Inside Rival: How We Use What We Build
This blog is part of Inside Rival, a series where our team shares how we use Rival in our everyday work. From favorite tools to AI agents and automations, we're showing you what we're building, how we're using it and what we've learned along the way.
Three clients. Three projects. Three sets of deadlines. And somehow, everyone needs something from you right now. Sound familiar? Last quarter, I was juggling a major client launch alongside two other projects when one missed deliverable set off a domino effect. Three more tasks slipped, and nobody caught it until the status meeting. That's when I realized the problem wasn't that I needed to work harder. I needed a smarter way to manage multiple clients without losing my mind (or my entire day to project trackers).
The challenge isn't getting the work done. It's keeping everything moving across multiple clients (or even projects) without losing track of the details. Every client has different expectations, timelines and tools. Trackers need updating, dependencies need monitoring and deadlines need managing. And when you're the person responsible for keeping it all organized, the administrative work can quickly take over the time you should be spending on your clients.
The Real Cost of "Work About Work"
This isn't just your problem or mine. According to Asana's 2023 Anatomy of Work Global Index, knowledge workers spend roughly 60% of their time on "work about work" — the coordination, status-chasing, and admin that orbits actual project work. Only about a quarter of the week goes to skilled work, and just 13% to strategy.
Now translate that to project life. Every tracker update. Every "any update on this?" email. Every manual re-prioritization. That's the 60%. It gets worse when work lives across tools: the brief in Google Docs, approvals in email, tasks in Asana, milestones in your calendar. Every switch between tools is another small tax on your attention.
Why a Single General Assistant Agent Falls Short
The obvious answer is an AI project management assistant. One agent, all your projects. Sounds clean. It isn't. A general assistant can't deeply understand three different projects at once. It knows a little about all of them and a lot about none of them.
Worse, the context bleeds. Ask it about Project A, and it's still carrying Project B's priorities: the stakeholders, the deadlines, the vocabulary. Before long you're correcting an agent that was supposed to save you time. A jack of all trades is master of none, and that's exactly the problem with relying on a single AI agent for project management across every project. Project leads know that phrase. It applies to agents too.
The Per-Project Agent Approach: One Agent, One Project, One Expert
Instead of one assistant for everything, I create per-project AI agents. Each one is a dedicated expert on a single project's ask, timeline, and end result. Sounds like more work. It's actually less, because on Rival, creating an agent takes minutes, not days. And the payoff is that each agent knows one project cold.
Specialized Expertise Changes the Game
A per-project agent holds only one project's context: the stakeholders, the milestones, the dependencies, the definition of done. No context bleed. No "which project are we talking about?" moments. When you ask it something, the answer is about the project you're both working on, because that's the only project it knows.
How It Works in Practice
Right now I run dedicated agents for either big projects or one large task per week. Each one was briefed on its specific scope: stakeholders, timeline, success criteria. Then I connected the tools the project lives in. The agent handles the tracking. I handle the judgment calls.
How to Set Up Your AI Agent for Project Management on Rival (in Minutes)
Here's how to create an AI agent when a new project lands, the whole loop takes about as long as your coffee.
Step 1: Start With Rivalbot
RivalBot is my best friend when it comes to creating these agents. I tell RivalBot what my project is and what I want the agent to be an expert on. RivalBot spins up the agent in minutes allowing me to connect my tools and customize what I need.
Step 2: Brief the Agent on the Project
Name it after the project. Then write instructions the way you'd brief a new teammate. Rival's How to Build an Agent on Rival guide spells out the framework: Purpose → Instructions → Procedure. What outcome is this agent responsible for? How should it behave? What steps should it follow?
I tell the agent the background information on the client, or for internal project management I provide background on the team working on the project to make sure the knowledge base knows everything I know. .
Step 3: Connect the Tools the Project Already Lives In
The connectors are super important for the success of my agent. For each project, I share the tracker and internal documents created for the unique project including Google Docs, and Asana project name. By doing this, I can ask the agent to make direct updates to trackers, create new tasks, respond back to tasks and share project updates via email.
Connect the project's home base:
Google Docs: the brief, the deliverables, the source of truth.
Email: client comms and approvals.
Asana: tasks, owners, and dependencies.
Calendar: milestones and hard deadlines.
Now the agent lives where the work lives.
Step 4: Let It Run
Once briefed and connected, the agent gets to work. It builds an internal tracker for the project. It creates tasks in Asana. It watches deadlines.
TL;DR: duplicate, brief, connect, run. You don't run the agent. It runs for you.
Proactive Deadline Monitoring
The agent watches task dependencies continuously. When one task slips, it recalculates what that means for everything downstream before the slip becomes a crisis.
Flagging Owner-Critical Items
This is the feature I didn't know I needed. The agent flags the items that need my action to unblock other people. "This approval is waiting on you." "This review is due tomorrow." It surfaces them before the status meeting, not during it.
Auto-Updating Trackers
Deadlines shift; the tracker follows. When dates move in Asana or your calendar, the agent updates its tracker automatically so you never hand a client a stale report. Running the agent day-to-day is as simple as opening it and asking.
Daily or Weekly Project Briefings
Set a ritual and the agent delivers a briefing on a schedule: what's on track, what's at risk, what needs you. It lands in your inbox like a note from a teammate who actually reads the tracker.
The Time You Get Back
Here's my before and after.
Before: 30 minutes every morning updating trackers. Another 30 chasing status through email and Asana. A scramble before every status meeting to reconstruct where things actually stand.
After: The agent updates the trackers. It chases the status updates. It hands me one briefing that says what needs me.
I don't spend less time on the project. I spend it on the parts that need judgment, the parts that decide whether the launch succeeds.
About the Author:
Alexa Rifkin is Associate Director of Enterprise Strategy at Rival, where she helps organizations put AI to work in practical, meaningful ways. From building AI agents and automating workflows to managing complex client projects, Alexa focuses on making technology easier to use and more valuable for the people behind it. She works closely with enterprise clients to turn big ideas into solutions that save time, simplify work and deliver real results.