Attention: You are using an outdated browser, device or you do not have the latest version of JavaScript downloaded and so this website may not work as expected. Please download the latest software or switch device to avoid further issues.
The International Geosynthetics Society (IGS) is seeking a supplier to create a service that lets anyone, member or not, ask questions of the IGS Digital Library in plain language and receive trustworthy answers drawn only from its 10,000+ technical documents, with every answer referenced back to the source papers.The IGS Digital Library is a unique resource of more than 10,000 documents spanning over 40 years of technical work, including peer-reviewed conference papers, journal content and general publications. While the Library is highly valued by the IGS membership, we want to make it more accessible, in support of our aim to promote the responsible use of geosynthetics worldwide. We envisage a new AI-powered tool that will enable anyone, anywhere, to ask aquestion in plain language and receive a useful, trustworthy answer drawn from the Library, turning the Library from an archive into a working tool that members reach for regularly. We are seeking proposals from potential suppliers by Thursday 10 September 2026. Full details and how to apply are in the Request for Proposals document outlined below.
AN AI DIGITAL ASSISTANT FOR THE IGS DIGITAL LIBRARY
The International Geosynthetics Society (IGS) is seeking a supplier to create a service that lets anyone, member or not, ask questions of the IGS Digital Library in plain language and receive trustworthy answers drawn only from its 10,000+ technical documents, with every answer referenced back to the source papers.
Proposals due: Thursday 10 September 2026
The IGS is a non-profit learned society that promotes the responsible use of geosynthetics worldwide. Our members include practising engineers, consultants, designers, academics, researchers, students and corporate members, organized through around 45 national chapters.
We want the IGS to be sustainable as an organization, representative of the broad community it serves, trusted as a source of well-founded technical knowledge, and influential in the discussions where that knowledge matters, including policy. This project is intended to serve these aims.
The IGS Digital Library holds more than 10,000 documents spanning over 40 years of technical work, including peer-reviewed conference papers, journal content, general publications, images and video. It is possibly the most complete record of geosynthetics research anywhere. For the purpose of this document, the “Library“ refers to the combined contents of the IGS Digital Library and the recently-published IGS Geosynthetics Handbook.
The Library is underused. Search relies on titles, authors, conferences and keywords, and members tell us it is hard to find anything. Awareness of the Library among our members is not as high as we want. The knowledge is there; people cannot access it easily. We want anyone, anywhere, to be able to ask a question in plain language and receive a useful, trustworthy answer drawn from the Library, with references that take them to the source documents. Done well, this turns the Library from an archive into a working tool that members reach for regularly, and a reason to join the society and stay.
A practising engineer advising a client can find relevant case studies and evidence in minutes, cite exactly where each point comes from, and answer the follow-up question their client asks next.
A trainer on one of our professional development courses can answer niche questions from the classroom without spending time hunting for a half-remembered paper.
A student or early-career researcher can find the peer-reviewed work that matters without already knowing the terminology used in a paper written in 1994.
Anyone researching a contested topic sees the range of published positions and can read the papers behind each one. Where our authors disagree, users should discover that disagreement, rather than have an AI-sanitised answer or single conclusion.
Over time, non-specialists (a client weighing a design proposal, a policymaker, a procurement officer wary of unfamiliar technology) can understand what four decades of published evidence says.
The Library becomes a service member's value and use often, which strengthens the case for joining and staying.
The IGS Digital Library remains the reference portal of choice for geosynthetics knowledge, for members and for the wider engineering world.
Common misconceptions about geosynthetics can be easily and quickly met with relevant published evidence.
Our reputation for reliability is protected. This outranks every feature; an assistant that gives wrong or invented answers under the IGS name would cost us more than having no assistant at all.
The points below describe what the service must achieve rather than how. The how is yours to propose, and section 7 asks you to explain and defend the choices you make.
Answer only from our Library. Every answer must be built from material in our collection, with nothing drawn from the wider internet and nothing invented. If the library does not contain an answer, the service must say so plainly rather than guess.
Reference everything. Users must be able to see which documents and passages an answer draws on, and get to those citations easily. A claim that cannot be traced to a source in the library should not appear.
Show disagreement honestly. Forty years of research contains contradictions; authors reach different conclusions, and some papers carry commercial interests while others are independent academic work. Where published views differ, the service must present the range with sources, never a single settled answer.
Never advise. Our members use the Library to advise their own clients. The service exists to get them to the evidence quickly; it must not offer engineering advice, recommendations or designs for a user's specific project, however the question is phrased.
