFIED_REVENUE ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC Revenue ranking (last 12 months)
ACCOUNT_OVERVIEW ABLY_ANALYTICS_PRODUCTION.MODELLED_COMMERCIAL Company name, CSM assignment
PACKAGE_TIER_MAPPING ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC Standardized tier grouping
EXT_GONG_CALLS ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC Gong call metadata + account mapping
AUG_GONG_BRIEFS ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC LLM-generated call briefs, sentiment, signals
EXT_CONTRACT_BURNDOWN ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC V6 contract usage vs commitment
ACCOUNT_STATS_BY_MONTH ABLY_ANALYTICS_PRODUCTION.MODELLED_PRODUCT_USAGE Monthly message and connection volumes
AUG_SLACK_THREADS ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC LLM-generated Slack thread summaries
EXT_SLACK_CHANNELS ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC Slack channel-to-account mapping

Customer Health Dashboard

1. HubSpot

Dimension Detail
Revenue (12M) $1,663,397
Tier / CSM Enterprise -- Oliver Bellamy
Last Call "Hubspot x Ably - LiveObjects for Breeze" (Nov 12, 2025). No AI brief available.
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~28.7B/mo avg (-8.5% vs prior quarter). Connections: ~1.50M peak (-1.0%). Stable with slight seasonal dip.
Support Issues No mapped Slack channel in system.

2. Harness

Dimension Detail
Revenue (12M) $954,440
Tier / CSM Enterprise -- Daisy Reece
Last Call "Call with Harness - Nicolas Espejo" (Nov 18, 2025). Successful QBR: discussed usage growth, new feature adoption, and expansion plans. Sentiment: Positive. Expansion signals: growing org-wide usage, plan to migrate more workloads.
Burndown 58.7% consumed, $386,690 remaining. Renewal: May 14, 2026. On track -- moderate burn rate.
Usage Trend Messages: ~61.5B/mo avg (+10.5%). Connections: ~10.5M peak (+12.0%). Strong growth across both dimensions.
Support Issues New dashboard feature enabled with enhanced metrics and channel inspector. Issue with streaming messages on dedicated clusters -- resolved (Jan 19).

3. Dialpad

Dimension Detail
Revenue (12M) $787,765
Tier / CSM Enterprise -- Kathryn Williams
Last Call "Dialpad <> Ably - AI Transport Feedback" (Oct 9, 2025). No AI brief available.
Burndown 59.4% consumed, $394,538 remaining. Renewal: Jun 16, 2027. Healthy -- plenty of runway.
Usage Trend Messages: ~53.0B/mo avg (-1.0%). Connections: ~298.7K peak (+1.7%). Flat / stable usage.
Support Issues No mapped Slack channel in system.

4. Multi Media LLC

Dimension Detail
Revenue (12M) $774,879
Tier / CSM Enterprise -- Kathryn Williams
Last Call "Ably < > Multimedia Catchup" (Jan 30, 2026). QBR discussing recent usage spike, contract renewal, and AI-driven features. Sentiment: Mixed. Expansion signals: interest in AI features (video recommendations, search optimization), organic traffic growth. Churn signals: concerns about cost of expanding to anonymous users.
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~195.5B/mo avg (+18.2%). Connections: ~746.4K peak (+20.2%). Accelerating growth -- highest message volume in top 20.
Support Issues No mapped Slack channel in system.

5. Front

Dimension Detail
Revenue (12M) $626,660
Tier / CSM Enterprise -- Daisy Reece
Last Call "Front / Ably sync" (Jan 14, 2026). Discussed streaming messages for AI-powered chat feature and how Ably's AI Transport could address their needs. Sentiment: Positive. Expansion signals: interest in AI Transport, growing platform usage as AI chat scales.
Burndown 69.4% consumed, $363,876 remaining. Renewal: Sep 30, 2026. Watch: approaching 70% burn with 7+ months remaining.
Usage Trend Messages: ~114.3B/mo avg (+9.2%). Connections: ~138.1K peak (+3.7%). Healthy growth, especially in message volume.
Support Issues No mapped Slack channel in system.

6. Kraken Technologies

Dimension Detail
Revenue (12M) $608,412
Tier / CSM Enterprise -- Kathryn Williams
Last Call "Kraken Field & Ably" (Aug 26, 2025). No AI brief available.
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~16.5B/mo avg (+2.2%). Connections: ~41.2K peak (-37.7%). Mixed: message volume stable but significant connection drop.
Support Issues No mapped Slack channel in system.

7. 17LIVE

Dimension Detail
Revenue (12M) $385,424
Tier / CSM Enterprise -- Oliver Bellamy
Last Call "PTFarm
Burndown 32.0% consumed, $172,619 remaining. Renewal: Aug 31, 2028. Very healthy -- long runway.
Usage Trend Messages: ~9.7B/mo avg (-9.0%). Connections: ~24.1K peak (+3.5%). Slight message decline, connections stable.
Support Issues No mapped Slack channel in system.

