Sep 29, 2026 · 42 min
Teams must redesign AI agents for shared work
How to Build Team Agents
As agents move beyond individual assistants, organizations must decide how to share knowledge, set guardrails, and fit new systems into existing tools.
- 1Team agents extend AI from private assistance into shared workflows shaped by organizational context.
- 2Expert agents are the most practical early use case because they reduce bottlenecks and add useful redundancy.
- 3Successful deployments depend less on novelty than on curated knowledge, agreed ground truth, guardrails, and ecosystem fit.
Don't miss
Nufar Gaspar argues that configuration and knowledge curation, rather than model selection, are the hardest parts of deploying a team agent.
The brief
AI agents are becoming mainstream, but most work still happens collaboratively. Nathaniel frames the central problem: private assistants are poorly matched to teams that share decisions, tools, and institutional knowledge.
Nufar Gaspar defines team agents as shared systems designed for an entire group, not simply personal assistants exposed to more users. That shift makes organizational context and intentional design unavoidable.
The discussion distinguishes four archetypes and finds a practical starting point: expert agents that reduce bottlenecks and provide useful redundancy are already easier to implement than agents coordinating across teams.
The sharpest operational lesson is that configuration and knowledge curation are harder than choosing a model. Teams should establish ground truth, guardrails, and use cases before selecting a tool adjacent to their existing ecosystem.
Team agents are therefore less a product category than an organizational design problem: the winning implementation fits current workflows while making shared expertise more accessible and dependable.
What was said on this episode
39 statements · 25 positive · 8 negative · 2 mixed · 4 neutral
2026 has been the year when autonomous agents became mainstream.
“2026 has been the year of agents.”
Listen at 0:00
AI agents have moved from anticipated technology to present reality.
“we really have gone from agents being the next big thing to just being here.”
Listen at 0:16
Shared-workspace agents are beginning to emerge.
“we're gonna start to see, and I guess we're starting to see early evidence of, more agents that live in between people's shared workspace.”
Listen at 3:17
Some agents should remain private while others become team-level agents.
“some agents should stay yours and private while others should become the teams level agents.”
Listen at 4:29
Team agents can reduce knowledge bottlenecks and cross-team coordination problems.
“an agent that is built for the whole team can help with both of these problems”
Listen at 6:12
AI-forward companies have begun merging individual agents into team-level agents.
“They started to merge some of those agents into team-level agents.”
Listen at 7:00
A team agent is one shared agent with common knowledge, memory, and configuration.
“one agent that many people talk to. With shared knowledge, shared memory, and one configuration.”
Listen at 8:23
Team agents handle diverse team tasks, unlike skills focused on specific tasks.
“a skill is a playbook for a specific task, where a team agent is something that your whole team works with on diverse set of tasks.”
Listen at 8:46
A skill library and a team agent are distinct systems.
“they are not one and the same.”
Listen at 9:10
Not every agent should be shared.
“not every agent should be shared.”
Listen at 13:42
Shared knowledge with private agents is often the easiest starting point.
“the easiest place to start”
Listen at 14:27
Data agents can answer questions spanning multiple company departments.
“it can be the data agent that can answer any data question across multiple departments in the company”
Listen at 15:58
Agents where personal taste outweighs standards should remain private.
“those need to remain private agents”
Listen at 19:18
Unowned or disputed shared knowledge causes team agents to drift rapidly.
“If the team cannot agree on how the work is done or nobody is willing to own and maintain the shared knowledge over time, the agent will drift within weeks, sometimes within days.”
Listen at 19:29
Shared agents drift quickly without knowledge ownership and maintenance.
“the agent will drift within weeks, sometimes within days.”
Listen at 19:36
Teams should delay team agents when complexity exceeds their benefits.
“Don't build it, at least not until you are able to untangle some of the complexities.”
Listen at 20:03
Teams should initially limit agents to reading and drafting.
“ideally start with narrower scope, reading and drafting”
Listen at 22:51
A shared folder is the simplest way to begin sharing an agent.
“The simplest method will just to create a shared folder with your own tools”
Listen at 23:21
Acting with the requester's access is generally the safest team-agent access model.
“the safest choice”
Listen at 30:29
Teams should not use one person's login for shared agents except narrowly.
“So I would not recommend to go down that path.”
Listen at 31:00
Teams should not use one person's login for a shared agent except narrowly.
“I would not recommend to go down that path.”
Listen at 31:01
Sensitive team-agent answers should be delivered privately to the requester.
“if it's sensitive, the answer has to go to the person who is asking privately and not in a shared channel.”
Listen at 32:38
Sensitive agent answers should be delivered privately to the requester.
“the answer has to go to the person who is asking privately and not in a shared channel.”
Listen at 32:39
Each team agent should have one owner or a small ownership group.
“I want to have one person or a very small group of people who owns the priorities, maintain the agent”
Listen at 33:31
Each team agent should have one owner or a small ownership group.
“one person or a very small group of people who owns the priorities, maintain the agent”
Listen at 33:32
Teams should pilot agents with a small group and repeat test questions after changes.
“you will probably start with a small— ideally, you should start with a small pilot group and keep a few test questions that you can rerun whenever something changes.”
Listen at 34:47
Teams should launch team agents with a small pilot group.
“you should start with a small pilot group”
Listen at 34:49
Teams should pilot agents with a small group and repeat test questions after changes.
“you should start with a small pilot group and keep a few test questions that you can rerun whenever something changes.”
Listen at 34:49
Team-agent tools will increasingly formalize identity, permissions, and ownership.
“we will see more and more formalization of what we just covered and more tools and features that will help us get it even better and easier around identity, permissions, ownership, and so on”
Listen at 37:45
Agents operating between individuals or teams are hardest to execute.
“it's the hardest to execute.”
Listen at 38:37
Bridge agents are generally not teams' first adoption choice.
“the ones that will, from what I'm seeing, are not the first to go for.”
Listen at 38:40
Expert agents can add redundancy and reduce organizational bottlenecks.
“I've seen various very successful versions of the first one, of the expert agents that help create more redundancy in a team, redundancy in the good sense, and relieve some of the burden on those bottlenecks within the company.”
Listen at 38:44
Expert agents already have many implementations and should become easier to build.
“I've seen a ton of implementations already, and I think the more the tools make it more accessible, the easier those will be to build.”
Listen at 38:58
Expert agents already have many successful implementations.
“I've seen a ton of implementations already”
Listen at 38:58
Internal knowledge-hub agents have been successful early use cases for companies.
“companies have had a ton of success with as just an early, easy, fast use case right from the beginning.”
Listen at 39:41
Configuration and knowledge curation will remain the main work of team-agent implementation.
“the heavy lifting is always gonna be the configuration and the knowledge curation.”
Listen at 40:29
Teams should choose a team-agent tool adjacent to their existing ecosystem.
“I would select the one tool that is adjacent the most to your existing tool ecosystem”
Listen at 40:35
Native team-agent products will become more accessible and capable.
“all of these are coming and will probably be made very accessible and very smart.”
Listen at 41:21
Organizations’ durable advantage lies in information vendors cannot access.
“Your moat is probably in everything that these companies cannot tap into.”
Listen at 41:26
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.