On Wednesday 16 of September, in the Central Saint Giles offices of Google, Syrena held The AI Workforce Summit, alongside Lua. The evening brought together about 200 growth-stage founders, operators, technology leaders and investors who already run, or are now building, AI-first functions in sales, operations, growth and hiring. The purpose was to work out what changes when a company stops buying AI tools and starts hiring AI workers.

Attendees at a Syrena AI Workforce Summit evening

Founders and operators at a Syrena AI Workforce Summit evening. Photo: Syrena Network

The org chart is the new product

Most companies still buy AI the way they buy software. A licence per seat, one more tab in everyone's browser, one more login for IT to manage. The question in that room at Google was simple and slightly uncomfortable: what if that is the wrong unit?

Syrena's answer is in its own tagline. It calls itself "the professional network built for AI workers", and its product is not an app. It is a worker called Ren, which finds people, qualifies them, makes introductions and follows up. Lua, the evening's partner, describes itself as a platform where people and AI agents work in one governed workspace. Two companies, one bet: the next big change in work is not a better tool. It is a new kind of colleague.

That is the thesis of this piece too. The companies that win the next two years will not be the ones with the most AI subscriptions. They will be the ones that write job descriptions for their agents, give each agent an owner and check its output like they check a new hire's.

Why this matters now

The numbers explain why a room of founders gave up a Wednesday evening for this. In June 2025, Gartner predicted that at least 15% of day-to-day work decisions will be made autonomously through agentic AI by 2028, up from 0% in 2024. In the same release, Gartner said 33% of enterprise software applications will include agentic AI by 2028, up from less than 1% in 2024.

Gartner also gave the warning that every founder in that room needs to hear. It predicted that over 40% of agentic AI projects will be canceled by the end of 2027, and advised companies to pursue agentic AI only where it delivers clear business value. So the market is moving fast, and a large share of the projects will fail. That is exactly the gap a "worker, not tool" model tries to close: a worker has a goal you can measure, so you know quickly if it earns its place.

Money is following the thesis. In April 2026, Lua announced a $5.8M seed round led by Norrsken22, with Flourish Ventures, 20VC, P1 Ventures, Phosphor Capital and Y Combinator participating. Lua's own page lists more than 10,000 agents deployed, more than 50 million messages processed and 99.999% uptime. In his SiliconANGLE interview at the time of the raise, co-founder and CEO Lorcan O Cathain put the idea in one line: "The org of the future is a 10-person human team with 30 agents."

Lua, partner of the AI Workforce Summit

Lua, the evening's partner, builds one workspace for people and AI agents. Photo: Syrena Network

Read that sentence twice. It does not say "a team that uses 30 tools". It says the agents sit on the org chart. That is the shift the summit was built around.

Inside the room

The format was the first signal that Syrena practises what it sells. You did not just register. You applied, and the application asked for one thing: the single most important objective for your business in the next six months. Places were curated.

Then the objectives did the agenda planning. The stated goals were used to organise introductions, so the opening round of networking was built around what people needed rather than where they worked. Then came "The Thesis", a short opening from Aziz Ben Allala, Syrena's founder and CEO, on what changes when AI becomes labour rather than software.

Founders talking at a Syrena AI Workforce Summit evening

Introductions built around each attendee's six-month objective. Photo: Syrena Network

The panel "Building the AI Workforce" brought together founders and operators who already run AI-first functions. On stage with Aziz were Jason Butcher of Orbit Capital, Lorcan O Cathain of Lua and Danielle D'Lima of Moneybox. The brief for the panel was three questions: what an AI worker should own, what teams look like now and what comes next.

I like that brief because it starts with ownership. Most AI conversations start with capability: what can the model do? The better question is what the agent is accountable for. A tool does not own anything. A worker owns an outcome.

The Building the AI Workforce panel format

The panel asked what an AI worker should own. Photo: Syrena Network

The idea I heard most often, on stage and in the hallway afterwards, was to give agents a job, not a seat: teams of agents with defined roles inside each function, rather than another AI tool on everyone's desk.