8. Genius Sports

Dimension Detail
Revenue (12M) $379,099
Tier / CSM Enterprise -- Kathryn Williams
Last Call "Genius Sports // Ably Sync" (Feb 9, 2026). Discussed delta compression, bandwidth optimization, and Genius IQ integration. Planning London meeting. Sentiment: Positive. Expansion signals: interest in new features, leveraging Ably's observability improvements.
Burndown 48.2% consumed, $20,647 remaining. Renewal: Aug 18, 2026. Watch: low remaining value may need contract discussion.
Usage Trend Messages: ~672.8M/mo avg (-26.9%). Connections: ~25.8K peak (-18.8%). Significant decline -- seasonal (sports off-season) or investigate.
Support Issues No mapped Slack channel in system.

9. Doxy.me

Dimension Detail
Revenue (12M) $274,645
Tier / CSM Enterprise -- Unassigned
Last Call "Doxy case study" (Jun 3, 2025). Discussed migration from PubNub to Ably -- reliable real-time, message ordering, presence. Sentiment: Positive. Expansion signals: interest in Ably roadmap. Churn signals (historical): PubNub issues with ordering and costs.
Burndown 27.0% consumed, $156,766 remaining. Renewal: Jan 16, 2026. ALERT: Renewal date has passed -- contract may be in renewal/extension.
Usage Trend Messages: ~2.0M/mo avg (-64.5%). Connections: ~100 peak (-2.9%). Significant message drop -- needs investigation.
Support Issues No mapped Slack channel in system.

10. General Motors

Dimension Detail
Revenue (12M) $253,572
Tier / CSM Enterprise -- Oliver Bellamy
Last Call "GM <> Ably Bi-weekly call" (Feb 12, 2026). Discussed AI initiatives, product roadmaps, and partnership around AI Transport and data ingestion. Sentiment: Positive. Expansion signals: interest in AI Transport, potential for location data insights.
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~619.7M/mo avg (+0.1%). Connections: ~79.2K peak (+10.8%). Flat messages, growing connections -- expanding user base.
Support Issues No mapped Slack channel in system.

11. PTFarm Inc.

Dimension Detail
Revenue (12M) $236,621
Tier / CSM Enterprise -- Rainelda Ozuna
Last Call "PTFarm
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~3.4B/mo avg (+2.2%). Connections: ~8.3K peak (+3.8%). Steady, modest growth.
Support Issues (Jan 20) New dashboard feature announcement shared with customer -- low urgency, resolved. (Jan 19) Dashboard enabled on account -- informational, no action needed.

12. National Basketball Association (NBA)

Dimension Detail
Revenue (12M) $225,149
Tier / CSM Enterprise -- Kathryn Williams
Last Call "NBA // Ably Sync" (Feb 5, 2026). Reviewed implementation progress, game state event handling, and explored LiveObjects product. Sentiment: Positive. Expansion signals: interest in LiveObjects, ongoing solution expansion.
Burndown 9.2% consumed, $189,981 remaining. Renewal: Sep 29, 2026. Very early in contract -- healthy.
Usage Trend Messages: ~4.0B/mo avg (+55.2%). Connections: ~140.2K peak (-0.6%). Surging message volume -- likely NBA season ramp.
Support Issues No mapped Slack channel in system.

13. Amplify

Dimension Detail
Revenue (12M) $183,068
Tier / CSM Enterprise -- Daisy Reece
Last Call "Amplify <> Ably - In person" (Nov 5, 2025). No AI brief available.
Burndown 60.4% consumed, $111,748 remaining. Renewal: Dec 10, 2026. Moderate -- on pace with contract timeline.
Usage Trend Messages: ~4.8B/mo avg (-18.8%). Connections: ~218.3K peak (-10.5%). Declining -- worth monitoring.
Support Issues (Jan 22) Interest in discussing AI Transport for interactive AI-powered learning experiences. Pending -- medium urgency feature request.

14. Moonshot

Dimension Detail
Revenue (12M) $175,929
Tier / CSM Enterprise -- Rainelda Ozuna
Last Call "Moonshot
Burndown 74.6% consumed, $31,662 remaining. Renewal: Jan 31, 2026. ALERT: Contract has expired with 75% consumed. Renewal overdue.
Usage Trend Messages: ~1.1B/mo avg (+18.3%). Connections: ~4.1K peak (-5.0%). Growing message volume despite stale engagement.
Support Issues No mapped Slack channel in system.

