Former Workplace from Facebook engineers launch AI start-up Slashwork to banish teamwork slop

It doesn't happen very often that you have customers before having the product.

Workplace communication platform Slashwork has launched out of stealth today with $3.5 million funding and an advantage that most startups can only dream of — an extensive rolodex of prospective customers eager to try out its as-yet unreleased product. That's because the team building Slashwork includes some of the key players previously involved in Facebook's much-loved enterprise teamwork platform, Workplace, which its owner Meta shut down last year after unexpectedly deciding to pull out of the enterprise teamwork market. The comment above comes from Julien Codorniou, General Partner at London-based specialist media and tech investor 20VC, which led the investment, and who was previously VP of Workplace. He elaborates:

A lot of ex-Workplace customers — you know, we had more than 11 million paying subscribers — they still miss Workplace. And the reason why we started this business with [CEO] Jackson [Gabbard] and the team is that they were still calling me, you know, six months ago, to tell me, 'Hey, we miss Workplace. Facebook killed it. We tried everything, and nothing really works.'

That's not to say Slashwork will simply be a me-too re-incarnation of Workplace, even though its three co-founders — Jackson Gabbard, CEO, David Miller, CTO, and Josh Watzman, Chief Engineering Officer — each spent between five and 11 years as software engineers at Facebook, where Watzman led the technology build-out for Workplace. Angel investors who've joined 20VC in the current funding round include Cal Henderson, co-founder and former CTO at Slack; AJ Tennant, former sales chief at Slack and more recently Glean; Des Traynor, co-founder and Chief Strategy Officer at Intercom; Soleio, who led product design at Dropbox and later advised Figma; and a clutch of Facebook alumni including Sheryl Sandberg, Meta's former COO. Codorniou says that the aim is use AI to go beyond earlier generations of workplace communications product:

We don't want to do Workplace 2.0 or a better version of Slack. We want to build something first principles in the age of AI and Cursor and Claude Code in 2026. While Workplace and Slack were focused on talking to your employees, we also want AI and Slashwork to focus on talking to your employees, but also your systems. This is a great use case for AI, and AI is a great platform to build this product.

What this means is that Slashwork will be much more of a platform than an app, providing a robust infrastructure and set of services and connections that customers will be able to call from LLMs such as ChatGPT or their own custom or vibe-coded apps. The two dozen or so early access customers who are design partners for the product will need to be comfortable playing with the technology, as Gabbard explains:

We built our product on top of an introspectable GraphQL API, which is important because it's an API that LLMs are extremely good at using. We have access tokens, and so we expect — we know — that people are going to use our product from other places. They're going to use it from ChatGPT, they're going to use it from the other tools that they have, and they're going to funnel stuff in. They're going to do things in our platform. We're welcoming that. We expect more and more power to go into the hands of people who have an LLM and an inkling of how to get something vibe coded, and we're leaning hard into that, to make this a very hackable experience.

Codorniou adds that the end result will be a product with very broad appeal:

The people building this product have built software products used by billions of people. And so I think the magic of Slashwork will be that the interface will look very simple, will look magical, but the platform is extremely complicated. These guys have done it many, many times, across many different products. And so using AI, but making it in a way so that even my mum could use it without feeling that she needs to get a PhD to use the product. I think this is the magic that this product team is bringing to the table.

Making a difference

The difference that AI can make will come in areas such as feed ranking and notifications, where generative AI can do a much better job of figuring out what matters to each individual user than traditional predictive algorithms, or in natural language prompting, where users can simply describe the outcome they want and the AI will figure out how to deliver the requested information or action. Starting out with a business mission will also avoid some of the constraints that Workplace suffered from. Gabbard explains:

One of the tricky things about Workplace was that the Facebook infrastructure was sort of forced onto Workplace and had to work the same way. Facebook is trying to keep your attention. They want to show you more ads. They want to keep you on the platform. We don't have this constraint at all. We're designing this around getting you the most relevant content, so you can do your work. That's the stuff you need to respond to. That's the posts that are really actually interesting for the company progress, for the things that are happening. It's not that thing that's going to keep you clicking the next one. If you want to refer to it in algorithmic terms, we get to bias our algorithm towards utility, not towards attention.

The team also recognizes the high standards of accuracy and confidentiality that a business context demands. Gabbard says that they have put a lot of work into building instruction loops that constantly test the responses generated by an LLM, to check that they match the requested outcome. Watzman brings his experience of building out Workplace with robust enterprise identity and access management. Gabbard goes on:

That includes a huge amount of work that they did on identity management, on privacy rules, on making sure that an enterprise-tier product has all of the connections that it needs to have, imports and synchronizes all the ways that it's supposed to synchronize, and doesn't give up the goods to people who aren't supposed to have it. We just we have that knowledge in our team already... If we hold true to this path, we should be able to, when it's time for enterprise, have a product that we know is really trustworthy.

The company is currently working with a couple dozen customers, focused initially on knowledge workers in tech-savvy companies, but with a view to ultimately expanding its reach to the same broad user base that Workplace attracted, including frontline workers. But such is the commitment to recreate the product loyalty that Workplace achieved, all of Slashwork's early design partners have been selected from people already familiar with Workplace. Gabbard explains:

We believe that the ex-Workplace people are probably the best poised to just immediately understand the value proposition. There's some magic to Workplace that is lost in other tools. And the funny thing about magic that you haven't seen is, you don't believe in it. So we go to people who've never heard of Workplace, they never tried it. They're like, 'Yeah, but like, you know, I've got Teams, like, that's good enough, right?' [We're] like, 'Oh, it's like, it is so catastrophically bad. I can't explain to you the colors of the rainbow you haven't seen.'

So we're starting with ideally small, tech-oriented, ex-Meta-founded companies, and a small subset of that will also be non-Meta, ex-Workplace customers... people who really will tell us what they want, and we build around that. As we get solidity there, as we get product-market fit in a really deep sense, we'll know when it starts to feel like Workplace felt — when we have that same amount of love.

My take

Even though it sounds as though the Slashwork team want to build something with the same cross-enterprise reach and customer loyalty that Workplace achieved, I get the impression that this product will look very different. It may not even have a look at all, if customers end up calling on its functionality from other apps and AI assistants. But there is still an unmet demand to be satisfied — despite years of promises from a succession of teamwork vendors — for a team communications platform that surfaces what really matters to people at the right place and time, without overwhelming them with messages and notifications. You'd think that this is something that generative AI would be good at, but it's actually really difficult and depends a lot on the underlying infrastructure — including something I like to call a System of Knowledge. As Codorniou suggests, high--performing infrastructure is something this team ought to be able to do a good job of building.

As former Workplace users, we at diginomica have seen the colors of this particular rainbow. It's true that there's still nothing quite like it in the market. But it's also true that Workplace had a lot of shortcomings, too. It would be good to believe that Slashwork can recreate the same magic while using AI to dispel some of the downsides. There's certainly a gap in the market to be filled, and perhaps a pot of gold to be found by the team that can get to the end of this particular quest.