Evidence-Governed AI
Stop assembling reports.
Stop assembling reports.
Start answering questions.
The answers your board and funders ask for already exist inside your organisation — scattered across the systems, records, reports and project files you already run on. Cylentis turns that evidence into cited answers, board-ready reporting and honest gap maps. Findable. Defensible. Honest.
Built for CRCs, universities, research centres, government programs, not-for-profits & evidence-heavy commercial teams.
10 wks
Reclaimed per year
45 sec
Portfolio question answered
100%
Evidence-backed or flagged
Live
In production since 2025
Sound Familiar?
All that evidence. No clear way to see it.
These are the problems we hear from the leaders we talk to — the ones accountable for what gets reported.
“
I spend two and a half weeks every quarter pulling together board packs from five different systems.
“
A funder asks a question and I have to call three people and wait a week before I can answer.
“
We've got millions in funded work and I can't tell you where to put the next $500K.
“
I don't know which of our 40 projects are on track. I don't even know how many technology buckets we have.
“
Someone on the board asked about our translation pipeline and I had to wing it.
“
We tried Copilot. It made things up. We can't put hallucinated numbers in a report we have to defend.
What Changes
From weeks of assembly to seconds of answers.
Same evidence. Same systems. Completely different capability.
Before Cylentis
Weeks every quarter assembling board and funder packs by hand
Can't answer board questions without calling multiple people
No view across technologies, partners, programs or project status
Copilot and ChatGPT hallucinate in reporting and governance contexts
Spreadsheets and slide decks as the system of record
→
After Cylentis
Board packs generated in minutes with full evidence chains
Any portfolio question answered in 45 seconds with citations
Complete view of every project, technology and partner
The system tells you when it doesn't have enough evidence
One governed evidence base underneath the tools you already use
The Difference
If the evidence isn't there, we say so.
Every AI tool tells you what it found. We're the only one that formally tells you what it can't answer — and exactly what evidence would change that. That's not a limitation. That's governance.
query › What is the commercialisation readiness of Project SC-027?
⚠ PARTIAL ABSTENTION
Partner engagement and field trial evidence support moderate readiness (3 sources, directional confidence).
However, no IP protection filings or market validation data were found in the corpus.
→ Adding IP status and market analysis would upgrade this from Directional to Verified.
Sources: Progress Report Q3 2025, Partner Engagement Log, Field Trial Summary Newcastle
In Production
Not a prototype. Not a pitch deck.
Deployed and running with a live Cooperative Research Centre since September 2025.
48Projects governed
93Partners tracked
29Technologies mapped
503People indexed
Evidence ingested daily from the organisation's own systems — document libraries, CRMs, project records and more. Board packs, signal watches, technology readiness, case studies and portfolio queries, all from one governed evidence base.
Board packs: 2.5 weeks → minutes · Portfolio queries: days → 45 seconds · Hallucinations: zero
See It In Action
What portfolio leaders actually see.
Live governance surfaces from a deployed portfolio. Real structure. Real intelligence. Real evidence chains.
C.
🏠 Today
📊 Portfolio
📄 Cases
📋 Packs
✨ Ask
🔍 Intel
📈 Brief
🛡 Evidence
Needs attention now
What matters right now across your portfolio.
4 RED projects need rescue-or-exit decisions
Limited runway before wind-up. Each needs a clear decision: commit to a demonstrator pathway or prepare to close.
What to do: Run a 30-day rescue triage covering final output, partner commitment, and cost-to-complete.
CRITICALGREEN projects with hidden translation risk
Several GREEN projects carry unresolved Finance, Partnership, or IP flags. Likely to finish technically but may still miss adoption.
HIGHThe CEO opens this Monday morning. Every risk flagged before it reaches the board.
C.
🏠 Today
📊 Portfolio
📄 Cases
📋 Packs
✨ Ask
🔍 Intel
📈 Brief
🛡 Evidence
Portfolio health
48 active projects · Signal Watch on every entity
Green
41
Amber
6
Red
1
Total
48
Signal Watch WATCH
Research capability is advancing, but delivery pressure is now visible.
Every project scored. Signal Watch and three governance reads on every entity.
C.
🏠 Today
📊 Portfolio
📄 Cases
📋 Packs
✨ Ask
🔍 Intel
📈 Brief
🛡 Evidence
Ask anything
Natural language queries · governed evidence base
What are the main risks across the portfolio?
Ask
DIRECTIONAL20 records · 43 evidence items · GCA-v1
Main risks across the portfolio
Portfolio risks cluster into five themes: partner instability 14, delivery delays 9, commercialisation risk 21, budget timing 24, and research execution risk 27. The biggest shared risk is projects losing momentum because industry partners are disengaged. 35
Any question. 45 seconds. Every claim cited. Confidence scored.
C.
🏠 Today
📊 Portfolio
📄 Cases
📋 Packs
✨ Ask
🔍 Intel
📈 Brief
🛡 Evidence
Evidence posture
What you can and cannot defend this week.