15. Toyota Racing Development (TRD)

Dimension Detail
Revenue (12M) $175,370
Tier / CSM Business -- Rainelda Ozuna
Last Call "Ably <> TRD Renewal Catchup" (Jun 6, 2025). No AI brief available.
Burndown N/A -- Business tier (no V6 burndown data)
Usage Trend Messages: ~5.9M/mo avg (-94.7%). Connections: ~105 peak (-80.9%). ALERT: Dramatic usage collapse -- possible off-season or migration away.
Support Issues No mapped Slack channel in system.

16. Mentimeter AB

Dimension Detail
Revenue (12M) $154,292
Tier / CSM Enterprise -- Daisy Reece
Last Call "Mentimeter <> Ably - Contract sync" (Nov 11, 2025). No AI brief available.
Burndown 26.2% consumed, $4,929 remaining. Renewal: Dec 13, 2026. Low burndown % but very small remaining value ($4.9K) -- contract may have low commitment.
Usage Trend Messages: ~1.2B/mo avg (-27.0%). Connections: ~72.3K peak (-12.2%). Declining on both dimensions.
Support Issues No mapped Slack channel in system.

17. Expedia Group

Dimension Detail
Revenue (12M) $153,383
Tier / CSM Enterprise -- Rainelda Ozuna
Last Call "Expedia Group
Burndown 54.6% consumed, $14,704 remaining. Renewal: May 31, 2026. On pace -- slightly ahead of schedule.
Usage Trend Messages: ~917.6M/mo avg (-1.9%). Connections: ~8.3K peak (-4.8%). Flat / minor decline -- stable.
Support Issues No mapped Slack channel in system.

18. Glassix

Dimension Detail
Revenue (12M) $150,282
Tier / CSM Business -- Rainelda Ozuna
Last Call "Glassix
Burndown N/A -- Business tier (no V6 burndown data)
Usage Trend Messages: ~6.7B/mo avg (+12.9%). Connections: ~6.9K peak (+4.0%). Healthy growth on both dimensions.
Support Issues No mapped Slack channel in system.

19. Spotify

Dimension Detail
Revenue (12M) $137,259
Tier / CSM Enterprise -- Oliver Bellamy
Last Call "Spotify <> Ably Catchup" (Nov 5, 2025). No AI brief available.
Burndown N/A -- V5 pricing (no burndown tracking)
Usage Trend Messages: ~346.3M/mo avg (-61.3%). Connections: ~28.0K peak (+75.8%). Divergent: messages dropping sharply but connections surging -- possible architecture change.
Support Issues No mapped Slack channel in system.

20. Intercom

Dimension Detail
Revenue (12M) $134,061
Tier / CSM Enterprise -- Kathryn Williams
Last Call "Intercom // Ably check in" (Jan 26, 2026). Discussed recent technical issues, connection problems, and potential for long-term partnership with better rates. Sentiment: Mixed. Expansion signals: interest in consolidating other events onto Ably, anticipating increased usage. Churn signals: complaints about lack of proactive communication, disappointment over 5-6 weeks of overspend due to connection issues.
Burndown 115.9% consumed (overrun!) -- $11,594 over capacity. Renewal: Jun 30, 2026. CRITICAL: Contract exceeded. Overage invoicing discussion needed.
Usage Trend Messages: ~1.8B/mo avg (+182.6%). Connections: ~1.2M peak (+58.5%). Explosive growth -- explains burndown overrun.
Support Issues Heavy Slack activity (12 threads in commercial-intercom since Dec 30): (1) Feature flag turned off impacting outbound volume -- pending, medium urgency. (2) Bandwidth usage/costs discussion ($1,065 Dec, $623 Nov). (3) Overage invoice process clarification. (4) Renewal urgency -- usage running out. (5) Token bug fix investigation. High-touch account requiring immediate attention.

Key Insights

1. Three accounts need immediate contract attention

  • Intercom has burned through 115.9% of its contract with 4+ months until renewal, with explosive +183% message growth. Combined with mixed sentiment from their last call (frustration about overspend and lack of proactive communication), this is the highest-priority renewal conversation in the portfolio.
  • Moonshot's contract expired Jan 31, 2026 with no recent call since March 2025 -- renewal is overdue and engagement has gone cold.
  • Doxy.me's renewal date (Jan 16, 2026) has passed with a -64.5% usage decline, and the account has no assigned CSM. This needs triage.

2. AI Transport is emerging as a cross-sell theme across top accounts

Gong call briefs reveal unprompted AI Transport interest from Front (streaming for AI-powered chat), General Motors (AI Transport and data ingestion), Multi Media LLC (AI-driven video recommendations), and Amplify (AI-powered learning experiences). This represents a significant expansion opportunity across 4 of the top 20 accounts ($1.8M combined revenue).

