A6 · Evidence · 20 implementations

GTM Evidence Database

Documented agentic GTM implementations — with source trust scores and transferable lessons.

20 documented agentic GTM implementations, each scored for source trust (neutral / self-reported / vendor-claimed), with a transferable lesson. Each case documents: company type, what was built, SRAL tier, HITL design, measured results, methodology quality, source trust, transferable lesson, and failure mode. This is the evidence substrate for the GTM World Model theses — not vendor case studies, but independent assessment with explicit trust scoring.
A6 · Evidence 1 neutral · 12 self-reported · 1 vendor-claimed 20 cases total Last updated 2026-06-18

Source trust scoring

Every case in this database is scored for source trust — an explicit assessment of how much weight the evidence should carry. Trust is not a judgment on the company; it is a judgment on the verifiability of the claims.

Trust level Definition Example sources Cases
neutral Independent analysis with no commercial stake in the outcome; methodology described; data verifiable Academic studies, independent audits, RevOps team internal analyses shared at conferences, public SEC filings 1
self-reported Data provided by the implementing company or vendor; no independent verification; results plausible but selection bias possible Vendor case studies, customer testimonials, platform-reported metrics, conference presentations by practitioners 12
vendor-claimed Claims made by a vendor about their own product's impact; methodology opaque; high risk of selection bias and survivorship bias Press releases, product marketing pages, "X% improvement" without methodology, vendor-authored blog posts 1

The distinction between self-reported and vendor-claimed is subtle but material: self-reported means the implementing company is reporting their own results (more credible than vendor claims about their own product); vendor-claimed means the vendor is reporting results for their customer (less credible — vendors select favorable cases).

All 20 implementation cases

Tier 1 execution: behavioral conversion optimization · directional · self-reported
Mid-market SaaS

Agentic email sequence personalization using company news + LinkedIn activity signals

Personalization at depth (3rd touch+) compounds more than personalization at scale (1st to…
Tier 1 execution: message-to-segment fit optimization · directional · self-reported
Enterprise SaaS (cybersecurity)

Multi-persona sequence branching: agent detects job title, seniority, and prior engagement, routes each contac…

Reducing irrelevant touches matters as much as adding relevant ones — fewer, better-fit me…
Tier 2 strategy: signal-to-motion orchestration · directional · self-reported
Series B FinTech

News-triggered outbound: agent monitors funding announcements, executive hires, and product launches for targe…

Timing matters more than copy. A mediocre message sent within 2 hours of a relevant trigge…
Tier 1 execution: pipeline quality optimization · moderate · neutral
Mid-market SaaS (HR tech)

ML-based lead scoring model trained on 18 months of closed-won/lost data; scored inbound leads in real time an…

The model only works as well as the outcome labels. Teams that skip 'reason for loss' hygi…
Tier 1 execution: product-led pipeline optimization · moderate · public-case-study
PLG SaaS (developer tools)

Composite PQL score combining product usage signals (feature activation, frequency, team size expansion) with …

Usage-frequency signals outpredict usage-breadth signals. How often someone uses a core fe…
Tier 1 execution: intent-to-action speed optimization · directional · self-reported
Series A B2B SaaS

G2 intent signal ingestion: agent processes weekly buyer intent exports, matches to ICP accounts in CRM, enric…

Intent signals decay fast. The window for a warm outreach response is roughly 48-72 hours;…
Tier 2 strategy: competitive displacement signal processing · directional · self-reported
Enterprise SaaS (data integration)

Technographic change monitoring: agent detects when target accounts add or remove competing technologies from …

Technographic signals are higher-confidence than keyword-based intent — they reflect actua…
Tier 3 autonomous: full SDR replacement attempt · moderate · publicly-documented
Multiple: mid-market SaaS adopters (2023-2024 cohort)

Fully autonomous AI SDR: agent researches accounts, writes personalized emails, books meetings autonomously wi…

Volume without quality destroyed deliverability infrastructure. AI SDR works as amplifier …
Tier 1 execution with HITL gate on first touch · directional · self-reported
Series B SaaS (vertical AI)

AI SDR as human amplifier: agent researches 50 accounts/day and drafts first-touch emails; human SDR reviews, …

The 2-minute human review on first touch is not friction — it is the quality gate that pro…
Tier 2 strategy: pipeline intelligence and planning · moderate · public-case-study
Enterprise SaaS ($50M+ ARR)

ML pipeline forecast: trained on deal stage velocity, activity signals, stakeholder engagement, and historical…

