~50K
B2B SaaS companies with 10–500 reps and multi-feature product catalogs. Every one has a memory problem.
For investors
Institutional memory is the new bottleneck in B2B sales. We’re building the layer.
Salency is the structured, cited, queryable memory layer for what your accounts actually told you. The input every future revenue tool will inherit from. Every call compounds the graph.
Platform & moat
The structured memory layer for B2B sales accounts.
Salency is a system of record for what your accounts actually told you. Every call becomes durable, queryable context mapped to your product catalog, so the next rep, the next quarter, and the next pipeline review all inherit what was said, not what someone remembered to type.
The primitive is structured account memory. The loop compounds: every call makes every future call smarter, for every future rep on that account. Every surface we ship, starting with handoff docs today, inherits from the same core graph. One primitive. Everything else is surface.
Data primitive
Structured account memory, cited, scoped, and compounding, is the input every future revenue tool will be measured against.
Vs CRM, the wedge is the shape of the data. Contradiction pairs, pain evolution over time, confidence-ranked pain → product matches, cross-account pattern graphs. CRM rows hold one value per field; the graph holds rankings, timestamped citations, and cross-call comparisons. Two layers, distinct shapes of data. We sit on top of your stack, not inside it. Reps live in CRM for pipeline stages. Reps live in Salency for the qualitative layer, what the customer actually said, what contradicts what, which pains map to which products.
The uncopyable trio
Pain-product mapping, cross-call contradiction detection, and customer-stated timeline tracking. No direct competitor ships all three as an integrated stack. Together, wired into product-management workflows, the graph compounds per call.
40–50% of customers get visibly annoyed when asked to repeat themselves; 20–30% explicitly express frustration.
Market
A $100M ARR opportunity, on a graph we own.
Bottom-up
Capture target
0.5–1% in 3 years
250–1,000 teams. Revenue scales with category-standard enterprise ACVs.
Validation
Clay: $1M → $100M
AI sales tooling category proves the scale. Clay hit $100M ARR in 2 years (CapitalG-led Series C, $3.1B valuation).
Counterexample
11x.ai raised $74M on autonomous AI SDRs and lost 70–80% of customers. Autonomous bots without structured customer memory don’t hold the relationship. That’s the gap we fill.
The field
Closest threats, and where the shape doesn’t fit.
Pain extraction, pre-call briefing, and handoff briefs are now table-stakes. Gong, HubSpot Frame AI, and AskElephant ship versions of all three. No single competitor ships pain-product mapping, cross-call contradiction detection, and customer-stated timeline tracking as an integrated stack. The gap is the shape of the join, not the individual capabilities.
WhoThreatWhy they’re not there yet
Gonghigh12–18 month window. Already ships pain Smart Trackers (pre-2024), AI Briefer for handoffs (May 2025), Account Console (Feb 2026). Doesn’t ship pain → product matching against a customer-maintained catalog (no Vault primitive) or cross-call contradiction detection.
HubSpot Breezemod-high12 month window. Frame AI acquisition (Dec 2024) and Smart Deal Progression (Apr 2026) put them ~12 months from packaging the matching engine. 200K customers, freemium.
Salesforce Einsteinwildcard18–24 month window. Account Research Agent (Dreamforce 2025), Data 360 customer graph, Agentforce 360. Recipe published in Architect docs but not packaged as a SKU. 150K customers.
AskElephantmoderate6–12 month window. Pre-call briefs already ship per third-party reviewers. $6M seed (May 2025) with a scale-AI-agents mandate. $99/mo anchor. Doesn’t ingest a product catalog or do pain-product matching.
Clari + Salesloftlow12–18 month window. Merged Dec 2025, $450M combined ARR. Integration paralysis: 2 CI stacks plus 2 engagement stacks to rationalize. CEO Cox’s public language is 100% forecast-execution unification. FAQ says full unification “coming years.”
03.
Why this team, on this problem, now.
Why now
The bottleneck is about to shift.
Sales teams are consolidating tools and cutting rep headcount while deal complexity keeps rising. The surviving reps own more accounts with less tribal knowledge, and AI agents will start drafting their outreach inside the next eighteen months.
Both of those depend on structured account memory as the input. Not a raw transcript pile, not another CRM field. The agent doesn't know what the buyer has already told you. The rep inheriting the account doesn't either. The asset that makes either one useful is the same asset.
Every LLM commoditizes extraction. What doesn’t commoditize is the customer-built graph of pains, products, and contradictions, and that graph compounds per account, with every call. The defensibility is the specific join, pain → product → contradiction across time, wired into the PM tools where roadmaps already live.
Salency is building that layer. The longer a team runs it, the more context compounds. The graph keeps getting richer with every call.