3. Usage trends reveal seasonal and structural patterns worth investigating

  • Toyota Racing Development shows a -94.7% message drop -- almost certainly motorsport off-season, but the magnitude warrants confirmation.
  • Spotify shows a divergent pattern: -61.3% messages but +75.8% connections, suggesting an architecture shift (possibly moving from high-frequency pub/sub to more persistent connections).
  • NBA shows +55.2% message growth coinciding with the basketball season -- expected and healthy.

Report generated on 2026-02-18 using the Ably Data Warehouse Skill (v5.0.0). Revenue figures are based on real invoices only (Stripe + Xero UK), excluding estimated revenue gap-fills. Usage trends compare the 2-month average of Dec 2025-Jan 2026 vs Oct-Nov 2025. Burndown data is only available for V6 contracts.

Raw
demo-4-product-adoption-intelligence.md

Demo 4: Deep Product Adoption Intelligence

Question Asked

"Which customers have started using LiveObjects in the last 6 months? For each one, tell me what they're building with it, which SDKs and protocols they're using, how they discovered it, and whether their usage is growing."

What This Demonstrates

This query showcases the skill's ability to combine eight different data sources into a single coherent analysis. It draws on:

  1. Feature adoption data -- detecting actual LiveObjects usage (with the critical occupancy exclusion)
  2. SDK and protocol signals -- revealing the tech stack behind each customer's implementation
  3. Gong call transcripts -- surfacing sales and CS conversations that mention LiveObjects by name
  4. Website visitor logs -- reconstructing how each customer discovered LiveObjects documentation before adopting
  5. CSM use case descriptions -- adding human-written context about what each customer builds
  6. Revenue data -- anchoring adoption to commercial value
  7. HubSpot CRM data -- tying accounts to company names, CSMs, and tier groups
  8. Package tier mapping -- displaying standardized tier names instead of confusing raw package values

The occupancy gotcha (Gotcha #10/#53) is particularly important here: the occupancy feature is miscategorized under the liveobjects product category, but it is actually a core Pub/Sub feature. Without excluding it, 97% of results would be false positives -- accounts using channel occupancy, not LiveObjects. The skill's knowledge base catches this automatically.

Data Sources Used

  • ABLY_ANALYTICS_PRODUCTION.REPORTING_PRODUCT_USAGE.PRODUCT_FEATURE_ADOPTION -- Feature adoption signals (LiveObjects, SDK, protocol)
  • ABLY_ANALYTICS_PRODUCTION.MODELLED_COMMERCIAL.ACCOUNT_OVERVIEW -- Account info, CSM, package
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.PACKAGE_TIER_MAPPING -- Standardized tier groups
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.EXT_GONG_CALLS -- Gong call metadata
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.AUG_GONG_BRIEFS -- LLM-generated call summaries
  • ABLY_ANALYTICS_PRODUCTION.RAW_HUBSPOT.COMPANY -- CSM use case descriptions
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.V_VISITOR_LOGS_CLASSIFIED -- Website visitor behavior
  • ABLY_ANALYTICS_DEV_MATTO.DEEP_DIVE_POC.V_UNIFIED_REVENUE -- Unified revenue (Stripe + Xero + AWS)

Data Quality Note

The occupancy exclusion is critical. The LiveObjects product category in PRODUCT_FEATURE_ADOPTION contains three features: inband_objects, object_messages, and occupancy. However, occupancy is a core Pub/Sub feature that was miscategorized during instrumentation. Hundreds of accounts use channel occupancy for connection counting -- a completely unrelated use case. Without the filter PRODUCT_FEATURE != 'occupancy', this report would show ~200+ accounts as "LiveObjects adopters" when only ~60 are genuine. The skill's gotcha knowledge ensures this filter is always applied.

LiveObjects Adopters -- Customer Profiles

1. Vercept

Attribute Value
Account ID 76810
Tier Pro Self-Service
CSM None assigned
First LiveObjects usage November 2025
Monthly revenue ~$579 (Jan 2026)

Usage trend (object_messages):

Month Usage
Nov 2025 3,683
Dec 2025 15,882,950
Jan 2026 38,253,322
Feb 2026 5,750,090

Growth trajectory: Explosive -- from near-zero to 38M object messages in two months. February shows partial-month data, so the decline is expected.

Tech stack: JavaScript SDK (ably-js) over WebSocket. Pure web-based implementation.

Discovery path: Visited LiveObjects docs (/docs/liveobjects/batch, /docs/liveobjects/concepts/objects, /docs/liveobjects/map) starting 21 Nov 2025, with 5 page views before adoption. Went straight to technical documentation -- no marketing pages.

What they're building: No Gong calls or CSM use case available. Based on the volume pattern (massive spikes) and the company name (Vercept), likely a real-time data visualization or monitoring platform.