7 portfolio questions need qualification before board or funder reporting.
Need qualification
7
Follow up
48
Fixable now
130
Closed
0
AI executive lens
130 evidence gaps across 48 projects. Main pressure: portfolio freshness, document promotion, and evidence depth.
Board-facing questions need qualification
Freshness issues may distort status
Evidence depth thin in multiple projects
130 evidence gaps ranked by impact. What your portfolio can't defend — yet.
Built For
Made for evidence-rich organisations.
If your answers have to stand up to a board, a funder, or an auditor, this is built for you.
CRCs & Research Centres
You report to funders, manage dozens of projects across many partners, and spend weeks every quarter on reporting that should take minutes.
Universities & Research Offices
You oversee hundreds of research streams and can't get a clear picture of where the strongest translation opportunities are without asking ten people.
Government Programs & Not-for-Profits
You need to know which initiatives are on track, where outcomes are real, and where the next dollar should go — backed by evidence, not instinct.
The same governed evidence layer works anywhere decisions have to be defended with evidence — including evidence-heavy commercial operations.
How It Works
Six weeks to set up. Zero disruption.
Your evidence stays where it is. We structure it underneath. No new tools for your team to learn.
Week 1–2
Connect
We connect to the systems you already run on — document libraries, CRMs, project tools, finance records, wherever your evidence lives. Nothing moves.
Week 3–4
Structure
We build your governed evidence base — every project, partner, technology and milestone mapped with provenance and confidence.
Week 5–6
Deliver
Your portfolio goes live. Board packs in minutes. Any question answered. Gaps flagged. Intelligence your board can actually defend.
Ongoing
Compound
Every question asked makes the system smarter. Each quarter we deliver new intelligence surfaces based on what your organisation actually needs.
Cylentis does not replace your experts.
Your people create the work. Partners drive delivery. Funding enables impact. Cylentis is the layer that makes the evidence visible, governable and actionable — so the people making decisions can actually defend them.
“Cylentis took the time to deeply understand our requirements along with the environment and systems that we operate within, and translated that into an AI-enabled interrogation tool that allows us to efficiently analyse large volumes of qualitative project data. The quality of the solution has exceeded our expectations, delivering clear, actionable insights that are already strengthening how we monitor and manage our portfolio.”
Clare Tubolets · CEO, SmartCrete CRC
Who Built This
Aaron Bost
Founder, Cylentis AII built Cylentis because I watched a CEO spend two and a half weeks every quarter pulling together board packs from five different systems. The work was excellent. The governance infrastructure wasn't. So I built the governed evidence layer that sits underneath — connecting the document libraries, CRMs, project systems and records an organisation already runs on into one source of truth. Every answer is cited. Every gap is flagged. If the system doesn't have the evidence, it says so. Based in Sydney. One product. One problem. Doing it properly.
FAQ
Questions we get asked.
Straight answers to what boards, funders and reporting teams ask before they commit.
What is evidence-governed AI?
It's AI that only answers as far as your authoritative evidence allows — citing every claim, scoring its own confidence, and abstaining when the evidence isn't there. The opposite of a tool that sounds confident regardless of support.
How is this different from ChatGPT or Microsoft Copilot?
General assistants generate fluent text and fill gaps with plausible invention. Cylentis only reports what the governed evidence supports, cites every claim, scores confidence, and formally flags what it can't defend yet.
How is it different from RAG?
Retrieval makes an answer relevant; it doesn't make it defensible. Cylentis uses retrieval to find evidence, then adds source authority, citations, confidence and abstention — the controls that make an answer safe for a board or funder.
Which systems can Cylentis connect to?
The ones you already run on — document libraries, CRMs, project and delivery tools, finance systems, spreadsheets, forms, portals and bespoke internal systems. Cylentis isn't tied to any single platform, and nothing has to move.
What happens when the evidence isn't there?
It says so. Cylentis abstains — fully or partially — and tells you exactly what evidence would let it answer with confidence, instead of inventing a response.
What is an evidence-gap map?
A live view of what your organisation can defend, what needs qualification before external reporting, and the specific evidence that would close each gap.
How fast is an answer?
Around 45 seconds for a portfolio-wide question, with citations and a confidence level attached.
Who is Cylentis for?
Evidence-rich organisations whose answers must stand up to scrutiny — CRCs, universities and research centres, government programs, not-for-profits, and evidence-heavy commercial teams.
How long does setup take?
About six weeks, with no disruption. Your evidence stays where it is and your team keeps working in the tools they already use.
Is our data safe? Does anything move?
Nothing moves. Cylentis structures a governed evidence layer over the systems you already run on; your records stay in place.
Your organisation already has the answers.
Book a 20-minute demo. Bring a question about your portfolio. We'll answer it live — or show you exactly why we can't.
© 2026 Cylentis AI — Sydney, Australia · aaron@cylentis.com