The Salesforce PM rewriting how enterprise teams use AI in project delivery
Brian Galardo, Program and Project Manager at Salesforce, shares how AI is transforming project delivery — not by replacing human judgment, but through practical workflows that move teams from experimentation to real adoption.
Cross-functional Initiatives: Brian Galardo emphasizes the importance of cross-team collaboration for effective project delivery and operational efficiency.
AI Integration: AI helps streamline project management by automating routine tasks and enhancing decision-making capabilities.
Structured Processes: Implementing structured intake systems minimizes misunderstandings and improves clarity between business and technical teams.
Operational Efficiency: AI tools reduce the need for manual tracking, enabling real-time collaboration and ongoing alignment among teams.
Future of Project Management: Project management is evolving from task tracking to decision structuring, prioritizing clarity over mere activity.
Brian Galardo is a Program and Project Manager in Revenue Operations at Salesforce, where he leads cross-functional initiatives across billing, collections, and customer self-service. He has designed governance frameworks to evaluate AI and agent use cases, partnered with leadership to align innovation initiatives with strategic priorities, and led programs that improve operational efficiency across enterprise systems and business teams.
We talked to Brian about how project delivery is evolving and how he's leading the charge with AI. Here's what he had to say.
How cross-functional initiatives enhance project delivery
I’m a Program and Project Manager in Revenue Operations at Salesforce. My role sits between business strategy and system delivery, helping teams define the problem clearly and reach a decision everyone understands and supports. I’m a transformation leader focused on helping organizations move from AI experimentation to scalable enterprise adoption. At Salesforce, I lead cross-functional initiatives that explore and implement AI-driven solutions across RevOps and Finance operations.
Unlock for Free
Create a free account to finish this piece and join a community of forward-thinking leaders unlocking tools, playbooks, and insights for thriving in the age of AI.
Have an account? Log In
My work sits at the intersection of AI strategy, governance, and operational transformation, helping teams identify high-impact opportunities to improve efficiency, automate workflows, and enhance decision-making.
My background spans supply chain, operations, and enterprise program leadership, giving me a practical perspective on how technology and AI can transform real-world processes. I believe the most successful AI initiatives combine strong governance, measurable business value, and thoughtful change management to ensure adoption across the organization.
How AI transforms tasks in project delivery roles
AI has reduced the time I spend writing updates, summaries, and first-pass plans. I now spend more time validating assumptions, preparing decisions, and aligning stakeholders early. The work shifted from tracking progress to shaping direction.
I use AI to handle the first pass of anything that would normally slow me down. After a meeting, I drop in rough notes or a transcript and have AI turn them into a clean summary with decisions, risks, and next steps. I also use AI to pressure test plans, asking it to poke holes in assumptions or outline edge cases before I bring something to stakeholders.
AI has reduced the time I spend writing updates, summaries, and first-pass plans. I now spend more time validating assumptions, preparing decisions, and aligning stakeholders
The setup is not overly complex. I’ve built a few repeatable prompts and templates that match how I run projects, and I reuse them consistently. It’s less about the tool itself and more about having a clear structure for information output, which allows me to move faster without losing quality.
Why structured intake improves operational efficiency
We introduced a structured intake and prioritization process for operational requests across several finance teams. We categorized, summarized, and prepared submissions before review, so meetings focused on decisions instead of clarification. Review time dropped significantly, and rework decreased because expectations were clearer at the start. The real win was fewer misunderstandings between business and technical teams.
We built this using a Slack workflow for intake, a shared tracker, and AI to standardize how requests showed up. Every request had to be submitted through a structured form with required fields like problem statement, impacted users, and urgency. From there, we used AI to summarize the request, highlight gaps, and flag potential risks before it ever made it into a review meeting.
Step-by-step, it looked like this: requests are submitted in Slack and then flow into a centralized tracker. From there, AI generates a clean summary and surfaces missing context, then business and technical leads asynchronously review them before the meeting. By the time we got to the meeting, we were choosing what to do, not trying to understand what the request even meant.
We cut review meetings roughly in half and saw a noticeable drop in back-and-forth after decisions. I’d estimate rework dropped by at least 30-40% because requirements were clearer upfront. It also reduced the number of follow-up meetings, which was probably the biggest quality-of-life improvement for the team.
We cut review meetings roughly in half and saw a noticeable drop in back-and-forth after decisions. I’d estimate rework dropped by at least 30-40%
How project leads can optimize AI integration
AI does not fix messy processes. It highlights them. When goals and ownership are clear, it helps a lot. When they are not, it creates faster confusion. Project delivery leads need to get the basics right before bringing AI into the process. That means clear ownership, a defined problem statement, and agreement on what a good outcome looks like. If those aren’t in place, AI just accelerates the noise.
