How Motive uses 1,000+ AI agents to boost productivity by 50%

In this episode of This New Way, Maddie Engelmier (AI leader at Motive) walks us through a three-tier AI adoption plan:

1. Democratize AI

2. Automate core workflows

3. Transform the tech stack

Her team supports a bold company OKR, increasing employee productivity by 50 percent, by rolling out secure, connected agents that sit on top of real work data.

🎬 Listen to Maddie’s episode

🍿 Subscribe to the YouTube Channel

Maddie spent this episode walking through how Motive uses Glean to build internal AI agents. Her workflows are based specifically on that system, but even if your company doesn’t use Glean, the same concepts apply to any enterprise AI platform that supports workflows and agents.

Tutorial 1: Ship a self-assessment agent that cuts review time by half

Why it matters: Performance cycles stall on recall. This agent compiles what happened, then prompts the human to reflect, so your reviews are higher quality without turning into AI paste.

1. Connect your data sources

   • In your enterprise AI platform, enable connectors for Slack, Google Drive, Gmail, Confluence, and your HR performance tool.

   • Enforce source permissions so users only see what they already have access to.
     

2. Create the agent skeleton

   • Trigger: ask the user for role, level, and links to their last review.

   • Inputs: time window (for example, last 6 months), job architecture or competency matrix, prior feedback doc.
     

3. Run time chunked company search

   • Loop across the date range, query each source, and extract artifacts tied to the user: projects, threads, docs, pull requests, decisions.

   • De duplicate and rank by impact signals (mentions, approvals, file activity, meeting notes that reference outcomes).
     

4. Map work to competencies

   • Load the role matrix, then classify each artifact to competencies such as domain expertise, communication, delivery, leadership.
     Produce an “evidence board” per competency with links.
     

5. Close the loop with human reflection

   • Generate a concise summary, then add 4 to 6 targeted reflection prompts.

   • Example prompts: “Where did you change a decision with data,” “What would you do differently in rollout,” “Which feedback from last cycle did you act on.”
     

6. Create the handoff

   • Output a clean doc: contributions table, evidence links, last cycle feedback vs actions, plus reflection questions.

   • Avoid auto writing the self review. Make the doc the thinking surface the person uses to write their own narrative.

Pro tips:

• Prefer retrieval by people and objects, not just keywords.
  Keep prompts short and specific. Long prompts often reduce recall precision.

• Gate this agent behind a simple form so input fields are consistent.
  

See it in action: Watch the demo here.

AI Resource: Security Checklist

Most meetings include confidential data -  from project details to client information. But not every AI notetaker protects that information with the security your company deserves. If the AI notetaker you’re using trains their AI models on your team’s conversations, you could be putting your company at risk without realizing it. This AI Meeting Notetaker Security Checklist helps you avoid that.

In just two minutes, you’ll learn the 7 checks to ensure your team’s AI meeting notes stay private and secure. Don’t let your meetings become someone else’s dataset:

Tutorial 2: Build an executive account summary agent for live customer insights

Why it matters: Leaders need fast, trustworthy context before a call or weekly business review. This agent compiles CRM, support, and internal chatter into one page in seconds.

1. Pick the single input
   

   • Make the only required field the Account Name or CRM record URL. Less friction equals more usage.
     

2. Query structured systems first
   

   • CRM: pull ARR, stage, renewal date, owner, next steps.

   • Support: open cases, severity, time to first response, recent CSAT.
     

3. Layer in unstructured signals
   

   • Slack: recent threads that mention the account, prioritize posts from account team and leadership.

   • Email and Drive: latest meeting notes, exec threads, EBR decks.
     

4. Extract risks and opportunities
   

   • Classify patterns: feature gaps, adoption risk, billing issues, expansion potential.

   • Include a short rationale with the source link for each callout.
     

5. Assemble the one pager
   

   • Sections: Overview, Open Opportunities, Open Cases, Risks, Watch Items, Recent Internal Notes, Next Best Actions.

   • Keep total length under 300 words, with links out for depth.
     

6. Deliver where work happens
   

   • Expose it as a slash command in Slack or a sidebar in CRM.

   • Optional: pre generate summaries for tomorrow’s meetings by reading calendars, then post to the owner’s DM.

Pro tips:

• Rank evidence by freshness and author reputation.

• Always show source links so leaders can verify quickly.

• Redact sensitive fields by default unless the requester is on the account team.
  

See it in action: Watch the demo here

AI tools mentioned:

• Glean (or your enterprise RAG platform): secure connectors, agent builder, permission aware retrieval.
  

• CRM and Support tools: Salesforce, Zendesk, or equivalents for structured data.
  

• Work graph sources: Slack, Google Drive, Gmail, Confluence.

Why these playbooks matter

Agents that remove recall, then prompt reflection, raise the floor for everyone. Start with high frequency, high friction workflows and you will see immediate impact.

👉 Subscribe to This New Way: YouTube Channel

 📩 Was this email forwarded to you? Subscribe here

Until next time,

Aydin Mirzaee
CEO at Fellow.ai & Host of This New Way