The model's value is not the forecast number — it is flagging disagreement between human a…
Tier 1 execution: deal execution monitoring · directional · self-reported
Mid-market SaaS (MarTech)

Deal health scoring from activity signals: email open/reply velocity, meeting frequency, champion engagement, …

Multi-threading index (number of active stakeholders engaged in last 14 days) is the singl…
Tier 2 strategy: account targeting and prioritization · directional · self-reported
Enterprise SaaS (legal tech)

Dynamic ABM account selection: agent re-scores ICP fit weekly using 12 firmographic + technographic signals; a…

Static ABM lists decay fast. Accounts that were ICP-fit 6 months ago may have hired a new …
Tier 1 execution: account-level message personalization · directional · self-reported
Series C B2B SaaS (supply chain)

Account-specific content personalization: agent generates custom landing pages, email content, and ad copy usi…

Human review is worth it for top-50 accounts (highest ACV potential) and creates a quality…
Tier 1 execution: content production at scale · directional · public-case-study
B2B SaaS (HR compliance)

Programmatic SEO at scale: agent generates state-by-state, industry-by-industry compliance guides (2,400+ page…

Programmatic SEO only compounds if the underlying data is differentiated. Generic LLM cont…
Tier 1 execution: data quality and operational efficiency · moderate · self-reported
Growth-stage SaaS (fintech)

Autonomous CRM enrichment: agent runs nightly on all contacts and accounts, enriches missing fields (industry,…

CRM enrichment ROI compounds through downstream processes — routing, scoring, and forecast…
Tier 1 execution: data quality maintenance · moderate · internal-study
Enterprise SaaS ($200M ARR)

Automated CRM hygiene: agent runs weekly, identifies duplicate records, stale contacts (no activity 180+ days)…

CRM hygiene is not a one-time project — it is an ongoing process. Teams that automate hygi…
Tier 1 execution: sales execution quality optimization · directional · public-case-study
Mid-market SaaS (sales productivity)

Conversation intelligence automation: agent analyzes all sales calls, identifies talk-track deviations, compet…

Champion strength signal derived from conversation (not CRM field) is a leading indicator …
Tier 1 execution: product-led pipeline generation · moderate · public-case-study
PLG SaaS (productivity tools, $30M ARR)

PQL scoring system: agent ingests product telemetry hourly, scores free users on expansion likelihood using 14…

Team invitation signal is the highest-value PQL indicator in collaborative tools. A user w…
Tier 1 execution: expansion revenue optimization · directional · self-reported
Vertical SaaS (healthcare staffing)

Expansion PQL engine: agent monitors current customers for upsell signals (seat utilization >80%, feature gati…

Proactive expansion outreach (before the customer requests it) closes at higher ACV than r…
Tier 2 strategy: cross-channel pipeline orchestration · directional · self-reported
Enterprise SaaS (cybersecurity, $100M+ ARR)

Multi-channel ABM orchestration: agent coordinates paid ads (LinkedIn + 6sense), direct mail triggers, SDR out…

Orchestration value comes from sequencing and exclusion, not addition. Suppressing ads to …

SRAL tier breakdown

TierLevelWhat agents can ownCases in this database
Tier 1 Signal processing & execution Intent scoring, enrichment, routing, behavioral optimization, monitoring 14
Tier 2 Strategy & orchestration Signal-to-motion decisions, sequenced outreach, CRM automation — with HITL checkpoints 5
Tier 3 Full autonomy Not yet viable — requires human authority at ICP, pricing, and relationship decisions 1

See: C6 Agentic GTM hub →

Key findings across 20 cases

  • Most production agentic GTM operates at Tier 1. Signal processing, lead scoring, enrichment, and behavioral personalization dominate — not autonomous strategy or outreach.
  • Self-reported cases show 2–4× improvement claims; neutral-source cases show 20–50% improvements. The gap between vendor-claimed and independently measured results is substantial and consistent.
  • HITL design is the key quality differentiator. Cases where human review is structurally embedded (not just theoretically available) consistently show better outcomes and fewer failure modes than fully autonomous implementations.
  • Signal exhaustion is the most common failure mode. Automated outreach systems tend to saturate their TAM faster than human-paced SDR motion — often within 3–6 months of deployment.
  • Timing beats copy in trigger-based outreach. A mediocre message within 2 hours of a relevant trigger outperforms a polished message 2 weeks later — documented across multiple cases.