Never endorse. Answers must not suggest a preference for or a recommendation of commercial products or brands.
Cope with our varied documents. Much of the value in our papers sits in equations, tables, technical diagrams and figures, often in dense two-column layouts, some of it decades old. Be specific about how much of this content your approach can read and use, and what would be lost. We would rather understand the limits now than discover them after launch.
Work across languages. Most of the collection is in English; some is in other languages, and users will ask questions in their own language. Tell us what is realistic and what challenges we may face.
Respect membership boundaries. A subset of the Library (some journal content) is available to members only and is published with independent publishers. While the service should still draw from the journals, proposals should explain how they would handle a collection where most material is open and some is restricted, including pointing non-members towards membership rather than exposing restricted text.
Be easy to update. We add a few hundred papers in bulk every year or two after each major conference, plus occasional individual documents, uploaded by non-technical staff. Adding material must be routine and must not repeat the cost or effort of the initial preparation.
Serve a global audience. Our users are spread across every timezone and usage will spike around our conferences and announcements. The service must stay available and reasonably quick under varied load.
Run without a technical team. We do not employ software or AI specialists and do not plan to. Whatever attention the service needs day-to-day must fit within the hours of a small staff team, and your proposal should say what that attention looks like.
The photographs and videos in the Library. Start with the PDF documents; multimedia may follow as a later phase.
More than 10,000 PDF documents. The core is peer-reviewed conference proceedings, often uploaded in batches after each conference. Some journals are published with independent partners and sit behind membership login. Metadata currently covers title, author, conference, year, country and keywords, and its consistency varies. The PDF documents can be rich in equations, tables, images and technical diagrams.
The current Digital Library runs on a WordPress website hosted by a third party. Our staff publish content to it; nobody on staff is a developer.
Modest, and imprecisely measured. Web analytics suggest engaged visitors in the low tens of thousands a year (a figure we treat with caution), and around 1,500 users downloaded roughly 2,700 documents over the last year, with strong access from Asia and the Americas.
The Library has never been promoted seriously. We expect usage to grow once it is easier to use and we begin telling people about it, and proposals should be positioned with that growth in mind.
Please structure your response to cover the eight areas below. Keep the document under 15 pages and written in plain English, with technical depth in appendices; independent technical review may be available to us while we assess proposals. We will treat unexplained jargon and undefended decisions as weaknesses, not as signs of expertise.
About you. Who you are, the team who would do the work, two or three examples of comparable projects, and references we may contact.
Your approach. How our documents would be read, prepared, stored and searched; how answers would be produced and referenced; and how the boundaries in section 4 would be enforced. Where you have made a choice on our behalf, name the alternatives you considered and explain why your choice suits our collection in particular.
Evidence. How will we know the service is doing the right thing at launch, and beyond? Tell us how you would test it before launch, including the part our subject experts should play, and how you would keep checking once it is live, making costs, and expectations from our internal resource clear.
Delivery plan. We favor staged delivery, with points where we can judge results before committing further. Propose the phasing you think is right, with what we receive and what we pay at each stage, and where we could choose to stop.
Costs. One-off costs, including preparing the platform, as well as ongoing operational and content update costs, with your assumptions stated. Fixed prices are welcome where the assumptions behind them are stated.
Ownership and exit. What the IGS would own, including the prepared and processed forms of our content; what happens if we part ways; and what moving to another supplier would cost and involve.
Hosting, security and privacy. Where our content and our users' questions would be held; your position under GDPR / DPA; a commitment in writing that our content and our users' questions will not be used to train anything beyond our own service; and how you would prevent the service being used to copy the Library wholesale.
Support and workload expectations. What you look after, what falls to us, a realistic estimate of the staff time we should expect to spend, and what happens when something breaks.
Questions about this document should be sent in writing to the contact below. Where a question is relevant to all bidders, we may share the question and our answer with everyone. We may shortlist, put further questions to shortlisted suppliers, and ask them to demonstrate their approach against sample documents we provide.
We are not bound to accept the lowest-priced proposal, or any proposal. We will treat your proposal in confidence, and we ask that you treat this document and anything we share with you the same way.
Contact: John Kraus, Executive Director, john.kraus@geosyntheticssociety.org
Any initial questions must be sent to us by email no later than Monday 31 August 2026. We will respond to all questions within a week of this date.
Proposals should be submitted by Thursday 10 September 2026 as a PDF, by email to the address above.
Once received the IGS will review and shortlist proposals. We will aim to notify shortlisted bidders by Friday 30 October, along with details of final selection steps.
IGS, August 2026