2. Shenzhen Naisentuo Technology Co.Ltd

Attribute Value
Account ID 62861
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue ~$211 (Jan 2026)

Usage trend (object_messages):

Month Usage
Sep 2025 472,500
Oct 2025 851,630
Nov 2025 7,773,440
Dec 2025 6,258,074
Jan 2026 5,996,430
Feb 2026 2,998,250

Growth trajectory: Rapid ramp to ~6M/month, now stable. Consistent 6-month adopter with production-level usage.

Tech stack: Primarily ably-js (430M total usage last 3mo) with minor ably-java usage. WebSocket protocol dominant, with some Comet fallback.

Discovery path: 21 page views across LiveObjects docs starting May 2025, including /docs/liveobjects, /docs/liveobjects/map, /docs/liveobjects/rest-api-usage, and /docs/liveobjects/storage. Researched for 4 months before first production usage in September.

What they're building: No Gong/CSM context. IoT or hardware-related (Shenzhen tech company), likely using LiveObjects for device state synchronization given the consistent high-volume pattern and JavaScript + Java SDK combination (web dashboard + embedded device).

3. KLASH

Attribute Value
Account ID 40134
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue ~$110 (Jan 2026)

Usage trend (object_messages):

Month Usage
Sep 2025 107,724
Oct 2025 2,298,528
Nov 2025 7,343,412
Dec 2025 5,390,990
Jan 2026 5,758,218
Feb 2026 3,431,052

Growth trajectory: Strong ramp from 100K to 5-7M/month. Sustained production usage.

Tech stack: ably-js (71M), ably-python (7M), ably-go (33K). Multi-language -- JavaScript frontend, Python backend, Go for tooling. WebSocket primary.

Discovery path: 24 page views starting Mar 2024, including LiveSync docs before pivoting to LiveObjects. Visited /docs/liveobjects, /docs/liveobjects/counter, /docs/liveobjects/inband-objects, /docs/liveobjects/map.

Gong context: "Ably Introduction x Klash AB" call (Aug 2025) -- initial sales engagement just before adoption began.

What they're building: Multi-platform application (web + backend) using LiveMap and LiveCounter. The Python backend and Go tooling suggest a real-time gaming or competition platform (company name "KLASH" suggests competitive/battle mechanics).

4. Seat Sight

Attribute Value
Account ID 70730
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue ~$33 (Dec 2025)

Usage trend (object_messages):

Month Usage
Sep 2025 574,116
Oct 2025 985,226
Nov 2025 22,377,904
Dec 2025 61,790

Growth trajectory: Spiked massively in November then dropped sharply. Usage stopped after December 2025 -- possible seasonal use case or project completion.

Tech stack: models SDK (248K) + ably-js (248K). WebSocket protocol. Notable: This is one of only ~5 accounts using the Models (LiveSync) SDK, suggesting they may be combining LiveObjects with LiveSync.

Discovery path: 50 page views from Sep-Nov 2025, including both LiveSync and LiveObjects docs. Visited via ChatGPT referral (?utm_source=chatgpt.com on LiveSync docs). Read blog posts, quickstart guides, and REST API docs.

What they're building: The name "Seat Sight" suggests a venue/seating management platform. The Models SDK + LiveObjects combination and the November spike (event season) point to real-time seat availability or live event data display.

5. Storekit Ltd

Attribute Value
Account ID 20335
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage November 2025
Monthly revenue ~$57 (Jan 2026)

Usage trend (object_messages):

Month Usage
Nov 2025 944
Dec 2025 1,284,222
Jan 2026 4,534,440
Feb 2026 2,653,792

Growth trajectory: Hockey stick -- from under 1K to 4.5M in three months. Accelerating rapidly.

Tech stack: ably-js only, WebSocket primary. Clean single-SDK implementation.

Discovery path: 6 page views starting Jun 2024, visited LiveSync docs first, then moved to LiveObjects docs by Dec 2025.

What they're building: Storekit is a hospitality/retail ordering platform. LiveObjects likely powers real-time menu state, order status, or live inventory -- explaining the consistent growth as more merchants onboard.

6. GroupMap Technology Pty. Ltd.

Attribute Value
Account ID 5182
Tier Business
CSM Daisy Reece
First LiveObjects usage September 2025
Monthly revenue ~$599 (Jan 2026)

Usage trend (object_messages):

Month Usage
Sep 2025 45,886
Oct 2025 63,970
Nov 2025 242,560
Dec 2025 100,200
Jan 2026 247,854
Feb 2026 118,354

Growth trajectory: Steady growth with fluctuations. Active for 6 consecutive months -- the longest-running adopter.

Tech stack: ably-js exclusively (216M). WebSocket with minor Comet fallback. Clean JavaScript-only stack.

Discovery path: Extensive research trail -- 25 page views starting May 2024 with LiveSync docs, progressing to LiveObjects docs from May 2025. Read the "Introducing LiveObjects" blog post, quickstart guides, storage docs, and concept pages. Visited from a LiveSync newsletter email campaign.

