Discovery Is Becoming an AI-Native Workflow: Why the KGL–Cashmere Partnership Matters

Across Hong’s recent Scholarly Kitchen essays (1, 2, 3), he has come back to one core idea: discovery in scholarly publishing is no longer just about helping someone find a document. It is increasingly about helping them complete a task.

Researchers, clinicians, students, and professionals still value the source record. They still need the article, the chapter, the dataset, and the evidence trail. But the way many people now begin is different. They want faster orientation, a clearer answer, and enough confidence to know where to go deeper. They want the path into the evidence to be shorter.

Discovery as a Workflow, Not a Feature

That may sound like a subtle shift, but it has major implications for publishing platforms. Discovery can no longer be treated as a search box sitting on top of a content repository. It is becoming a workflow layer that helps users move from question to understanding to evidence with less friction.

In that sense, the future of discovery is not just better search. It is a combination of retrieval, explanation, trust, and task support.

That is also why AI-powered discovery is quickly becoming table stakes. Much like keyword search became a basic requirement in an earlier era, natural-language Q&A and AI-assisted navigation are becoming expected.

This may not always create a new revenue stream on its own, but it will increasingly shape competitiveness, retention, and renewals. If a platform cannot support this kind of experience, customers will start to see it as behind the market.

Integrated Infrastructure, Not More Features

At the same time, I do not think the answer is simply to add more AI features. The industry does not need more disconnected experiments and duplicated solutions. It needs integrated infrastructure.

Precision search, metadata filtering, known-item lookup, and direct access to the authoritative record still matter. The opportunity is to combine conversational discovery with strong retrieval, rights, provenance, and governance underneath. The goal is not to replace search. It is to modernize discovery without losing trust.

Why the KGL–Cashmere Collaboration Matters

This is where the KGL–Cashmere collaboration becomes important.

One reason is practical. Most AI discovery systems today are variations of the same retrieval-augmented pattern, but the underlying technology keeps changing. Indexing methods evolve. Retrieval strategies improve. Ranking, generation, and reflection layers continue to shift.

A platform provider should not have to rebuild its roadmap every time the discovery stack changes. That complexity is better handled by a specialist partner whose job is to keep pace with the technology.

That creates a clearer division of responsibility. Publishers and platform providers should focus on what they are uniquely responsible for: maintaining a trusted single source of truth.

That means content quality, metadata integrity, provenance, entitlements, rights clarity, and the authoritative version of record. A specialist partner can then provide the technical and commercial rails that make that content usable in modern AI environments.

Understanding “Local LLM” Discovery

This also helps explain the growing interest in “local LLM” discovery. In my view, local LLM is not really a separate AI category. It is a deployment and governance choice.

The core pattern is still retrieval plus generation. What changes is the corpus being searched, where the system runs, and who controls the content, logs, and user interactions.

For many society and independent publishers, that matters a great deal. They want discovery grounded in their own content, they want engagement to stay on-platform, and they want stronger control over trust, access, and data handling.

Enabling Dual-Channel Discovery

That is why this partnership matters beyond licensing alone. It enables KGL to support a governed layer for publisher-controlled discovery now, while also creating the option for broader AI distribution later.

That flexibility is important because the market is clearly becoming dual-channel. More users will begin in broad AI environments, just as they once began in search engines or Google Scholar.

But that does not make the publisher platform less important. It changes its role. The publisher site becomes the authoritative source, the entitlement gate, the canonical record, and the engagement hub.

Leveling the Playing Field for Smaller Publishers

This matters even more for small and mid-sized publishers. Large publishers may be able to build internal AI capabilities or negotiate direct relationships with major AI platforms. Most smaller organizations cannot.

They still need a credible path into AI discovery and licensing, but without having to build a bespoke infrastructure stack themselves. An independent partner model helps level that playing field by providing shared capabilities for indexing, authentication, licensing, delivery, tracking, and monetization.

How KGL Adds Value

That is where KGL is in a strong position to add value. KGL can help publishers do three things well:

• Maintain a trusted content, metadata, and rights layer as the source of truth for both human and AI use
• Enable governed AI discovery across publisher platforms, workflows, and external channels while keeping publishers in control
• Help customers accelerate dissemination, reporting, and monetization by connecting content to AI applications from production onward, without waiting for hosting deployment or requiring them to build their own AI infrastructure

Trust Infrastructure Becomes Critical

More broadly, if AI is becoming a new kind of reader, then the trust framework becomes as important as interface design.

What matters is not just whether an assistant can generate an answer, but whether that answer is faithful, attributable, governed, and measurable. That requires richer metadata, machine-readable rights, stronger provenance signals, and new ways of capturing engagement beyond the traditional page view.

The Convergence of Discovery, Rights, and Monetization

For years, publishing treated discovery, rights, and monetization as related but separate functions. AI is collapsing those boundaries.

The same infrastructure that improves discovery quality can also improve governance. The same signals that protect IP can also make content more usable in machine-mediated environments.

That is why we see the KGL–Cashmere collaboration as more than a feature partnership. It is a step toward the infrastructure scholarly publishing now needs.

The Real Question for the Industry

The real question is no longer just, “Can users search our content?” It is, “Can trusted AI systems discover, use, and direct value back to our content in ways that preserve control, attribution, and trust?”

That is the challenge in front of the industry now. And that is why this collaboration matters.

KnowledgeWorks Global Ltd. (KGL) is the industry leader in editorial, production, online hosting, and transformative services for every stage of the content lifecycle. We are your source for society services, research integrity, intelligent automation, digital delivery, and more. Email us at info@kwglobal.com.