neesh Inc.
FOUR AI ENGINES WE BRING TO PROJECTS

Each product is an instrument you can watch working.

Four AI engines we bring to projects: purpose-built, and configured to your use case or built on for your own. One is available now and three are in build; this page says which is which, per product and per capability. Every panel below runs its product’s real decision, including the case where it declines to act.

AskBase The Research Instrument AVAILABLE_NOW
AskBase answering, and declining to guess A console showing two exchanges. The first question is answered and cited from three sources the asker is permitted to see, with two further matches filtered out by permission. The second question uses an ambiguous business term, so AskBase refuses it at a confidence of 0.42 against its 0.55 floor, says "Not sure, please confirm", and lists what it searched. QUERY_01 ANSWERED Which client MSAs auto-renew before 31 March? Four agreements auto-renew before 31 March. Two of them need 60 days’ notice to stop. That window closes 30 January. MSA_Thornfield.pdf §7.2 · renewal & notice renewals-2026.xlsx rows 11–14 #legal-ops 14 Feb · thread CITED FROM 3 SOURCES · 2 MORE MATCHED, FILTERED OUT BY PERMISSION QUERY_02 REFUSED How many active clients do we have? Not sure, please confirm. CONFIDENCE 0.42 · FLOOR 0.55 “Active” resolves three ways here: billing status, last engagement, or an open matter. SEARCHED · crm.accounts · billing.subscriptions · 4 policy documents
AskBase answering, and declining to guess A console showing two exchanges. The first question is answered and cited from three sources the asker is permitted to see, with two further matches filtered out by permission. The second question uses an ambiguous business term, so AskBase refuses it at a confidence of 0.42 against its 0.55 floor, says "Not sure, please confirm", and lists what it searched. QUERY_01 ANSWERED Which client MSAs auto-renew before 31 March? Four agreements auto-renew before 31 March. Two of them need 60 days’ notice to stop. That window closes 30 January. MSA_Thornfield.pdf §7.2 · renewal & notice renewals-2026.xlsx rows 11–14 #legal-ops 14 Feb · thread CITED FROM 3 SOURCES 2 MORE MATCHED, FILTERED OUT BY PERMISSION QUERY_02 REFUSED How many active clients do we have? Not sure, please confirm. CONFIDENCE 0.42 · FLOOR 0.55 “Active” resolves three ways here: billing status, last engagement, or an open matter. SEARCHED · crm.accounts billing.subscriptions · 4 policy documents
READS QUERY_02 is the text-to-SQL engine (the capability marked in build below), and the 0.55 refusal floor is its own. The permission-filtered count is how every answer reports itself. The two questions are illustrative.

Ask. The answer resolves with its sources.

FlowBase The Command Center IN_BUILD
FlowBase routing two emails to two different outcomes A decision trace: classify, then judge, then a branch to either auto-execute or the review queue. Below it, two emails. A routine maintenance request scores 0.91 against the 0.85 threshold and is auto-executed. A lease-termination email scores 0.62, stops short of the threshold, and is held in the review queue with its full context. DECISION_TRACE THRESHOLD 0.85 · PER WORKFLOW CLASSIFY JUDGE AUTO-EXECUTE REVIEW QUEUE “The kitchen tap in 4B is dripping again.” MAINTENANCE · ROUTINE AUTO-EXECUTE ≥ 0.85 0.91 EXECUTED “We're ending the lease on 1 June. What happens to the deposit?” LEASE · ELEVATED 0.62 HELD A THRESHOLD OF 1.00 MEANS NEVER AUTO-EXECUTE. LEGAL WORKFLOWS SHIP THAT WAY.
FlowBase routing two emails to two different outcomes A decision trace: classify, then judge, then a branch to either auto-execute or the review queue. Below it, two emails. A routine maintenance request scores 0.91 against the 0.85 threshold and is auto-executed. A lease-termination email scores 0.62, stops short of the threshold, and is held in the review queue with its full context. DECISION_TRACE THRESHOLD 0.85 · PER WORKFLOW CLASSIFY JUDGE AUTO-EXECUTE REVIEW QUEUE “The kitchen tap in 4B is dripping again.” MAINTENANCE · ROUTINE AUTO-EXECUTE ≥ 0.85 0.91 EXECUTED “We're ending the lease on 1 June. What happens to the deposit?” LEASE · ELEVATED 0.62 HELD A THRESHOLD OF 1.00 MEANS NEVER AUTO-EXECUTE. LEGAL WORKFLOWS SHIP THAT WAY.
READS The 0.85 threshold and the "1.00 means never" sentinel are FlowBase’s own defaults. The two emails are illustrative.

One email in. Judged, routed, done.