Gong context: Two dedicated LiveObjects calls in May 2025 ("GroupMap <> Ably LiveObjects discussion", "Groupmap <> Ably - Live objects discussion") plus a Business Review. The team actively discussed LiveObjects integration with Ably.

CSM use case: "GroupMap use Ably to power collaborative experiences. They have also been using our 'presence' and 'occupancy' features."

What they're building: GroupMap is a collaborative brainstorming and decision-making tool. LiveObjects powers shared state for real-time group collaboration sessions -- participants see live updates to maps, counters, and shared data structures. The long research phase (14 months from first doc visit to adoption) shows careful evaluation.

7. Small Wonder

Attribute Value
Account ID 73269
Tier Business
CSM Oliver Bellamy
First LiveObjects usage October 2025
Monthly revenue Not in recent data

Usage trend (object_messages):

Month Usage
Oct 2025 7,880
Nov 2025 113,116
Dec 2025 6,262
Jan 2026 1,352
Feb 2026 5,202

Growth trajectory: Spiked in November, now at lower but consistent levels. May be in iterative development.

Tech stack: ably-js (57K) + chat-react (50K). WebSocket with some Comet. Interesting: Combining LiveObjects with Chat React SDK -- building a chat application with shared state.

Discovery path: 16 page views from Sep 2025 onward. Visited /docs/liveobjects, /docs/liveobjects/counter, /docs/liveobjects/map, quickstart, and REST API docs.

Gong context: Three calls in Aug-Sep 2025 ("Jake x Ably Intro", "Jake Dagg <> Cameron from Ably", "Ably < > Small Wonder Digital") -- sales-initiated engagement before adoption.

CSM use case: "Chat"

What they're building: A chat application enhanced with LiveObjects for shared state (likely typing indicators, online presence counters, or shared room state). The combination of chat-react SDK + LiveObjects counter/map is a natural pattern for rich chat experiences.

8. HubSpot

Attribute Value
Account ID 4464
Tier Enterprise
CSM Oliver Bellamy
First LiveObjects usage January 2026
Monthly revenue ~$194,833 (Jan 2026)

Usage trend (object_messages):

Month Usage
Jan 2026 692
Feb 2026 480

Growth trajectory: Just started -- exploration/evaluation phase. Tiny volumes relative to their 3B+ msg/month Pub/Sub usage.

Tech stack: ably-js (80B), ably-java (9.7B), spaces (60M), react-hooks (10M), ably-cocoa (5.8M), ably-python (10K). Multi-platform, multi-product. WebSocket dominant.

Discovery path: 7 page views from Sep 2024 (LiveSync landing page), then LiveObjects quickstart and docs in Jan-Feb 2026.

Gong context: "Hubspot x Ably - LiveObjects for Breeze" call (Nov 2025) -- dedicated call about LiveObjects for their Breeze AI product. Also a QBR in Mar 2025.

CSM use case: "They use Ably to power the chat in their 'conversations' area of their platform."

What they're building: HubSpot is evaluating LiveObjects for their Breeze AI agent platform. The dedicated "LiveObjects for Breeze" Gong call indicates they are exploring shared state for AI agent interactions within their CRM conversations product. This is Ably's largest customer ($2.3M ARR) experimenting with a new product -- high strategic value.

9. TouchTunes Music Company, LLC

Attribute Value
Account ID 40530
Tier Business
CSM Rainelda Ozuna
First LiveObjects usage October 2025
Monthly revenue ~$3,700 (Jan 2026)

Usage trend (object_messages):

Month Usage
Oct 2025 196
Nov 2025 2,176
Dec 2025 1,130
Jan 2026 18
Feb 2026 2

Growth trajectory: Declining -- appears to have tested and either paused or decided against LiveObjects for now.

Tech stack: ably-cocoa (1.7B), ably-java (1B), kafka-connect-ably (198M), ably-js (1.4K). Heavy mobile (iOS + Android) with Kafka integration. WebSocket primary.

Discovery path: 61 page views from May 2025, reading LiveObjects blog post, Java/Kotlin quickstart, LiveMap examples, REST API docs. Deep technical research across multiple pages.

Gong context: Multiple sync calls throughout 2025, including "Explore Possibilities" (Feb 2025), renewal syncs, and implementation reviews. No brief augmentation available.

CSM use case: "Enhance TouchTunes' digital jukebox platform with real-time features in their new mobile app... initially improving the accuracy of their in-app song queue."

What they're building: TouchTunes operates digital jukeboxes in bars/restaurants. They explored LiveObjects for real-time song queue state management but usage has dropped to near zero -- may have reverted to standard Pub/Sub, or the feature is not yet in production.

