# Gong Panel Exercise: Research Notes ## WhatsApp vs. Slack Connect Messaging Integration --- ## THE BOTTOM LINE This is more nuanced than "pick WhatsApp or Slack Connect." Both have significant technical and strategic complexity. The winning answer is probably about sequencing, scope, and what problem you're actually solving. --- ## WHATSAPP IN APAC ### Market Position - WhatsApp has 58% share across APAC messaging - 3.27B monthly active users globally; Southeast Asia grew 19.4% YoY - APAC generates $127M+ in WhatsApp Business revenue (2026) ### Country-Level Reality | Country | Dominant App | WhatsApp Position | |---------|-------------|-------------------| | Singapore | WhatsApp (73%) | Dominant, 97% penetration | | Japan | LINE (~70-100%) | Minimal presence | | South Korea | KakaoTalk (near-universal) | Minimal presence | | Malaysia | WhatsApp | Strong (51% SME adoption) | | Indonesia | WhatsApp | Growing (39% business adoption) | | Thailand | LINE (50M+ users) | Minor | **Critical finding:** WhatsApp alone only covers Southeast Asia (Singapore, Malaysia, Indonesia). Japan needs LINE. Korea needs KakaoTalk. "WhatsApp for APAC" is really "WhatsApp for SEA." ### B2B Sales Usage - 67% of Singaporeans prefer WhatsApp for business communication - B2B teams report 3-5x higher reply rates vs. cold email - 18-25% conversion rates (3x higher than email's 6-7%) - Used for: initial outreach, deal negotiation, proposals, relationship management - Case studies: Bank Central Asia (5x conversions), HDFC (13K leads in 12 months) ### Technical Blocker: End-to-End Encryption - WhatsApp E2EE prevents server-side access to message content - No native conversation archival API (unlike Slack, Teams, Zoom) - No competitor has solved WhatsApp ingestion for conversation intelligence - Would require client-side capture or user-initiated message logging - This is likely why no CI platform has built this integration yet ### Platform Risk: Meta's AI Lockdown (Jan 2026) - Meta banned third-party general-purpose AI chatbots from WhatsApp Business API (effective ~Jan 2026) - Goal: Make Meta AI the sole AI assistant on the platform; consolidate AI presence - Updated terms explicitly prohibit businesses from using WhatsApp data to train or fine-tune AI models - Exceptions: Businesses can still run specialized chatbots for direct customer interaction (bookings, orders, support) - Regulatory backlash: Ban paused in Italy and Brazil due to market dominance scrutiny - Meta AI is integrated directly into WhatsApp and cannot be removed by users - Interactions with Meta AI are used by Meta to train its own models - Strategic direction: Meta wants WhatsApp Business focused on commercial, revenue-generating activities (marketing, support), not third-party AI analysis **Implication for Gong:** Even if Gong solved the E2EE technical barrier, Meta's platform policy actively blocks the use case. Building a WhatsApp integration means building on a platform whose owner is moving against third-party AI access. This is a compounding risk: technical + regulatory + platform. ### Privacy/Compliance - Singapore (PDPA): All-party consent required for recording (strictest) - Japan (APPI): One-party consent for personal; consent for business use - South Korea (PIPA): One-party consent; third-party recording illegal without both parties - Cross-border data transfers restricted under all three frameworks --- ## SLACK CONNECT IN ENTERPRISE SALES ### Adoption - 80% of Fortune 100 use Slack Connect - 91K+ customers; 10M+ active users - 100M+ inter-company messages exchanged weekly - 35% usage surge YoY ### How Sales Teams Use It - **Deal rooms:** Dedicated channels for specific deals (auto-created at $100K+ thresholds) - **Customer success:** Post-sale onboarding, VIP support channels - **Cross-company collaboration:** Proposals, contracts, proof of concept work - Snowflake case study: One AE increased deal size 3x YoY using Slack Connect - 15% increased productivity in deal closure ### Revenue-Critical Content - Deal progression, competitor mentions, buyer sentiment - Proposal feedback, objection handling, contract negotiation - Budget/pricing discussions, stakeholder concerns - This is genuine deal activity, not just operational chatter ### Technical Feasibility: Moderate but Constrained - API rate limiting: 1 request/minute, 15 objects max per request - External member data excludes email addresses (privacy by design) - Channel IDs can change (data consistency challenge) - Must use Conversations API (older APIs don't work reliably) - Rate limits suggest Slack doesn't intend high-volume data extraction ### Privacy/Governance Concerns - Cannot prevent external partner from accessing/exporting shared channel data - Shared channel = customer owns their data too - External parties need consent for conversation analysis - Financial services have compliance objections to unmonitored shared channels --- ## GONG'S CURRENT STATE ### Channels Currently Captured - Phone calls, video meetings (Zoom, Teams, Google Meet) - Email interactions - SMS (Zoom Phone, Dialpad via native