Spekit, Glean and the End of Looking Stuff Up at Work

Somewhere in your company, right now, the answer exists. The updated pricing. The security questionnaire from last quarter. The one Slack message where someone finally explained how the renewal process works. It is all there. The problem was never that the knowledge went missing. The problem is that finding it costs a person twenty minutes and their train of thought, and by the time they do, they have opened six tabs and forgotten what they were doing.

McKinsey has put a number on this. Knowledge workers spend roughly a fifth of the workweek - close to one full day - searching for and gathering information. It is the office's quietest tax. Nobody wrote it on a job description. Nobody celebrates paying it. And two startups, coming at it from opposite directions, have decided that this tax is the entire business.

One is Spekit, out of Denver, which puts the answer where you already are. The other is Glean, a company now valued at $7.2 billion, which indexes everything you own and lets you ask it a question like it is a person. They rarely bump into each other in a sales cycle. They are chasing the same ghost.

01 / THE PROBLEMA hoarding problem, not a storage problem

For a decade, software sold companies on storing knowledge. Wikis. Content libraries. Learning platforms. Shared drives with folders inside folders inside folders. The pitch was always the same: capture what your people know before they leave. Companies bought it, dutifully, and ended up with more places to put things and no faster way to get them back.

That is the insight both of these companies are built on. A wiki nobody can search is worth about the same as no wiki at all. The value was never in the storing - it was in the retrieval, the last thirty seconds where a person needs one specific thing and either gets it or gives up. Most tools fail exactly there, and the failure is invisible, because "I couldn't find it so I guessed" does not show up in any dashboard.

The value was never in the storing. It was in the last thirty seconds, where a person needs one thing and either gets it or gives up.

02 / SPEKITBring the answer to the person

Spekit was founded in 2018 by Melanie Fellay and Zari Zahra. Its early bet, sharpened during the pandemic, was almost stubbornly narrow: a new sales rep should not have to leave the deal to learn how to close it. Instead of sending someone off to read a wiki, Spekit surfaces the relevant snippet - a playbook, a piece of content, a bit of coaching - inside the tool the rep is already in. In Salesforce. In the browser. At the moment of need, not two clicks and one lost thought away.

The company has since repositioned around what it calls "Rep Acceleration," a governed layer of go-to-market knowledge that delivers deal context and trusted content directly where selling happens. The customer list has grown into names like Snowflake, PagerDuty and Outreach. In 2024 the company reported revenue of about $19 million, up from $9.4 million the year before, run by a team of roughly 100 people. In early 2025 Gartner named it a Visionary in the first Magic Quadrant it ever published for revenue enablement platforms.

Spekit's raise reflects the focus: about $62.8 million total, including a $45 million Series B led by Craft Ventures, with Renegade Partners, Bonfire Ventures, Felicis, Foundry Group and others along for it. It is not trying to answer every question in the building. It is trying to make sure the one person under the most pressure - the rep with a deal on the line - never has to stop and hunt.

There is a discipline in that narrowness that is easy to underrate. Enablement content rots faster than almost anything else in a company - a pricing change, a repackaged product, a new competitor, and half the playbook is quietly wrong. By living inside the workflow and governing what gets surfaced, Spekit is really selling trust as much as speed: the rep who pulls up a snippet mid-call needs to know it is current, not a screenshot from two quarters ago. Get that wrong and reps stop trusting the tool, which is the failure mode every enablement product dies from.

03 / GLEANIndex everything, then ask

Glean starts from the other end of the telescope. Founded in 2019 by Arvind Jain, who spent years working on search at Google, it does not pick a single tool to live inside. It connects to all of them - Google Workspace, Microsoft 365, Slack, Salesforce - builds a knowledge graph of who works on what and how documents relate, and then puts a plain-language box in front of the whole thing. You ask. It reads across the company and answers, with the sources attached.

The market has responded to that pitch with real money. Glean's annual recurring revenue passed $100 million for its fiscal year ending January 31, 2025, then roughly doubled to about $200 million over the following nine months. In June 2025 it raised a $150 million Series F led by Wellington Management, lifting its valuation to $7.2 billion - up from $4.6 billion just nine months earlier. Sequoia, General Catalyst, Altimeter, Khosla Ventures and Sapphire were all in.

Glean's whole value proposition fits in four words: ask, don't search.

What makes the number less strange than it looks is how unglamorous the job is. Glean is not trying to write your novel or replace your analysts. It answers "where is that thing" for the entire workforce, and it turns out companies will pay a great deal to stop losing a day a week to that question. The generative-AI wave gave the interface a boost - you type a sentence instead of guessing keywords - but the underlying product is retrieval, done well, at company scale.

The harder part is the part buyers don't see in a demo: permissions. An enterprise search tool that surfaces the wrong document to the wrong employee is not a productivity feature, it is a security incident. So the real engineering under Glean is not just the language model on top; it is a knowledge graph that respects who is allowed to see what, kept in sync across dozens of systems that each define access differently. That is unglamorous, slow to build, and hard to copy - which is a large part of why the company can charge what it does and why investors keep marking it up. Jain's framing, in the Series F announcement, was about pushing past search into agents that take action across the same graph, which is the next place the money is pointed.

04 / THE SPLITSame enemy, opposite doors

Set them side by side and the philosophies come into focus. Spekit assumes the person is mid-task and should not have to break it. It brings a curated answer to them, in context, and it aims that discipline at one high-stakes role. Glean assumes the person has a question they can't predict, about anything, anywhere in the company, and it makes the whole organization answerable on demand.

Neither is obviously the "right" answer, and that is what makes the category interesting. A curated, in-context layer is precise and trustworthy but narrow - it only knows what someone thought to load into it. A company-wide index is broad but leans on the messy reality of your actual data, which means it inherits every duplicate doc and stale page you have. One is a sniper, the other is a floodlight.

05 / THE TAKEAWAYRetrieval is the product

There is a lesson here worth stealing if you build anything for the workplace. The most valuable feature is often the one that removes a step rather than adding one. Both of these companies win by making people stop looking - Spekit by getting there first, Glean by making the search unnecessary. Neither is selling intelligence in the flashy sense. They are selling the end of the hunt.

Whether the future belongs to the floodlight or the sniper - or to some larger platform that swallows both jobs - is genuinely unsettled. Microsoft's Copilot looms over Glean; a dozen enablement tools crowd Spekit. But the direction is not really in doubt. The office that keeps making its people search for answers it already has is the one paying a tax it doesn't need to. The companies betting you should never search twice have found, so far, a very large room full of buyers who agree.