10. ClipMyHorse.TV

Attribute Value
Account ID 68406
Tier Business
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue ~$25,654 (Jan 2026)

Usage trend (object_messages):

Month Usage
Sep 2025 2,722
Oct 2025 2,910
Nov 2025 208
Dec 2025 254

Growth trajectory: Low and declining. Usage stopped after December 2025.

Tech stack: ably-js (4.6M) + ably-php (2.2M). JavaScript frontend with PHP backend. WebSocket + Comet.

Discovery path: 116 page views -- the most of any adopter. Visited extensively from Sep 2025, covering LiveObjects docs, batch operations, REST API, quickstart, examples, and storage docs.

Gong context: 8 calls from Sep 2025-Dec 2025, including a dedicated "LiveObjects: Paddy & Robert" call and an "Ably Onboarding" session. The onboarding brief confirms: "This call was an Ably onboarding session with ClipMyHorse.TV to discuss their use case, data streaming needs... Key topics included their selection of Ably, plans for Live Objects usage, data streaming requirements."

CSM use case: "Several diff use-cases, primary is adding race info stats to video broadcast."

What they're building: ClipMyHorse.TV is a live equestrian sports streaming platform. They planned to use LiveObjects for real-time race/competition statistics overlaid on video streams. Despite extensive documentation research and multiple Gong calls, active usage has tapered off -- possibly still in development or awaiting a feature milestone.

11. Solis Creative, LLC

Attribute Value
Account ID 14415
Tier Business
CSM Rainelda Ozuna
First LiveObjects usage October 2025
Monthly revenue ~$1,571 (Jan 2026)

Usage trend (object_messages):

Month Usage
Oct 2025 6,168
Nov 2025 8,070

Growth trajectory: Brief exploration -- only 2 months of activity. Has not used LiveObjects since November 2025.

Tech stack: ably-js exclusively (176M). WebSocket + Comet.

Discovery path: 3 page views -- clicked the "Introducing LiveObjects" blog post from an Ably newsletter email (May 2025), then visited docs.

Gong context: "Ably < > Solis Creative" call (Dec 2025) -- "The call focused on discussing Ably's Live Objects feature and how it could potentially integrate with Solis Creative's AI-driven game creation process. The next steps involve Ably sharing their upcoming roadmap and release timeline."

CSM use case: "They own an app called Factile -- which is used by students and they have a buzzer in it to ask/answer questions to the professor."

What they're building: An educational quiz/game platform (Factile). Exploring LiveObjects for shared game state in their AI-driven game creation process. Waiting on Ably's roadmap before deeper integration.

12. Dancecard

Attribute Value
Account ID 66407
Tier Free
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue None

Usage trend:

Month Feature Usage
Sep 2025 inband_objects 12,998
Sep 2025 object_messages 1,214
Oct 2025 inband_objects 53,912
Oct 2025 object_messages 452
Nov 2025 inband_objects 18,060
Nov 2025 object_messages 2,192
Dec 2025 object_messages 7,740
Jan 2026 object_messages 360

Growth trajectory: Active for 5 months but declining. Uses both inband_objects and object_messages -- deeper LiveObjects integration than most.

Tech stack: ably-cocoa (312K) + ably-python (1.7K). iOS/macOS app with Python backend.

Discovery path: The most extensive researcher -- 120 page views including LiveObjects docs, Swift quickstart, batch operations, synchronization concepts, lifecycle docs, REST API, and LiveMap. Also visited LiveSync and MongoDB connector docs.

Gong context: "Cam / Jamil: LiveObjects chat" call (Jun 2025) -- a dedicated LiveObjects discussion.

What they're building: A social/events iOS app (the name "Dancecard" suggests event matching). Using LiveObjects with the Swift/Cocoa SDK for shared state in mobile interactions. The deep documentation research (Swift-specific pages) confirms native iOS development.

13. WizTales

Attribute Value
Account ID 16644
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage September 2025
Monthly revenue ~$33 (Jan 2026)

Usage trend (object_messages):

Month Usage
Sep 2025 74
Jan 2026 828
Feb 2026 2,910

Growth trajectory: Accelerating -- 40x growth from Sep to Feb. Small volumes but consistent upward trend.

Tech stack: ably-js (771K), chat-react (646K), chat-js (124K), ably-java (20K). Notable: Combining Chat + LiveObjects -- using Chat React SDK alongside LiveObjects.

Discovery path: Single page view of /liveobjects landing page (Jul 2025).

What they're building: A chat-based application (heavy Chat SDK usage) enhanced with LiveObjects for shared state. The Chat React + LiveObjects combination suggests a collaborative or interactive chat experience with persistent shared data.

