Podcast Summary

Podcast: Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
Episode: 99 — Using AI Automation to Build Smarter Workflows Across Your Organization with Marc Boscher
Date: April 13, 2026
Host: Chris Daigle
Guest: Marc Boscher (Founder & CEO of Unito)

Episode Overview

This episode dives deep into how companies can make the transition from AI as an individual productivity tool ("single player mode") to a true organizational asset that automates and optimizes workflows across teams and systems ("multiplayer mode"). Marc Boscher, CEO of Unito, shares insights on what holds organizations back, why context is the new battleground for AI effectiveness, and how leaders can overcome integration and change management hurdles to unlock ROI from generative AI tools.

Key Discussion Points & Insights

1. From Personal Productivity AI to Business-Wide AI

The Challenge:

• Most companies think they're "using AI at work" if employees have access to ChatGPT, Gemini, or other tools.
• Reality: This is usually "single player mode," with isolated, manual AI usage, not true workflow automation.
  Marc Boscher (03:24):

"How do we leverage AI today to go from a personal productivity use case to more of a business use case? ... There's a big gap from using AI locally or just for your use cases to actually rolling this out in a cross team or cross organization use case."

Key Insight:
The real payoff comes when AI is deployed at the workflow level, automating tasks between people, departments, and IT systems (e.g., sales to delivery to support).

2. The Context Gap: Why Most AI Initiatives Stall

• The big unlock isn't better prompts but better context: feeding the right, up-to-date, and relevant information into AIs automatically.
• Humans act as middleware when copy-pasting info for AI; this is a bottleneck for workflow automation.

Marc Boscher (06:30):

"The largest multiplier in terms of productivity is when you ... build an agentic workflow that's able to act across your organization ... Unsiloing the agent. The place where this is getting complicated is there's a big trust component. How do you let the agent have the right information at the right time? ... The humans are the ones bridging, giving the context to that AI. ... Once you're able to remove that, where the agent is able to get its context on its own ... then it is able to also take the right decisions without you crafting it."

• Context = any data, documents, or insight that an agent/human needs to do their job effectively (e.g., sales playbooks, brand guidelines, customer records).

Marc Boscher (08:28):

"Context is really just information or data ... when you hire someone and you ask them to do something, if they don't have the right context, they're going to do a shitty job, however smart they are ... you gotta give them the right information or the ways to go and get it themselves."

3. What Is Good AI Context in Practice?

Dynamic vs. Static Context:

• Static: "Snapshot" documents (often outdated, needs manual updating).
• Dynamic: Live, updated data drawn automatically from connected systems (CRM, Slack, support tickets, etc.).

Sales Example (11:34):
Marc describes building an AI deal advisor agent (11:34):

"You could say ... I'm kind of stuck. What would be tactics I could go about to unlock this deal ... this is a something that requires a fair amount of context ... information about the opportunity itself, like who's the customer, who are the contacts ... transcripts of all the conversations ... not just that, LinkedIn profiles, open tickets ... all of that context is really valuable. The same LLM or AI will give you an order of magnitude better result or insights if it has the right context. But all the work is just building that context."

Context Libraries (10:13 & 11:10):

• Moving beyond "prompt libraries" to organized, up-to-date context repositories (documents, data, policies).

4. The Mechanics: Making AI Context Work

• True AI workflow automation needs integrations (not just point tools), connecting all sources of truth inside the business.
• Platforms like Unito provide two-way live sync so an agent in one platform (e.g., Salesforce) has access to info from another (e.g., Asana, Jira).

Marc Boscher (17:45):

"Context is like, we're making the word more complicated than it should be. It's the same thing as with a human ... you need context to deliver on anything."

Fast vs. Slow Approaches (26:53):

• Relying on API calls ("MCP"—multi-platform connections) can be slow and unreliable.
• Live data syncs make agent performance faster and more predictable.

"The context building part ... was taking up to an hour with the MCP step and we went to another approach ... it takes seconds and it's always up to date. ... 90% of the work was happening just to get the context ... That's what the context gap is."