Networking at the AI Workforce Summit

Curated networking first, then open networking for the rest of the evening. Photo: Syrena Network

The worker showed up before the drinks

Here is the moment I keep coming back to. Early in the evening, while people were still arriving at Google, an email landed in my inbox. It was from Ren. "I'm Ren, your dedicated worker in Syrena's network," it said, and it explained that a member of the Syrena team had handed my brief over. My stated objective was to interview AI founders and entrepreneurs for our readers, and Ren wrote that it was reading the brief and working out who in the network was relevant.

That is the product doing its job live, at its own launch event. It is also the best possible test of the thesis. Three days later, I wrote back to Ren because I had not yet received a list of candidates. Ren replied, asked a sharper question about our audience and kept working the brief.

I do not tell this story to score a point. I tell it because it is exactly how you should manage an AI worker. You give it a clear objective. You check what it delivers. When the output is not there yet, you say so and give it more context. That is management, not software support. If you treat an agent like an app, you wait for it to work. If you treat it like a hire, you manage it.

Ren, Syrena's AI worker for professional relationships

Ren, Syrena's AI worker for professional relationships. Image: Syrena Network

What it means

For senior professionals, the shift is personal. Your value moves from doing the task to designing it and checking it. The skill that grows is writing a clear objective, defining what "done" looks like and reviewing the work. Those are management skills, and most individual contributors have never been trained in them. Start now, with one agent and one task you already understand well.

For founders, the summit was a warning about positioning. If your product is "a tool with AI in it", a buyer who thinks in labour will compare you to a hire, not to another app. That changes the price conversation. A seat licence is compared with other seat licences. A worker is compared with a salary. That is a much bigger budget, and a much harder bar to clear.

For media, and I include AI Central here, the lesson is about distribution. Syrena did not run a conference with a stage and a sponsor wall. It ran a matching engine with a panel in the middle. Every attendee walked in with a stated goal and walked out with introductions tied to that goal. The content was the hook. The relationships were the product. Media brands that only publish, and do not connect their readers, will feel that pressure.

And for everyone, Gartner's 40% cancellation forecast is the reason to start small. The projects that fail will be the ones with no owner and no measure. The ones that survive will look a lot like the summit's own format: one objective, one worker, one person who checks the result.

The AI Workforce Summit: Revenue, chapter two

Chapter two, The AI Workforce Summit: Revenue, 22 October at Level39. Image: Syrena Network

Syrena is already moving fast. On 22 October it opened chapter two, The AI Workforce Summit: Revenue, at Level39 in Canary Wharf with Deloitte, for about 200 growth-stage founders, revenue leaders and investors. Syrena later said that more than 1,000 people joined the first summit at Google and online.

Key takeaways

  • Think labour, not software - an agent needs a role, an owner and an output you can check

  • Give agents a job, not a seat - assign agent teams to defined roles in each function instead of one more tool for everyone

  • Start from one objective - Syrena seated a whole summit around each attendee's six-month goal, so do the same before you pick an agent

  • Manage it like a hire - when an agent's output is late or thin, give feedback and more context, as you would with a new colleague

  • Plan for failure, measure early - Gartner expects over 40% of agentic AI projects to be canceled by the end of 2027

What's next

The thesis I left with is simple. Stop counting AI subscriptions. Start counting AI workers, and give each one a job description, an owner and a result you can check.

Thank you to Aziz and the Syrena team for the invitation, and to Lua and Google for hosting the room.

Want the interviews and our practical AI tutorials every week? Join 300,000+ readers.

AI Central covers AI, tech and startup events, from global stages like London Tech Week to local meetups and founder dinners. AI doesn't have to be the only topic on the agenda. If there's an AI angle worth sharing, we bring it back to our readers

Running an event, big or small? Write to [email protected] and tell us about it