AI can't make your research for you
Anyone can now generate a plausible report. Only you did the fieldwork. Certant turns your research into a library that people and their agents can query.
Has any of the research looked at algal blooms in reservoirs under five metres deep?
Yes, three projects between 2011 and 2022. Two found blooms peaked once the water stayed above 25°C for a week. The 2022 study saw the same pattern at 23°C where nutrient levels were high.
The answer draws on three reports and gives the page in each.
Your members can't find your research
You commissioned the studies, ran the fieldwork and published the reports. Then they went into a library as PDFs, filed under project numbers and titles someone chose years ago.
A member with a question searches, gets forty results, opens two and gives up. Plenty of them then ask a public chatbot, which answers from whatever the web says about the topic.
Certant reads every report in the library, so the member gets the finding, the project it came from and the page it is on.
- Project 1047 Final Report.pdf2013
- FS-22 Fact sheet (superseded)2015
- Project 1132 Appendix C.pdf2018
- Workshop notes, Adelaide2019
- Project 1190 Summary.pdf2022
And 38 more results. None of them says which report answers the question.
Original data is worth more now
People have called data the new oil for twenty years, and the comparison now undersells it. Text, images and code cost almost nothing to produce, and more of what gets published each year was generated rather than found out.
Measurements still come from people: someone sampling a reservoir in the summer of 2014, running a ten-year trial, or surveying four hundred members about their costs. The conditions a 2014 study measured have since changed, so that study stays the only record of them.
The organisations that did that work hold what everyone else's AI depends on. Your archive is worth more now than it was five years ago, provided people and their software can get at what's in it.
Connected to the Certant MCP server · key: members, read-only
ask_knowledgebaseAnswers a question from your reports, with the page each part came fromsearch_knowledgebaseFinds passages by what they mean, whatever words the author usedfind_textFinds an exact phrase or project number across the libraryget_documentOpens a report the key is allowed to read
Your library becomes a set of tools that any MCP client can call, and every call runs under the key you issued. See the MCP server
"Proprietary data or content: data nobody else has, or unique content you can create"
"The problem is if nobody creates new content anymore and just asks their AI every question, the AI has nothing to train on anymore"
See what sits under the chatbot
Most organisations start with a chatbot, and it is the smallest part of what Certant builds. The knowledgebase and the map of your research underneath it are what make the work usable by other software.
Members ask in their own words
Put a chatbot on your website or member portal. It answers from your reports, shows the page each sentence came from, and says so when the library doesn't cover the question.
Because every answer names the report, your organisation is credited each time someone uses it.
Is there guidance on disinfection by-products for small rural systems?
Yes. The 2020 guidance note recommends quarterly sampling for systems serving fewer than 1,000 people, and monthly sampling for a year after any change of water source.
Line up findings across studies
Ask which treatment methods were tested for a contaminant and you get a table back, with a row for each study, its result and the page the result is on.
Certant reads the tables and appendices as well as the text, so results buried on page 40 come back with the rest.
| Method | Result, and where it says so |
|---|---|
| Granular activated carbon | Above 90% for long-chain PFAS Project 1158 · table 6 · page 22 |
| Ion exchange | Above 95%, resin replaced every eight months Project 1201 · page 41 |
| Reverse osmosis | Above 99%, leaving a concentrate to dispose of Project 1201 · page 44 |
| Ozonation | No useful removal Project 1158 · page 30 |
If a study doesn't report a figure, its row says so.
See how your research connects
Certant builds a knowledge graph of the library: projects, sites, contaminants, methods, authors and the findings that link them. Each topic gets a page written from everything that mentions it.
That page shows where studies disagree and what nobody has studied yet, which helps when you plan what to commission next.
Let other people's agents ask too
Connect through the Certant MCP server, or publish an agent as an API, and a member's own AI tools can query your library directly. Each call runs under a key you issue and can switch off.
Your own staff get the same tools through the AI FDE, the assistant built into Certant. It uses the MCP server's tools from inside the platform, so on a sovereign deployment it runs entirely on your own hardware.
Ion exchange removed more than 95% of PFAS in the 2021 pilot. The resin needed replacing every eight months.
Decide who gets to ask
Open the library to the public, keep it as a member benefit, or sell access to it. Certant controls who can ask and what each of them can reach, and what you charge is up to you.
Members, staff, funders and the public can each reach a different part of the library.
Each partner or agent gets its own key, limited to what it may read, and you can revoke it at any time.
Each answer names the report and opens it at the page, so your work is cited wherever the answer travels.
Certant Cloud, your own cloud account, or fully air-gapped on your own hardware.
Questions and answers are logged, and Certant keeps a list of the questions your library couldn't answer.
File drops, scheduled exports and the API add new work as you publish it, and Certant tells you when a newer report replaces an older one.
Built for organisations that publish knowledge
Research bodies
Decades of funded projects, final reports and fact sheets that members paid for and rarely find.
Industry associations
Submissions, member surveys and the advice your staff give out every day.
Standards and guidance publishers
Codes and guidance that change over time, where people need the version in force on a given date.
Peak bodies
Policy positions and evidence that go back further than anyone on staff.
Before you bring us your library
Our library is decades of PDFs, and some are scans. Will that work?
Yes. Certant reads scanned pages, tables and complex layouts, and picks a reader for each kind of file. You can change which one it uses.
Can members check where an answer came from?
Yes. Every answer shows the report it came from. Click the citation and the document opens at that page.
Can some reports stay restricted?
Yes. You decide which people, teams and keys can reach each knowledgebase, so restricted work can sit apart from the public library.
How do our members' own agents connect?
Through the Certant MCP server, or through an agent you publish as an API. Each connection uses a key you issue. The MCP server is included on the Business and Enterprise plans and on every dedicated and sovereign deployment.
Does the library have to leave our building?
No. Certant runs in Certant Cloud, in your own cloud account, or fully air-gapped on your own hardware.
See what your library can answer
Book a demo on a sample of your own reports and see what your members could ask them.
Not since 2019. The last study, Project 1164, covered two sites in north Queensland and recommended a follow-up that was never commissioned.
When the library has a gap, the answer says so.