AI does not fix messy processes. It highlights them. When goals and ownership are clear, it helps alot. When they are not, it creates faster confusion.
In practice, I make sure every piece of work has a clear owner, a simple definition of success, and a structured way to capture inputs before automating anything. Once that foundation is there, AI adds value instead of creating more confusion.
Recaps, documentation drafting, and request triage work well with automation because they follow patterns. Prioritization, conflict resolution, and change adoption still depend on human judgment. People align with people, not outputs.
Why AI tools are pivotal in modern tech stacks
Our core tools include Slack, Salesforce platform tools, collaborative whiteboards (LucidSpark), structured intake forms, and AI agents for drafting and analysis. AI supports summaries, communications, and early risk identification. Over the past year, we reduced static trackers and moved toward shared real-time context. Fewer manual handoffs drove this improvement.
I mainly use AI tools like Gemini and Slack-native workflows, and Salesforce tools like Agentforce for more structured use cases. I’m not building complex agents from scratch. It’s mostly prompt-based setups and lightweight workflows that handle summaries, draft communications, and early risk flags based on inputs we already capture.
I’m not building complex agents from scratch. It’s mostly prompt-based setups and lightweight workflows that handle summaries, draft communications, and early risk flags based on inputs we already capture.
For example, after meetings, I’ll use Gemini to turn notes or transcripts into a structured recap with decisions and next steps. Within Slack, workflows handle intake and route requests into a shared tracker, where AI helps standardize how each request is presented. Agentforce comes into play more on the Salesforce side when we’re tying data and workflows together at scale.
We moved away from static status decks and spreadsheets updated weekly. Instead, everything lives in shared spaces like Slack Canvas, trackers, and backlog views with continuous updates. This means anyone can see the current state of work, decisions, and risks without waiting for a meeting or report. It reduced the need for status meetings and made alignment more ongoing instead of periodic.
Anyone can see the current state of work, decisions, and risks without waiting for a meeting or report.
We replaced large project plans with shared backlogs, canvases, and decision logs. Teams interact directly with the work instead of waiting for status updates. We transitioned gradually by retiring documents that nobody referenced. Adoption followed because the process felt easier.
How AI reshapes core delivery rituals
We now begin with multiple interpretations of the problem to test assumptions early. Meetings focus on decisions since we prepare context in advance. We validate sooner because risks surface earlier. We monitor execution less and confirm understanding more.
We monitor execution less and confirm understanding more.
We generate those interpretations by running the initial problem statement through AI to produce several angles. For example, I’ll ask it to frame the request from a business, technical, and customer perspective, or to outline what could go wrong if we misunderstood the ask. It’s not about trusting the output; it’s about quickly surfacing assumptions we might not have called out ourselves.
We then bring those assumptions into early conversations with stakeholders to validate scope before committing to anything. Instead of asking “Does this look right?” we’re asking “Which of these is closest and what are we missing?” That shifts the discussion from agreement to clarity. It helps us catch gaps earlier, reducing rework and keeping execution more focused once work starts.
How to streamline delivery with AI-driven workflows
We are experimenting with automated intake triage and communication routing. We targeted areas where delays resulted from unclear ownership rather than technical effort. Early results show fewer escalations and more asynchronous decisions.
We focused on intake triage first because that's where most confusion occurred. Requests came in through different channels, ownership was often unclear, and teams spent time figuring out who should review them. We set up a Slack-based intake workflow where every request follows a consistent structure, then used AI to summarize it, flag missing context, and suggest where to route it.
From there, we automatically push requests into a shared tracker and assign them to the right group based on type and impact. We also improved communication routing; updates now happen in consistent threads or channels instead of scattered messages. It’s still evolving, but we’ve seen fewer “who owns this” conversations and more decisions made asynchronously before meetings.
It’s still evolving, but we’ve seen fewer “who owns this” conversations and more decisions made asynchronously before meetings.
Why decision-structuring is the future of project management
Project management will shift away from tracking tasks and toward structuring decisions. Coordination overhead will shrink while alignment skills become more important. The role becomes more strategic than administrative.
My advice to delivery leaders is to improve the process before adding automation to it. Clear ownership and expectations matter more than speed. Once clarity exists, efficiency follows naturally. Stop writing documents just to show activity. Build workflows where progress is visible without explanation.
Follow along
You can follow Brian's work on LinkedIn.
More expert interviews coming soon on The Digital Project Manager!
