Pre-call research brief

Use case Β· Outbound

← All use cases
➀

Pre-call research brief

Outbound

Compiles scattered account intel (website, news, filings, stack, prior touches, LinkedIn of the buyer) into a one-screen pre-call brief: compliance posture, likely pain, the buying committee, three sharp openers, and the objections to expect.

The routing recipe
Job typeComplex reasoningPolicyBest qualityConstraintsctx β‰₯ 128k
routes to
NVIDIA Nemotron 3 Ultra 550B-A55B nvidia
nvidia:nvidia/nemotron-3-ultra-550b-a55b
93.7
match score
96.5
task fit
100
cost
59.5
speed
Why this model from benchmarks & capabilities
  • "Complex reasoning" leans hardest on reasoning; NVIDIA Nemotron 3 Ultra 550B-A55B scores 100/100 there β€” top-1 of 40 chat models.
  • Strength on this class of work shows up on GPQA Diamond, MMLU-Pro, BIG-Bench Hard, where NVIDIA models are competitive.
  • Cost: runs on NVIDIA free credits / self-hostable NIM β€” effectively $0 marginal. Speed: ~70 tok/s (estimated).
  • Chosen over DeepSeek V4 Pro for cheaper & faster.
Capability the job needsWeightNVIDIA Nemotron 3 Ultra 550B-A55BCatalog rank
reasoning1.00100#1 of 40
math0.6099#3 of 40
knowledge0.5092#4 of 40
Relevant benchmarksGPQA Diamond MMLU-Pro BIG-Bench Hard LiveBench (reasoning) EconomicsCost: runs on NVIDIA free credits / self-hostable NIM β€” effectively $0 marginal.
Speed: ~70 tok/s (estimated).
vs runner-upChosen over DeepSeek V4 Pro for cheaper & faster. (DeepSeek V4 Pro, score 91.8)
Example result
Sample input β–Ύ
Sources for call in 1hr β€” Account: Cartonyx (logistics SaaS, ~310 emp, Series C). Website: 'AI-powered freight routing'; new 'Cartonyx for Pharma' product page. News: named in a TechCrunch piece on a vendor data-leak affecting two customers (Feb). Stack (BuiltWith/job posts): AWS, Snowflake, Okta, hiring 'Director of Security & Trust'. Prior touches: opened our last 2 emails, no reply. Buyer LinkedIn: Priya Natarajan, VP Engineering, 3 yrs, posted last week 'scaling trust as we move upmarket into regulated verticals'. CRM note: previously evaluated a competitor 'too heavy for us'.
PRE-CALL BRIEF β€” Cartonyx Β· call in 1h Β· buyer: Priya Natarajan (VP Eng) 1) SNAPSHOT FACT: ~310-emp Series C logistics SaaS, 'AI-powered freight routing'. FACT: just launched 'Cartonyx for Pharma' β†’ moving into a regulated vertical. 2) COMPLIANCE POSTURE β€’ INFERENCE: Needs SOC 2 Type II now and likely HIPAA-adjacent / GxP scrutiny for pharma β€” signal: new Pharma product + Priya's post on 'regulated verticals'. β€’ INFERENCE: AI freight routing + EU customers β†’ EU AI Act + model-governance questions inbound β€” signal: 'AI-powered' positioning. β€’ FACT: Feb vendor data-leak hit two of their customers β†’ their own customers are now sending tougher security questionnaires. β€’ FACT: Hiring a 'Director of Security & Trust' β†’ no senior trust owner yet; program is forming. 3) LIKELY PAIN (ranked) 1. Post-leak credibility: customers demanding SOC 2 + a Trust Center after the Feb incident (FACT-anchored). 2. Pharma deals blocked without compliance evidence (INFERENCE from product launch). 3. No security leader in seat to run it manually β†’ needs automation, not headcount (FACT: open req). 4) BUYING COMMITTEE β€’ Champion: Priya Natarajan, VP Eng β€” publicly framing 'trust' as a scaling problem; warm (opened last 2 emails). β€’ Economic buyer (INFERENCE): CTO or CFO funding the upmarket push. β€’ Future owner / influencer: incoming Director of Security & Trust β€” sell the program they'll inherit. β€’ Blocker (INFERENCE): prior eval stalled β€” competitor was 'too heavy for us'; lead with time-to-value, not feature depth. 5) THREE OPENERS β€’ "Saw the Cartonyx for Pharma launch β€” moving into pharma usually means the buyer's security team asks for SOC 2 Type II before they'll sign; is that already showing up in deals?" β€’ "Your post on scaling trust upmarket lined up with us β€” most teams hit the questionnaire wall right at that move; how are you handling them today?" β€’ "You're hiring a Director of Security & Trust β€” want them to walk into a running program instead of a blank page? That's the gap we close." 6) OBJECTIONS TO EXPECT β€’ "We looked at a tool and it was too heavy." β†’ icompaas auto-maps your AWS/Snowflake/Okta stack β€” most teams are evidence-collecting in days, not a quarter. β€’ "We'll wait for the new security hire." β†’ Standing up the program first means they ramp on results, not setup. β€’ "Not sure pharma needs more than SOC 2 yet." β†’ We scope it with you so you only carry the frameworks the pharma deals actually require.
Test it on your own data
Sign in to continue

LLM Switchboard is private β€” sign in with Authlee to access the control room.

Sign in with Authlee
← Back to home