integrations) - WhatsApp: No direct native integration (third-party automation only) - Slack: Push notifications TO Slack, not data ingestion FROM Slack ### Recent Product Evolution - **Mission Andromeda (Feb 2026):** AI coaching (Enable), conversational AI (Assistant), account management, MCP support. No new messaging channel ingestion. - **MCP Support:** Bi-directional: external data INTO Gong features, external AI agents querying Gong. Partners include HubSpot, Salesforce, Microsoft. - **Revenue Graph:** Stitches calls, meetings, emails, texts into connected intelligence layer ### Key Gap Gong is still fundamentally a synchronous conversation intelligence platform (calls + video). Asynchronous messaging (Slack, WhatsApp, chat) is secondary or absent. --- ## COMPETITOR LANDSCAPE | Platform | Messaging Capabilities | |----------|----------------------| | Chorus (ZoomInfo) | Calls, video only. No messaging. | | Clari | Email, calendar, LinkedIn (via Outreach). No chat/messaging. | | Salesloft + Drift | Only platform with omnichannel messaging (chat, SMS, email, voice) via Drift acquisition | | Outreach | Email, calls, SMS, LinkedIn. Slack via Zapier only. | | Revenue.io | Zoom, Teams, Google Meet, Gmail. No messaging at all. | **No major CI platform is ingesting Slack Connect or WhatsApp data.** The industry is still call/video-first. Salesloft + Drift is the closest to omnichannel, but that's chat-widget messaging, not Slack/WhatsApp ingestion. --- ## MARKET CONTEXT: THE DARK FUNNEL - 70-80% of B2B buying decisions occur in untracked "dark funnel" channels - Private Slack conversations, LinkedIn DMs, WhatsApp messages, community groups - $100M+ in open pipeline discovered in dark funnel channels before deals hit CRM - Gartner: By 2028, 60% of seller work happens via conversational AI interfaces - Buyers use average of 10 different interaction methods across their journey - Enterprise messaging market: $8.7B (2024) growing to $34.2B (2033) at 16.5% CAGR --- ## TECHNICAL CONSIDERATIONS: MESSAGING vs. CALL ANALYSIS | Challenge | Calls/Video | Messaging | |-----------|------------|-----------| | Structure | Linear, turn-based | Non-linear, threaded, async | | Transcription | Solved (90%+ accuracy) | Not needed, but threading/context reconstruction required | | Sentiment | Clear from tone, pace, emphasis | Ambiguous across short messages; emoji interpretation | | Context | Sustained narrative | Fragmented; participants respond hours/days later | | Speaker tracking | Clear labels | Implicit or missing; multiple parallel threads | | Coaching signals | Talk ratio, filler words, objection handling | Response time, message length, escalation patterns | **Key challenge:** Async conversation analysis requires explicit methods for dialogue act detection, adjacency pair identification, and participant role assignment. These are implicit in synchronous meetings. --- ## REGIONAL B2B COMMUNICATION PREFERENCES - **North America:** Direct, speed-oriented, digital-first. Slack is native to tech sales workflows. - **APAC:** Relationship-intensive, messaging-heavy (WhatsApp in SEA, LINE in Japan, KakaoTalk in Korea). Higher AI adoption at scale than other regions. - **EMEA:** Longer decision cycles, higher governance requirements, platform preferences vary by country (Xing in Germany, LinkedIn elsewhere). --- ## KEVIN'S ORACLE ANECDOTE (USE IN DECK) At Oracle, Kevin ran the social management platform (Engage) and managed network integrations. Sales teams would come in and say "We want Snapchat." When Kevin pressed for use cases, they fell silent. He'd tee them up: Do you want post scheduling? Analytics? Listening? Nothing concrete came back. He also pointed out that Snapchat had no API at the time. But he didn't just say no. He told them: give me solid use cases and outcomes you're looking to drive, and that's enough to go back to Snapchat for a potential partnership conversation. **Why this matters for the deck:** This is the exact same pattern as the WhatsApp request. APAC customers say "We want WhatsApp." But what do they actually need? Conversation recording? Activity visibility in the deal timeline? Metadata in the Revenue Graph? The answer changes the solution dramatically, and right now, the platform constraints (Meta's API lockdown, E2EE, data training prohibition) make the obvious interpretation impossible, just like Snapchat's missing API. **How to use it:** This anchors the "how I evaluate customer requests" section. It shows Kevin has lived this pattern before and his instinct is: don't dismiss the request, don't blindly build toward a wall, interrogate the use case. That's the framework. **Extra context from Kevin's site:** At Oracle, Kevin owned Engage AND network integrations specifically. He shipped LinkedIn, Instagram, Sina Weibo, and Tumblr, doubling supported networks. So this isn't a PM who happened to get a request. He was literally the person whose job was to evaluate which networks to integrate, ship the ones that made sense, and say no to the ones that didn't. Four shipped, Snapchat declined. That's the track record. --- ## KEVIN'S BACKGROUND CONNECTIONS TO THIS CASE These proof points from Kevin's background map directly to the panel exercise. Use them naturally, not as a list. **Network/platform integration evaluation (Oracle):** This is the direct precedent. Kevin owned Engage AND network integrations. He partnered directly with social networks to evaluate and ship LinkedIn, Instagram, Sina Weibo, and Tumblr across Admin, Publish, Engage, and Analyze modules. Declined Snapchat (no API, no concrete use cases). Doubled network coverage from 4 to 8. 