How Manus became “James,” the AI chief of staff behind Ascendea’s 90× output shift
Ascendea Team
A 90x increase in workforce output. Not over years, but over months. Ascendea did it with ten people and Manus as an AI chief of staff, the CEO's right hand for turning priorities into action.
James is what Rob Arnold, Ascendea's CTO, calls the Manus agent he put at the center of the company's operations. Based at the University of Warwick Science Park, Ascendea builds AI-powered CRM, agents, and automation systems for UK small businesses. Rob regularly tests new AI tools by assigning them real side projects. In most cases, they do not make it into Ascendea's production workflows.
What earned Manus a permanent seat was the output on complex work. Team members who had never written a prompt were getting usable results on multi-step tasks without the usual rounds of revision. The gap between a good idea and a working system, the part that used to take weeks of manual execution, started closing. Soon, Rob gave Manus a name: James.
How James Became the Chief of Staff
Ascendea placed Manus at the top of a three-layer system for the company's most complex, context-heavy work. Simpler and more repetitive tasks stayed with specialized Slack and local agents.
Ascendea learned to reserve Manus for the complex work where its execution capacity mattered most.
In the first month, Rob built apps, generated reports, and designed automations just to test the limits. The pattern he kept seeing: the more complex and multi-faceted the task, the less back-and-forth it took to get a usable result. That was enough to move Manus out of the side-project phase and into production.
Rob put Manus near the center of Ascendea's operating model. Ascendea matches each tool to the complexity of the job. "You wouldn't use Manus for the basic stuff. So we've created a hierarchical system where Manus sits on the top." The result is an AI workforce spread across three layers, each tuned to a different kind of work:
•Top layer - Manus (James): Handles all complex, context-heavy work - strategy documents, technical architecture, research, and anything that needs deep reasoning or multi-step execution.
•Middle layer - Victor, Victoria, and Ava: Conversational agents living in the internal Slack workspace that serve as the team's human-facing interface. Victor is the AI support; Victoria covers operations; and Ava handles customer service. Anything beyond their scope routes up to James.
•Local layer - AI Crew: Six local agents running autonomously on a dedicated server that handle repetitive, high-frequency work on a fixed schedule.
Ascendea uses the three layers as one operating system. The team uses Manus to draft strategies, SOPs, and technical blueprints, while local agents handle scheduled tasks against those plans. Slack agents give employees a conversational interface. When a task exceeds a local agent's defined scope, Ascendea routes it to Manus for a more complex response and keeps human review in the process.
The first win: 303 SOPs in a day
Ascendea used Manus to turn scattered process knowledge into 303 SOPs in one day. Rob estimated that documenting the same material manually would have required two people for about 14 weeks.
Documentation was the obvious place to start. Because Ascendea runs a different process for every service, work quality started to vary as the client base grew. New hires needed weeks of hand-holding, and Rob remained the single point of reference for almost everything. To fix this, Ascendea used Manus to build a comprehensive SOP system.
Ascendea used Manus to turn scattered process knowledge into 303 SOPs in one day, which according to Rob, "would have been probably about a 14-week project with 2 people working on that."
The payoff showed up quickly. When Ascendea hired three SDRs in April 2026, the SOP library helped them begin client work within days, according to Rob. The same documents also became instruction sets for the company's local agents. This allowed the human team to spend less time assembling routine work and more time setting direction, reviewing outputs, and making final decisions. A new hire reads an SOP to learn the process; an agent reads it to execute it.
Beyond 303 standalone SOPs, Ascendea built the internal onboarding hub at https://manus.ascendea.ai/ to roll this work out to the rest of the team. Employees now have a simple path to use it in their roles. Ascendea transitioned Manus from a personal experiment into a core part of its operating system.
From a single project to the whole operation
After the SOP project, Ascendea expanded its use of Manus across reports, automations, and a prospecting workflow.
The SOP project proved what Manus could do, and Ascendea widened its remit from there. It started with Rob, then spread to the team. Not everyone adopted it overnight, but once each role had a tailored SOP showing what Manus could handle for them, most did. "Generally everyone is now using it day to day," Rob said.
For Rob, Manus became a technical thought partner. It challenged assumptions, compared tools, generated reports, and shaped implementation roadmaps. "It challenged some of my ideas," he said, "which was great to have someone that would say, how about doing it this way, and what about using these other tools."
Beyond Rob's desk, different teams now lean on Manus for the documents and reports they used to assemble by hand. The client-facing side uses it to draft structured proposal documents. The account and delivery side uses it to produce weekly performance reports. Over on the strategy side, it writes strategic analysis like competitor breakdowns and market reviews that previously took an analyst days to compile. Manus assembles the initial drafts and reports from Ascendea's authorized data sources, while the team reviews the output and owns the final judgment.
That same execution capacity soon pointed outward, shifting from internal documents to business development. Ascendea's M&A arm needed a consistent way to identify digital marketing agencies that matched its acquisition criteria. Rob asked Manus to design the workflow, write the code, and document how the system should operate. The automated workflow analyzes publicly available company data, ranks potential matches, and prepares records for review. Once Ascendea's team has reviewed and approved a recommendation, the workflow automatically adds the record to the company's CRM. What used to be slow, manual research now runs quietly in the background.
When ambition no longer waits for headcount
Today, Ascendea runs on just 10 human staff, supported by an AI workforce that includes James (Manus), 3 Slack Agents, and its 6 other agents. That setup serves 112 active clients across accounting, finance, tech, and coaching, and runs 28 to 33 live AI workflows coordinated through James. A small company now produces the output of a much larger one. Manus gives people who understand their clients the capacity to act on that knowledge at the speed the market demands.
With AI carrying the execution, the capacity of the team is no longer limited by headcount.
