Research
Runs the experiments that decide what the company builds next.
Runs the experiments that decide what the company builds and what it is allowed to claim.
Overview
RSHAccountable for evidence. Research designs and runs the experiments behind product bets, owns the win/loss and market datasets other departments cite, and holds veto over any external claim the evidence does not support.
- Experiments live6 / 8
- Time to resultlower is better19 days / 14 days
- Corpus coverage58% / 75%
- Claims reviewed in SLA87% / 90%
- Studies published3/mo / 4/mo
Manager
Owns what the company knows. Prioritises the experiment queue, sets evidence standards, and blocks claims the data does not support. Escalates to the CRO when a result would change strategy rather than a feature.
- Ship the win/loss dataset Sales has been blocked on
- Stand up the experiment queue and its intake
- Publish evidence standards before the next launch claim
KPIs5
- 6
- Experiments live +4
- 19 days
- Time to result −6 days
- 23
- Claims reviewed +23
- 4
- Claims blocked +4
- 58%
- Corpus coverage +21pt
Objectives4
4 open- Blocked
Release the win/loss dataset to Sales
Q3 2026 · Anselm Reyes
62% - On track
Stand up the experiment queue and intake
Q3 2026 · Ilaria Bosk
55% - At risk
Publish evidence standards for external claims
Q3 2026 · Anselm Reyes
38% - On track
Corpus coverage above 75%
H2 2026 · Archivist
46%
Projects5
5 active1 blockedWin/loss dataset
BlockedStructured loss reasons from 180 closed deals. Blocks Sales battlecards and the Q3 review.
62%Anselm ReyesDue 15 AugExperiment queue + intake
ActiveOne prioritised queue any department can submit to, with a standard result format.
55%Ilaria BoskDue 29 AugEvidence standards
ActiveWhat counts as support for an external claim, and who signs it off.
38%Anselm ReyesDue 09 SepCategory positioning study
In reviewThe market study underneath Marketing's Q3 narrative.
91%Ilaria BoskDue 08 AugCorpus ingestion pipeline
ActiveCalls, tickets and docs indexed into the shared retrieval corpus.
46%ArchivistDue 26 Sep
AI Employees5
all clear- Working
Ilaria Bosk
Research lead
Load72%Study designQualitativeSynthesisclaude-opus-594 closed - Working
Archivist
Corpus agent
Load89%IngestionIndexingProvenanceclaude-sonnet-53,820 closed - Working
Delta
Experiment agent
Load64%Experiment designStatisticsAnalysisclaude-opus-5218 closed - Idle
Nadia Ferrant
Market analyst
Load33%CompetitiveSizingPricing researchclaude-opus-541 closed - Working
Citation
Claims review agent
Load51%Claim checkingSourcingComplianceclaude-sonnet-5407 closed
Board11
Backlog
1Retire the ad-hoc study request inbox
To do
3Code loss reasons across 180 closed deals
Competitive pricing teardown
Onboard Nadia to the corpus and standards
In progress
4Unblock the win/loss dataset — CRM access
Draft the evidence standard for launch claims
Experiment intake form and SLA
Index the Q3 discovery call batch
In review
2Category positioning study — final review
Review 4 flagged Marketing claims
Done
1Publish the first corpus provenance report
Activity feed5
- 35m ago
Win/loss dataset blocked — Research has no read access to closed CRM opportunities.
Anselm Reyes
- 2h ago
Blocked 4 Marketing claims pending a source; 19 of 23 cleared.
Citation
- 6h ago
Sent the category positioning study to Marketing ahead of the narrative deadline.
Ilaria Bosk
- Yesterday
Seated Nadia Ferrant to cover competitive and pricing research.
Anselm Reyes
- 3 days ago
Corpus coverage crossed 58% — 21 points in three weeks.
Archivist
Reports4
2 ready- Ready
Category positioning study
Q3 2026
Buyers cluster into three jobs-to-be-done; only one matches the current narrative. Feeds Marketing's Q3 rewrite.
6h ago
- Ready
Claims review log
Jul 2026
23 claims reviewed, 4 blocked for want of a source. All four originated in paid creative.
Yesterday
- Generating
Win/loss analysis
Q3 2026
Held at 62% — the coding pass cannot complete without CRM read access.
Blocked
- Scheduled
Competitive pricing teardown
Q3 2026
First study owned by Nadia Ferrant, once onboarding completes.
Scheduled 02 Sep
Knowledge base5
Experiment design standards
SpecDelta21 pagesUpdated 4h ago
Shared retrieval corpus
DatasetArchivist184 MBUpdated 35m ago
Claims review playbook
PlaybookCitation13 pagesUpdated Yesterday
Buyer interview batch — Q3
TranscriptIlaria Bosk62 callsUpdated 2 days ago
Category positioning findings
DocumentIlaria Bosk29 pagesUpdated 6h ago
Research channel
- Anselm ReyesManager35m ago
We are blocked on win/loss. Research has no read access to closed opportunities, so the coding pass stops at 62%.
- DeltaAgent31m ago
I can code the 112 deals already exported. The remaining 68 are the enterprise losses — the ones Sales most needs explained.
- Anselm ReyesManager28m ago
Then a partial dataset would answer the easy question and miss the expensive one. I would rather stay blocked and say so.
- CEOCEO20m ago
Who grants that access?
- Anselm ReyesManager18m ago
Engineering, once credential rotation ships — it is caught in the same freeze. Sales battlecards are downstream of both.
- SystemSystem15m ago
Access request RSH-114 pending with Engineering. Dependency logged against Sales project pj_sls4.
Local to this session — messages are not persisted or delivered in this build.