36 features in 12 months. His approach: partner directly with each network, evaluate API capabilities, assess customer demand, build repeatable GTM playbooks. This is exactly what evaluating WhatsApp vs. Slack Connect requires. From the case study: Kevin used "war room" execution to complete 5 epics in a single week at critical milestones. Managed two agile teams across Austin (Platform), Atlanta (Frontend/Design), and India (Engineering). Built async-first documentation with strategic sync meetings to bridge time zones. This maps directly to working with Gong's Israel and Ireland teams. **International platform partnerships (Sendoso):** Deeper than just "47 countries." Kevin negotiated directly with vendors (Tango, XOXODAY, Gyft, NGC, TOPPS) rather than relying on aggregators, which yielded better terms and faster onboarding. Landed Square via WeGift API integration, enabling small business targeting in 4 markets. Managed different APIs, terms, and regional requirements across multiple partners simultaneously. Coordinated 12 market launches with varying currencies and catalog requirements. Final catalog: 1,764 total eGifts (1,420 international, 344 U.S.). Key insight from case study: "Direct vendor negotiations outperform aggregator-dependent models." Relevant when discussing whether Gong should go through a WhatsApp BSP or build direct. Also consolidated three parallel redemption flows into one unified experience. This "build one framework that handles multiple channels" thinking is exactly what the messaging integration strategy needs. **Enterprise integration architecture (Lever):** HRIS Sync was a strategic sequencing decision. Kevin's approach: Phase 1 (operational efficiency: self-serve bulk import to unblock deals immediately), Phase 2 (platform capabilities: HRIS Sync as enterprise infrastructure). Chose SCIM over Merge.dev because "every enterprise IT stack supported it" and it provided long-term flexibility. This standard-over-shortcut thinking matters for the messaging integration question. Results: 37% Q1 engagement. Migration from days to minutes. Built team from zero to seven. Ran CAB with 12+ customers. Partnered across teams without direct authority. Key insight: "Sequencing matters: quick wins build organizational confidence before tackling complex platform work." This could frame the phased recommendation in the deck. **Co-branded partner platform (Rocket Lawyer / Covea):** Built first enterprise partnership reaching 11.5M partner customers. Designed for replication, not one-off integration. Chose OpenID Connect (long-term scalability despite short-term complexity). Built a Configuration API for runtime control over branding/pricing/features without release cycles. Created a custom connector as isolation layer for partner-specific exceptions. Result: 16% of EU traffic from partner channels. Framework powered subsequent partnerships. Kevin's approach: "absorbing communication complexity to protect engineering focus." This matters for the deck because it shows Kevin builds platforms, not point solutions. The messaging integration recommendation should follow the same pattern: build a messaging abstraction layer, not a WhatsApp-specific hack. **Regulatory-to-product translation (GridStrong):** Translated complex energy compliance standards into clear product plans. Same skill needed here: PDPA, APPI, PIPA, Meta's ToS all need to be translated into product implications. **Internationalization as product work (Oracle):** Internationalized across 30+ languages with RTL support and locale-aware formatting. Kevin positioned this as "a product investment rather than simple localization." Personally mastered non-English interfaces to advocate for quality. Integrated Sina Weibo for Chinese market. **Bridge builder / matrixed influence:** The prompt describes helping a peer PM set direction. Kevin's whole career is this: "I do my best work between teams that don't usually talk to each other." At Lever, partnered across PS, implementation, CS, and support without direct authority. At Oracle, coordinated across four offices in three time zones. This is the role. --- ## STRATEGIC IMPLICATIONS FOR THE DECK ### The Framework Answer Customer request evaluation should weigh: signal strength (how many customers, how much revenue), strategic alignment (does it serve growth goals), technical feasibility (can we actually build it), opportunity cost (what are we not building), and time-to-value (how fast does this create impact). ### The Specific Analysis Neither WhatsApp