5. Determinism, Trust, and the Human Factor

• Reliable AI outputs depend on moving from probabilistic results to more deterministic workflows (less randomness).
• Uniform context = more reliable, less synthetic results.
• Change management—getting people to actually use new workflows—is the main barrier now, not technology.

Marc Boscher (43:47):

"For AI to be adopted at scale from solo, hidden behind the scene to multiplayer ... you have to have trust, right? And how are you going to get the trust? Well, you need a reliability. You need to know that this is not going to mess it up. And there's a direct ... the better the context, the higher you're going to get trust. Yes, it's a direct, direct correlation..."

• Leadership should focus on removing friction, making context accessible right where people work, and leveraging existing platforms where users are already comfortable.

6. Steps for Business Leaders

• Audit your company's “sources of truth”: What documents, systems, playbooks, and data do you rely on? Where are these stored?
• Clean up and organize context documents—make them accessible and linked to your AI workflows.
• Choose integration platforms that are no-code and user-friendly to avoid bottlenecks.
• Prioritize dynamic (live-updating) context for processes that change frequently.
• Train your teams to think in terms of context supply, not just prompt crafting.
• Empower teams to build agents inside platforms they already use; ensure those platforms have the right data feeds.

Notable Quotes & Moments

• On context vs. prompt engineering (19:09):

  "Prompt engineering ... came and went pretty fast ... I don't know if the term then became context engineering because it's all about ... what's the information you give around it."

• On the analogy to training people (29:34):

  "It's like, instead of having to go through a training program for two weeks, you just give them the file. Right. Like that's like, it's ridiculously fast training. You just swap the context."

• On the inevitability (37:59):

  "Every software platform is an agentic platform or is becoming one. Our humans have developed expertise in the software that you work in every day. So ... the lowest barrier adoption ... is to leverage the platforms they already know best."

• On human change management (35:58):

  "The friction to adopting technology ... has never been technical. It's been purely human change management."

Important Timestamps

| Timestamp | Segment | |-----------|-------------------------------------------------------------------------| | 03:24 | The distinction between personal productivity AI vs. business-wide AI | | 06:30 | The "single player" to "multiplayer" transition—why it matters | | 08:28 | What is "context" in practical terms | | 11:34 | Practical example: building a sales deal advisor agent (context needs) | | 13:43 | The difference between static and dynamic context | | 17:45 | Context is just data—don't overcomplicate it | | 26:53 | Why slow integrations kill AI value; the "context gap" in action | | 29:34 | The value of context in onboarding/training analogy | | 43:47 | Trust, variability, and making AI output reliable | | 46:49 | Next steps for learning and implementing Unito's approach |

Key Takeaways for Listeners

• Upgrade from “prompt engineering” to “context engineering.”
  Modern AIs are smart enough—your job is to make sure they have access to the right, up-to-date information.

• Treat context as the new bottleneck.
  The biggest gains come from making organizational knowledge accessible to automation, not from better prompts.

• Automation will spread fastest where your people are already working.
  Start with integrations and enrich existing systems with live data syncs.

• Adoption is a human issue, not a tech one.
  Make it easy for teams to use AI tools in their existing workflow.

• Trust = reliability = better context.
  When context is live, accessible, and consistent, business-wide automation compounds.

Connect with the Guest

• Marc Boscher on LinkedIn: Link
• Unito: unito.io

Memorable Moments

• Chris’s realization about moving from prompt to context libraries:
  (11:10) Chris Daigle:

  "In 2023, oh, I need a prompt library in 2026, like a context library, a document of context elements ... I hadn't really put that into that contact context before..."

• Marc’s analogy to plugging knowledge into the back of the head:
  (29:34) Marc Boscher:

  "It's like ... instead of having to go through a training program for two weeks, you just give them the file ... it's ridiculously fast training. You just swap the context ... plug it into the back of the head."

Further Resources

• Unito: Cross-platform workflow automation
• Chief AI Officer: AI executive and team training
• Follow Chris Daigle and Using AI at Work on LinkedIn for more AI business insights.

For executives and managers asking, “How do I make the jump from scattered AI tools to real business transformation?”—this episode is a hands-on playbook for making AI a true multiplier in your operations.