Stephen Phillips How BASF puts knowledge at the centre of innovation
Jinfo Blog

28th January 2026

By Stephen Phillips

Abstract

A recent Jinfo Community session, with Dr Stephan Ballenweg of BASF, explored the development of “QKnows”, a global knowledge and information platform for 10,000 R&D researchers.

As a result of the session, Jinfo has produced a detailed report that gives an overview of the development of the QKnows platform, and eight actions for information leaders.

This blog gives some highlights from the report, including two actions, plus details of a forthcoming demo.

Item

The "QKnows" project at BASF is a world-class benchmark for positioning information at the heart of an organisation's innovation work.

By consolidating legacy applications, repositories, and third-party information products, BASF has created a single entry point for 400 million documents. This has created a powerful ecosystem for scientific discovery.

Here are two best-practice concepts from Jinfo’s new report that can reshape your own strategic roadmap:

Moving beyond the "time-saving" trap

Many business cases are built on the premise that we save our users time. However, BASF found that making the argument that researchers spend 25-30% of their day finding information was often met with indifference by their stakeholders.

Instead, BASF shifted the narrative to enabling novel capabilities.

They positioned the platform as a "sparring partner" for scientific brainstorming and idea validation. The value wasn't in finding a document faster; it was in the ability to conduct "deep research" across massive, disparate datasets – an activity that was previously impossible due to fragmented legacy systems and staff capacity.

Furthermore, stakeholders were motivated to mitigate the risk of repeating expensive R&D work, and recognised its potential as a competitive advantage.

Building trust with a "Walled Garden"

The rapid rise of GenAI presents a dilemma: how do we harness its power without compromising data integrity?

BASF's solution was to implement a "walled garden" – a highly-controlled, secure environment for both BASF's own proprietary content, and third-party sources of data and information.

This enabled the information team to curate the best content, and prioritise specific sources. This then immediately differentiated QKnows from generic commercial tools.

Crucially, they were able to influence the large language model not to synthesise answers when content is lacking. This minimised hallucinations, thus building the high level of trust in the platform required by the BASF global research community.

Strategic actions for information leaders

Here is a taste of two practical actions (from the eight in the report):

  1. Prioritise co-creation and radical transparency:

    Build your systems based on your community's articulated pain points, rather than preconceived technical solutions. If a search function operates as a "black box," users will reject it; show them the "why" behind the results.

  2. Embrace GenAI immediately:

    Start gaining experience now, by leveraging internal content where you already control the rights, while negotiating access to full-text external sources. Abstracts alone are not enough for meaningful AI-driven synthesis.

By learning from organisations like BASF, information leaders can build platforms that go beyond repositories, and become central to research and innovation strategy.

Read the report:

See a demo:

« Blog