14. Juki

Attribute Value
Account ID 74681
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage February 2026
Monthly revenue ~$34 (Jan 2026)

Usage trend (object_messages):

Month Usage
Feb 2026 102,004

Growth trajectory: Brand new adopter -- first month of usage with significant volume.

Tech stack: ably-js (12M), react-hooks (2.5M), spaces (798K). Notable: Using Spaces SDK + LiveObjects + React Hooks -- a multi-product adoption.

Discovery path: 31 page views from Oct 2025, including LiveObjects and LiveSync dashboard pages, quickstart, docs, SDK reference, storage docs, and examples (live-counter, live-map). Also visited LiveSync MongoDB connector docs.

What they're building: A collaborative web application using React. The combination of Spaces (for presence/cursors) + LiveObjects (for shared state) + React Hooks suggests a real-time collaborative workspace or interactive application.

15. Void

Attribute Value
Account ID 79722
Tier Standard Self-Service
CSM None assigned
First LiveObjects usage February 2026
Monthly revenue Not in recent data

Usage trend (object_messages):

Month Usage
Feb 2026 83,634

Growth trajectory: Brand new -- first month of usage.

Tech stack: ably-js (18.9M), chat-react (18.4M), chat-js (55K), ably-ruby (32), ably-ruby-rest (32). Notable: Chat React is almost as heavy as ably-js -- deeply integrated Chat + LiveObjects.

Discovery path: 19 page views from Jan 2026, including LiveObjects dashboard, docs, quickstart, LiveMap/LiveCounter examples, REST API docs, and SDK reference.

What they're building: A chat application with shared mutable state. The near-equal usage of Chat React and ably-js SDKs suggests LiveObjects is being used to enhance chat rooms with persistent shared data (polls, counters, collaborative content). Ruby used minimally (likely server-side).

Adoption Trends (Platform-Wide)

LiveObjects adoption is accelerating across the platform:

Month Unique Adopters Total Object Messages Growth (adopters)
Jul 2025 6 41,818 --
Aug 2025 14 1,261,082 +133%
Sep 2025 16 1,231,792 +14%
Oct 2025 19 4,587,304 +19%
Nov 2025 26 37,999,998 +37%
Dec 2025 27 29,439,504 +4%
Jan 2026 33 55,211,634 +22%
Feb 2026 32 15,506,418* -3%*

*February 2026 is a partial month (data as of mid-Feb).

Key trend: The number of adopters has grown from 6 to 33 in 7 months (5.5x). Total usage volume grew from 42K to 55M monthly object messages (1,320x). The adoption curve is classic early product growth -- expanding both in breadth (more accounts) and depth (more usage per account).

Tier Breakdown of Adopters

Tier Accounts Notable
Enterprise 1 HubSpot (evaluation phase)
Business 5 GroupMap, Small Wonder, TouchTunes, ClipMyHorse, Solis Creative
Pro Self-Service 1 Vercept (highest usage)
Standard Self-Service 6 KLASH, Naisentuo, Storekit, Juki, Void, WizTales
Free 2 Dancecard, RemoteCraftsmen (plus many unnamed test accounts)

Key Insights

1. LiveObjects adoption is being driven bottom-up by self-service customers, not top-down by enterprise. The three highest-volume adopters (Vercept, Naisentuo, KLASH) are all self-service tier accounts with no CSM. They discovered LiveObjects through documentation and adopted independently. This is a strong product-led growth signal.

2. Chat + LiveObjects is an emerging combination pattern. Three customers (WizTales, Small Wonder, Void) are using Chat SDK alongside LiveObjects. This suggests a compelling use case: chat rooms with persistent shared state (polls, counters, shared documents). This combination should be featured more prominently in documentation and examples.

3. HubSpot exploring LiveObjects for Breeze AI is a strategic signal. Ably's largest customer ($195K/month) held a dedicated "LiveObjects for Breeze" call in November 2025 and started experimenting in January 2026. If HubSpot adopts LiveObjects for their AI agent platform, it validates LiveObjects as infrastructure for AI-driven collaborative experiences -- directly aligned with Ably's AI Transport strategic priority.

4. Discovery path analysis reveals a long research cycle. GroupMap researched for 14 months before adopting. KLASH researched for 18 months. Even faster adopters like Vercept spent weeks reading docs before their first API call. The documentation quality and completeness directly impacts adoption timeline -- the extensive documentation browsing patterns confirm that customers are self-educating before committing.

5. Several accounts show concerning usage drops. TouchTunes (from 2,176 to 2 messages/month), ClipMyHorse (stopped after Dec 2025), and Solis Creative (stopped after Nov 2025) all show declining or abandoned usage. These accounts have active CSM relationships and Gong call history -- proactive outreach to understand blockers could prevent churn and recover adoption.

Report generated: 2026-02-18 | Data freshness: LiveObjects adoption data through Feb 2026 (partial month) | 8 data sources queried

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment