AI is supposed to be a cure for busyness — the key to smooth, stress-free working days. And yet, somehow, it often works in the opposite way. More automated notifications. More AI-generated messages. More tasks added to full to-do lists.
There’s a fix for this, and our latest Wrike AI Summer School session delivered it. Week one covered the AI foundations. Week two was all about personal productivity, which has become a whole different animal since we’ve all started sharing a desk with Claude, Gemini, ChatGPT, and more.
If you missed our latest session, here’s a quick recap: Our host, Melissa Kovach summed up the whole class in one line: if we want to start really improving productivity, we need to stop adding more work into our day and start clearing out the work that never needed us in the first place. Then she showed us how to do it.
1. Create genuinely helpful agents fast
The aim of the Wrike Summer School — a seven-part series — is to cut through the hype and deliver hands-on advice you can action the same day. This means live demos from product experts, ask-me-anything Q&As, and a straight-speaking approach that aims to meet you where you are.
Marco Bottaro, product manager for Wrike’s AI agent builder, kicked things off with a new release that makes setting up an agent a two-minute job.
The old way meant clicking through a step-by-step configuration form. Unless you’re a developer, that could get confusing fast — many people just gave up and asked an outside model for help instead, which means laboriously documenting every step of your workflow to a bot. Results were, at best, a coin flip.
The new Wrike conversational agent builder skips all of that. It already knows every inch of your work environment, so you don’t have to explain any of it. It can point right at the folder you mean, the tasks inside it, the workflow, and the people you can assign to.
It also knows how agents work and where they hit their limits, which is important — false promises cost you time at best, and revenue or reputation at worst. You tell it what you want in plain English, it asks a few smart follow-up questions, and it keeps going until the agent actually does the thing.
In Marco’s demo, an agent read an incoming brief, added a prefix, set the task type, and assigned the work to the right person. No manual setup. No hours of tweaking and adjusting.
The demo sparked a few great questions from our audience, which helped clarify some things:
- Building an agent doesn’t eat into your AI credits: It’s the same as playing around in the testing sandbox, so don’t worry if it takes a bit of back-and-forth to get it just right.
- Triggers are not limited to brand-new items: A changed custom field, a new assignment, or a schedule can all set an agent in motion. Set one to run every Monday at 7 a.m., for example, to get ahead of your working week.
- The old configuration system is still there: The form is still functioning in the highly intelligent way it always did, the conversational builder just fills it in for you in a natural, conversational way. Plus, you get to review everything before the agent runs, so there won’t be any surprises.
The result is an agent that actually understands how you work and what you need. One of the most popular use cases is request triaging. Build an agent to handle this for you, and all that soul-draining sorting work is done before you’ve even logged in.
2. Choose reasoned updates over abundant notifications
Quick question: How many unread notifications do you have in your inbox right now? It’s probably not a number you want to be heard saying out loud. One of the biggest culprits is automation, which sends a message every time something changes. Some of these are important — others, less so. The problem with too many notifications is that eventually, people start to ignore them.
This is where agents kick in. Where automation can be a blunt tool, agentic AI can get surgical.
The difference is reasoning. Because an agent can actually size up a change and analyze an action, it decides what’s worth a message. In Yulii Lyubeznikov’s demo, an agent looked at an incoming request, spotted that it was urgent and tied to a key customer, and posted to a team channel so the whole team knew why that task jumped the line.
The best part is that those smart notifications can now reach people wherever they are, inside Wrike or outside, in Slack, Microsoft Teams, or email.
The second great trick Yulii highlighted is an agent’s ability to hunt down missing details in half-filled request forms. In his demo, a request came in without enough information to start, so the agent emailed the requester directly, added a due date, and spelled out what was missing. In the past, that would’ve taken a full day of waiting and thumb-twiddling. Here, it just happened automatically, quietly, in the background.
Both examples chip away at the same headache: notification overload. When we strip back that noise, we can start to get productive rather than just busy.
3. Get instant, contextual answers with smart search
Nancy Boas, who leads AI product marketing, closed out our session with a tool you probably use a thousand times a day: Wrike search.
She explained that the latest changes are the result of real-world feedback. Customers said it needed work, so a dedicated team rolled up their sleeves and rebuilt it. Deeper filtering is in, relevance rankings and organizational structure got a rework, and they are already measuring a 22% improvement in getting that first click right.
There’s more on the way, too, including semantic and similarity search and inference that recognizes related terms. That means hunting for “summer school” could also surface “summer education,” which is helpful when you’re trying to see the big picture.
Wrike Copilot is the fast, casual cousin of traditional search. Ask it a question in plain language, typos and all, and it reasons its way to an answer. In the demo, a professional services consultant asked which project had the most billable hours, then which was most profitable, then who had the most tasks piling up in a folder. Answers came back instantly, and they were drillable right down to the projects underneath.
We talk a lot about context here in Wrike, but what it means is that our AI tools — like Copilot — are embedded in your actual workflow, not installed as an extension or integrated as an add-on. Every time you interact with our AI, it’s pulling from a long history of how things work in your organization. It knows where projects get bottlenecked, what clients need updates, and how assets actually get approved.
The demo showcased this brilliantly. Asked how to shift some work from one overloaded teammate to another, Copilot suggested specific tasks that could be reassigned and let the consultant swap them out on the spot. That one tiny action avoids burnout, safeguards deadlines, and maximizes resources — in an instant.
The combined result? Less busywork
Every one of these tools tackles a small, daily annoyance. Sorting requests. Chasing missing details. Digging for one statistic that’s buried three folders deep. On their own, agentic actions save you a few minutes. Stack them across a week, then across a team, and those minutes turn into the difference between surviving and thriving in your working week.
That is exactly why we focused on personal productivity in the AI Summer School. The small wins for one person are cumulative. If enough people make this part of their regular workflow, you’re going to have a super-efficient team, a high-performing department, and an organization that’s got a real competitive edge instead of a token-heavy tech stack.
Don’t miss your next class
Week two was just the second of seven, so there’s plenty of school left this summer. Classes run every Friday through August 28, 2026, starting at 8 a.m. PDT, and each one wraps in about 30 to 45 minutes. Here’s what’s on the syllabus:
- Week 3, July 31: Project management, including how to plan faster, prioritize with confidence, and catch risks before they turn into roadblocks.
- Week 4, August 7: Team collaboration that leads to fewer status meetings, better alignment, and updates that keep everyone on the same page.
- Week 5, August 14: AI agents and MCP: Where things get really interesting, connecting your tools and turning conversations into action.
- Week 6, August 21: Leadership insights, including how leaders use AI for visibility, governance, and faster decisions.
- Week 7, August 28: Graduation day: advanced workflows, real customer stories, and an expert Q&A.
Show up live to all seven, and you earn a Wrike AI Summer School Certificate of Completion to brighten up your LinkedIn profile, plus a spot in the draw to win a curated box of premium Wrike swag. And take it from us — we do swag well.
Save your seat for Wrike AI Summer School and pick up right where week two left off. We’ll see you in class!