nor Slack Connect is a simple "build it" decision: **WhatsApp:** - Strong market signal in SEA but technical architecture (E2EE) makes traditional CI ingestion impossible - Would require a fundamentally different approach (metadata analysis, opt-in message sharing, CRM sync rather than conversation recording) - Only covers part of APAC; Japan and Korea need different platforms - First-mover advantage is real since nobody has solved this **Slack Connect:** - Revenue-critical conversations are happening there (deal rooms, proposals, negotiations) - API constraints are real but not insurmountable for async analysis - Stronger alignment with existing Gong customer base (NA tech companies already on Slack) - Privacy/governance concerns around analyzing shared channel data ### The Recommendation Direction Could argue for a phased approach: 1. **Short-term:** Slack Connect metadata/signal integration (lower technical risk, immediate value for existing NA customers, aligns with original roadmap impetus) 2. **Medium-term:** WhatsApp Business API integration focused on what's technically feasible (metadata, CRM sync, opt-in message capture) for APAC expansion 3. **Long-term:** Build a messaging ingestion framework that can extend to LINE, KakaoTalk, Teams as market needs emerge But the beauty of the case is that there are multiple defensible positions. The panel wants to see the thinking, not a specific answer. --- ## GONG MCP CAPABILITIES (RESEARCHED APRIL 2026) ### Architecture: Two-Way Intelligence Gong's MCP support (announced as part of Mission Andromeda, Feb 2026) has two components: **MCP Gateway:** Pulls external data INTO Gong's features (AI Briefer, AI Ask Anything). Partners feed their data into Gong's intelligence layer. **MCP Server:** Lets external AI agents QUERY Gong directly. Salesforce Agentforce, Microsoft Copilot, HubSpot agents can all pull Gong's revenue intelligence into their workflows. This is bidirectional by design. Gong becomes a hub, not a silo. ### Launch Partners - Microsoft (Dynamics 365, Microsoft 365 Copilot) - Salesforce (Agentforce) - HubSpot (CRM) - Gong Collective: 300+ members, 80+ active technology/consulting partners ### What the MCP Server Exposes (Today) Based on open-source and third-party implementations: | Tool | What It Does | |------|-------------| | list_calls / search_calls | Retrieve calls with date filtering, participants, duration | | retrieve_transcripts / get_call_transcript | Full conversation transcripts with speaker labels, timestamps, topics | | get_call_details | Complete metadata: parties, outline, outcomes, key points, questions, topics, trackers | **What's NOT exposed:** Email content, deal data, CRM objects, or any messaging channel data. The MCP server is currently call/video-centric, mirroring Gong's core data model. ### Third-Party MCP Ecosystem - **Composio:** 53 tools for Gong including call management, transcripts, user management, CRM objects, analytics, scorecards. Broader than Gong's own MCP server. - **Workato:** 3 core tools (search_calls, get_call_details, get_call_transcript). OAuth 2.0 auth. Positioned for meeting prep and deal review. - **Qualified:** MCP integration feeding account engagement data INTO Gong's Account Briefer. "Coming Soon" status. Shows the Gateway direction working. ### Strategic Quotes Eran Aloni (EVP Product Strategy): "With MCP support, customers leverage Gong's data, insights, and AI agents directly within their existing revenue stack." Microsoft's Karan Nigam: MCP enables "Gong's revenue AI agents to connect into Microsoft's platforms and AI Agents." Salesforce's Josh Israels: Bidirectional access to "partner data, insights, and workflows inside both Gong and Salesforce." ### WHY THIS MATTERS FOR THE DECK MCP is a potential alternative pathway for messaging integration. Instead of Gong building native WhatsApp/Slack ingestion (with all the technical and policy barriers), MCP could enable: 1. **Gateway approach:** A WhatsApp BSP or Slack app pushes messaging signals INTO Gong via MCP Gateway, feeding AI Briefer and Ask Anything without Gong needing to build/maintain direct integrations. 2. **Server approach:** External messaging analytics tools query Gong's call data to create unified conversation views, with Gong providing the revenue context. 3. **CRM-mediated path:** WhatsApp messages stored in HubSpot/Salesforce flow into Gong through existing CRM MCP integrations, without Gong touching WhatsApp directly. **The honest assessment:** MCP is currently call-centric. Messaging data isn't in the protocol yet. But the architecture is designed to be extensible, and the Gateway pattern (Qualified pushing data IN) is exactly the model that could work for messaging metadata. This is a "build the rails now, add messaging cars later" argument. **For the deck:** MCP belongs in the recommendation as enabling infrastructure. It reframes the question from "should Gong build WhatsApp or Slack Connect?" to "how does Gong build a messaging abstraction layer that MCP partners can feed into?" That's a stronger, more platform-minded answer.