WatchBase The Observatory IN_BUILD
WatchBase narrating one change and filtering another A radar sweeping eight watched pages. The first pass finds no change. The second finds a pricing change scoring 0.82 against a 0.40 threshold, which blooms and is narrated. The third finds a navigation-copy change scoring 0.31, which stays dim and is filtered. The footer reads eight targets, forty-seven filtered, three narrated, four cents of model cost. SWEEP · 8 TARGETS DETECTION_LOG PASS_01 · 8 TARGETS NO CHANGE PASS_02 · competitor/pricing NARRATE ≥ 0.40 0.82 NARRATED PASS_03 · competitor/nav-copy 0.31 BELOW THRESHOLD · FILTERED Their compare page dropped the Alpine Shell 18% overnight. 8 TARGETS · 47 FILTERED · 3 NARRATED · $0.04 AI COST
WatchBase narrating one change and filtering another A radar sweeping eight watched pages. The first pass finds no change. The second finds a pricing change scoring 0.82 against a 0.40 threshold, which blooms and is narrated. The third finds a navigation-copy change scoring 0.31, which stays dim and is filtered. The footer reads eight targets, forty-seven filtered, three narrated, four cents of model cost. SWEEP · 8 TARGETS DETECTION_LOG PASS_01 · 8 TARGETS NO CHANGE PASS_02 · competitor/pricing NARRATE ≥ 0.40 0.82 NARRATED PASS_03 · competitor/nav-copy 0.31 BELOW THRESHOLD · FILTERED Their compare page dropped the Alpine Shell 18% overnight. 8 TARGETS · 47 FILTERED 3 NARRATED · $0.04 AI COST
READS The 0.40 significance threshold is WatchBase’s default and is configurable per organization. The pages watched and the sentence narrated are illustrative.

The change you would have found out about on Thursday.

DocBase The Editorial Studio IN_BUILD
DocBase drafting five sections and regenerating one Five document sections drafting at once on a shared elapsed axis, each bar stopping where its specialist finished, and a quality dial climbing to 6.6 out of 10. The commercials section comes back under the judge's 7.0 bar and is regenerated on its own, so it ends last and longest; the dial then settles at 7.4. The run is stamped revision one of two, and the note says the pipeline accepts a document as-is after two revisions rather than looping. GENERATION_RUN SECTIONS DRAFT IN PARALLEL SCOPE DRAFTED APPROACH DRAFTED TIMELINE DRAFTED TEAM DRAFTED COMMERCIALS DRAFTED 7.0 7.4 / 10 JUDGE BAR 7.0 ELAPSED · ALL FIVE START TOGETHER COMMERCIALS SCORED 5.8 · REGENERATED ON ITS OWN · RE-SCORED 7.4 QUALITY GATE · PASS REVISION 1 OF 2 AFTER TWO REVISIONS IT ACCEPTS THE DOCUMENT AND TELLS YOU IT STOPPED.
DocBase drafting five sections and regenerating one Five document sections drafting at once on a shared elapsed axis, each bar stopping where its specialist finished, and a quality dial climbing to 6.6 out of 10. The commercials section comes back under the judge's 7.0 bar and is regenerated on its own, so it ends last and longest; the dial then settles at 7.4. The run is stamped revision one of two, and the note says the pipeline accepts a document as-is after two revisions rather than looping. GENERATION_RUN SECTIONS DRAFT IN PARALLEL SCOPE DRAFTED APPROACH DRAFTED TIMELINE DRAFTED TEAM DRAFTED COMMERCIALS DRAFTED ELAPSED · ALL FIVE START TOGETHER 7.0 7.4 / 10 JUDGE BAR 7.0 COMMERCIALS SCORED 5.8 · REGENERATED ON ITS OWN · RE-SCORED 7.4 QUALITY GATE · PASS REVISION 1 OF 2 AFTER TWO REVISIONS IT ACCEPTS THE DOCUMENT AND TELLS YOU IT STOPPED.
READS The two-revision cap and the 7.0 judge bar are DocBase’s own. The sections and the scores are illustrative.

Specialists draft in parallel. One judge decides.

CAPABILITY_MATRIX

What each instrument is for

Core is the product’s primary value. Supporting means it is used but is not the reason to buy it.

Capability AskBase FlowBase WatchBase DocBase
RAG / knowledge retrieval Core Not applicable Not applicable Supporting
Permission-aware access Core Not applicable Not applicable Not applicable
Live database query (text-to-SQL) Core Not applicable Not applicable Not applicable
Schema curation Core Not applicable Not applicable Not applicable
LLM classification / triage Supporting Core Not applicable Not applicable
Confidence-based routing Supporting Core Not applicable Not applicable
Workflow orchestration Not applicable Core Not applicable Not applicable
Web scraping Not applicable Not applicable Core Not applicable
ML significance scoring Not applicable Not applicable Core Not applicable
Semantic caching Supporting Supporting Core Not applicable
LLM narration / summarisation Supporting Not applicable Core Supporting
Multi-agent parallel generation Not applicable Not applicable Not applicable Core
Quality gates (LLM-as-judge) Not applicable Not applicable Not applicable Core
Structured document output Not applicable Not applicable Not applicable Core
Slack / Teams integration Core Supporting Core Not applicable
Email integration Not applicable Core Not applicable Not applicable
Vector search (pgvector) Core Not applicable Core Core
Usage analytics / reporting Supporting Supporting Supporting Supporting

Core Supporting

Not sure which one fits?

That is what the assessment is for. Thirty minutes, a fixed agenda, and a one-page map back within two business days. If nothing here fits your problem, we say so.

Book Free Assessment