Dear ATIO Team,
Thank you for providing the opportunity to comment on the ATIO KB. Below, please find our submission on behalf of GIZ’s Fund for the Promotion of Innovation in Agriculture (i4Ag).
FSN Consultation
Contribute to shaping the design of the Agrifood System Technologies & Innovations Outlook (ATIO) Knowledge Base
Submission by Fabiana Woywod and Till Rockenbauch, Fund for the Promotion of Agricultural Innovations (i4Ag), Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH
1) Given the description of the ATIO KB, how do you think it can help you and users like you? Describe one or more specific use cases that you wish the KB would address, like “I imagine I would be able to find innovative products that support farmers with access to credit and insurance specifically for one country, and I would be able to see information on their readiness and how they fare against adoptability criteria” or “I would like to use statistics to show a correlation between level of inclusivity / co-design of the solutions and their levels of adoption”.
As an implementing agency GIZ, has vast experience in setting up innovation partnerships for the promotion of agricultural innovations. Potential use cases include scouting (innovations and partners) and ideation (project design). A particular use case could be support to the development of scaling initiatives / strategies (e.g. design of context-specific innovation bundles).
As a means to this end, we would ATIO expect to provide (additional) information on:
- what innovations are applicable / promising in a specific context (e.g. agroecological setting, national / regional or institutional context)?
- how innovations could be bundled for synergies?
- relevant/common challenges/hurdles to consider/overcome (differentiated by intended scaling context)
- blueprints/roadmaps for scaling and dissemination
In addition, ATIO would be of interest as an outlet to share/disseminate proven GIZ innovations.
2) What do you make of concepts like policy innovation and social innovation? Can you think of examples? Is it useful for you to be able to find such content? In which form do you expect to find them? How would you use them?
Social/institutional innovations (e.g. farmer self-help groups, machinery rings, etc.) are drivers of social change and empowerment and, hence, should be equally represented at the platform (besides technological innovations). In addition, policy innovations (approaches to inclusive decision/policy making and conducive regulatory frameworks) are prerequisite for successful scaling of innovation (bundles). The platform could feature/suggest policy innovations as enablers/facilitators of particular agricultural innovations.
3) How important is it to feature grassroots innovations? Looking at some records of grassroots innovations in the prototype, what would you like to see in the descriptions that you don’t see? Which dimension should we capture? What is most useful for grassroots use/application of innovations?
ATIO’s emphasis on grassroots innovations is appreciated, as the discourse around (scaling of) agricultural innovations (in many cases) is dominated by linear / top-down approaches. For innovations to be inclusive, it is prerequisite to strengthen grassroots actors’ capacity to innovate and to ensure their representation in scaling processes. Rather than classifying an innovation as grassroot/non-grassroot innovation it would be helpful to provide information on the suitability of innovations to foster inclusive scaling and/or recommendations on how to ensure inclusive scaling (e.g. by bundling technical with social innovations). In addition, the use of the platform by grassroot organizations should be ensured/promoted (e.g. by paying attention to language and easy applicability).
4) How do you think branded commercial products should be featured on the ATIO KB? Data sources of technology-related information often feature individual models of technologies (for instance, different models of solar-powered irrigation pumps). Should the ATIO KB feature models? What is the “innovation” unit you expect to find?
- Where possible, we suggest to avoid branding (in the presentation of the innovation).
- Innovations with an open access licensing could be highlighted
- You might consider including an additional section/category “copyright”
5) Here are two of the main taxonomies used in the prototype: types of innovations and use cases. Considering that there is no agreed standard for these categorizations, and that we are aligning them to those used in similar projects, are these “good enough” to start? Which major problems do you see? Please suggest changes or volunteer to help us improve them in the next months. Other taxonomies are here.
- The “use case” taxonomy helps to compare Innovations when listed in the result gallery. Could be placed on top of thumbnail.
- Types and use cases are very detailed, which is good to match keyword search – however feels overwhelming and time-consuming to go through. Lower scaling readiness stages (1-3) most likely not relevant for non-scientific users.
- Use of LLM for search related suggestions could help.
- Also, a digital dashboard to compare innovations could save time and increase user-friendliness, based on key metrics/scores (e.g. scaling potential, ease of use, cost-benefit ratio, etc. could further improve comparability
- Tools for identifying synergies and compiling innovation bundles could help building context-specific solutions
6) We are developing a chatbot-like search capability. Do you prefer the classic filter-based search or the chatbot search? Or the possibility of choosing either? Tell us how we can improve the search experience.
Chatbot search results often lack transparency. We would like to see an overview of the results to better control the search and tailor the assistance to the needs.
7) We use Artificial Intelligence (AI) to enrich and automatically categorize the records: you will see an AI stamp at the end of descriptions that have been generated by AI: how good is the text generated? Is AI enriching the records in a meaningful way?
- Description of innovations seems good but more conceptual clarity might be needed with regard to what actually constitutes particular innovations. Some “innovations” listed are resembling a range / bundle of (not necessarily well-defined) innovations.
- Subtitles or graphics would help for skim reading.
- AI-Reasoning reads consistent at a first glance; however, doubts remain due to vague/ambitious wording. Is AI checking the plausibility of the data base? How does it deal with vague/patchy data? For example, is AI able to distinguish between broad impacts claims (e.g. “the innovation improves income “) from reported impact (e.g. “the innovation has improved the income of x users by x percent”)? In addition, are there human plausibility checks (and if yes, how do humans deal with patchy data)?
We appreciate your efforts in drafting the ATIO KB, which we consider a promising tool with great potential for informed decision-making towards scaling the impact of agricultural innovations. We would be interested to stay updated on latest developments and would be glad (where possible and feasible) to accompany the conceptual refinement and implementation of the ATIO platform.
Dr. Till Rockenbauch