Inside Knowledge, Outside Truth
Enterprise AI works best when it understands the company.
That is Glean's strength: enterprise context. Glean helps teams find the right internal knowledge, reuse approved content, reason across company systems, and keep work grounded in the permissions and processes of the business.
But many high-value workflows do not begin inside the company. They begin on the live web: an RFP appears on a procurement portal, a county updates a deadline, a new attachment is added to a vendor system, a competitor changes pricing, or a regulatory page changes after the internal knowledge base was last indexed.
TinyFish can be now connected as a live-web tool inside Glean workflows. A Glean agent can call TinyFish to search public sources, fetch current pages and documents, extract structured facts, and bring that evidence back into the enterprise context layer where Glean can apply company knowledge.
This is a new partner pattern: Glean provides enterprise context, knowledge, permissions, reasoning, and reusable content. TinyFish provides live web search, web fetching, structured extraction, browser execution, and source evidence.
Why the boundary matters
Most enterprise workflows are neither purely internal nor purely external. They cross the boundary.
A proposal team needs approved internal answers, but the RFP itself lives on an external portal. A sales team needs account notes, but also current public news. A compliance team needs internal policy, but also the latest regulatory page. A market team needs company strategy, but also competitor pages that changed this week.
The useful workflow is the combination:
- Find current external evidence.
- Bring it into the enterprise context layer.
- Compare it against what the company already knows.
- Draft or recommend the next step.
- Keep a human in control when judgment, approval, or external submission is required.
Example: public sector RFP response prep
Public sector RFPs are a strong starting point because they are valuable, time-sensitive, and operationally messy.
An opportunity may appear on a state procurement site, a county page, a portal search result, a PDF attachment, a spreadsheet, or an amendment posted after the original notice. The decision to pursue depends on internal context: product capabilities, certifications, prior answers, legal constraints, pricing guidance, partner coverage, public sector references, and bid/no-bid rules.
With TinyFish connected, a Glean workflow can support the process in a grounded way:
1. Find relevant opportunities
TinyFish searches public procurement sites and agency pages for newly posted opportunities, returning structured metadata such as agency, title, posted date, due date, source URL, attachments, and visible fit signals. Glean compares those opportunities against internal pursuit criteria, ICP, disqualifiers, strategic accounts, and past guidance. The output is not "go bid on everything." It is a better shortlist.
2. Pull the RFP material into context
TinyFish fetches the live source material and extracts key external facts: deadlines, submission instructions, required attachments, security requirements, mandatory certifications, pricing format, and amendment history. Glean interprets those requirements against internal knowledge such as approved responses, product docs, security language, case studies, customer references, and prior proposals.
3. Create a qualification brief
Glean helps compare the RFP against the company's capabilities, readiness, certifications, references, delivery model, and known constraints. TinyFish can enrich that brief with current public context: agency background, related procurements, visible incumbent vendors, recent amendments, or similar past awards.
4. Prepare a first draft
If the team decides to pursue, Glean can pull approved internal source material and draft a first-pass response. TinyFish helps keep the draft tied to the external requirements by pointing back to the original RFP sections, required attachments, submission fields, and portal instructions.
Where TinyFish fits
The goal is not to use every capability in every workflow. The goal is to give Glean agents a reliable path from internal context to current external reality.
Closing
Enterprise AI needs both memory and reach.
Glean provides the memory: enterprise context, permissions, knowledge, and reasoning over how the company works.
TinyFish provides reach: the ability to find, read, extract, and operate against live web systems when the workflow depends on current external information.
Together, they can make RFP response prep more useful without over-automating the parts that still require human review: find the opportunity, pull the source material, compare it to internal knowledge, draft the qualification brief, prepare the response, and keep the human in control.
That is a grounded place to start, and a strong foundation for what comes next.
