Employee Onboarding in the Age of AI

Posted on by Brendan

I am four weeks into a new job where I actually boomeranged back to the company I was at 12+ years ago. This is my first time onboarding to a new company in the age of AI since I hadn’t onboarded somewhere since 2022. 

I decided to boomerang back to Seer Interactive for two reasons. First, for its growth and learning culture. It aligns deeply with who I am to my core. And also given the time we’re in with search and broader AI technology, I wanted to prioritize learning over anything else. I knew I’d get it here. 

Second is I wanted variety. What does that mean? To me, it means the variety of what I see (i.e. challenges, problems, opportunities), who I work with, and the pace in which we’re moving. After 11+ years driving organic search strategy in-house, I knew that beat so well it felt like it wouldn’t be much variety. 

It’s been great so far for a lot of reasons but one I want to share today is because of AI. 

Onboarding to a new company that has proper AI infrastructure in place (i.e. Seer) has been amazing. The point I am making is AI clearly drives ROI for employee onboarding. Assuming the company has the infrastructure in place. 

Within the first four weeks, here’s a few AI things that have helped me ramp quickly: 

  • Daily Briefings: my daily briefing agent helps synthesize context, priorities, and what matters across my portfolio. 
  • Client Pulse: Getting up to speed on client sentiment, nuances, historical goals, and current standing quickly versus sifting through old threads and having tons of internal meetings. 
  • Quick sounding board for acronym checks, team structure and basic questions. 

All the above preserves internal team members time, lets me have more strategy conversations with people, and allows me to add value faster. 

So a faster ramp within a company that’s made the right AI investments, doesn’t just benefit the business; it accelerates an employee’s ability to lean into growth, learning, and impact on day one.

How to measure this for anyone who needs to see the numbers? 

  • Time to First Value: think of something like date of first shipped deliverable, test launch, test read. Things that help drive actual impact. Metric: Date of First Shipped Deliverable – Start Date
  • Time to Full Utilization / Standard Output: Days until the hire reaches the target output expected of a tenured peer in that role. Metric: Date Target Output Sustained for 30 Days – Start Date
  • Time and Cost of Existing Employees Training New Hires: When tenured staff train a new employee, the true cost consists of direct time plus opportunity cost (billable or strategic work not completed).

This is a clear AI ROI case to me.

Thoughts on Dots

Posted on by Brendan

This week, OpenAI announced Dots. What they called, “remarkably capable, always-on agents built to handle everything.” And Dharmesh Shah had this to say about it:

I couldn’t agree more. Whether learning, working, collaborating, or parenting, my default mode is connecting dots. Just last week in a meeting, I referenced a report template that tied directly back to an all-hands release meeting. It took a moment for the person I was meeting with to bridge that gap, but once they did, the path forward was instantly clear. Sometimes I spot those intersections faster than others; other times, teammates help me connect them.

One of my personal goals is communicating those synthesis moments more effectively. That’s why tools like Dots excite me. Not just to automate tasks, but to help scale how we synthesize and share context.

I’m going to keep learning and leveraging AI to map those patterns faster and in ways my own brain can’t match alone. My one prediction? “Dots” won’t achieve the household-name stickiness of ChatGPT, but the utility underneath it is the real deal.

Workplace Agents vs. Personal Apps

Posted on by Brendan

After two years away from the blog, I’m trying to build the writing habit back into my weekly routine. I am starting with two different product marketing announcements that caught my eye this week:

Taking these posts in isolation, one wants to be your office manager; the other wants to be your personal lifestyle assistant.

I’m particularly interested in Claude’s “quiet” social post. In a world of loud announcements, a low-stakes social post is a genius way to gather real-time signals. They aren’t just announcing a feature; they are testing a hypothesis: Do people actually want AI in their personal lives, or just their work?

The Takeaway: Don’t get distracted by the pace of launches. The real work is following customer behavior and having the guts to ship “quiet” experiments mapped to big bets to see what sticks, like these companies are doing themselves.

Learnings & Thoughts on RAG

Posted on by Brendan

I enjoyed this essay on X about RAG and learned a lot. 

Part of LLMs advancing is a technology called RAG (retrieval augmentation generation). RAG is an evolving technology in the world of information retrieval for LLMs but it has problems. 

  1. For large models, freshness of source data for time sensitive queries is a limiter right now. A LLM needs to update it’s data source and put it back through the AI model and know to site the new source when a user writes a prompt or query. 
  2. RAGs has a similar problem for smaller models too. Amongst a corpus of internal documents with perhaps conflicting information because there are drafts how does the RAG know which once to use? 

I think one important quote from Aaron that ties to a post from Google is this: “The AI’s answer is only as good as the underlying information that you serve it in the prompt.” Google’s Liz Reid wrote: “With our custom Gemini model’s multi-step reasoning capabilities, AI Overviews will help with increasingly complex questions. Rather than breaking your question into multiple searches, you can ask your most complex questions, with all the nuances and caveats you have in mind, all in one go.” 

This is something we started talking about 12+ months ago at this point where the better the prompt the better the answer. It’s clear search engines and LLMs are trying to nudge humans to change search behavior in this way to improve answers and reduce hallucinations. 

The question on my mind is, will consumers change their search behavior to more complex searches or will tech companies need to find other ways to improve source identification during RAG to improve output? 

The Least Sexiest Thing

Posted on by Brendan

I find it human nature to chase the shiny new thing. Or make something overly complex and over-engineer it. I was reminded last week that sometimes the least sexiest thing is most useful. In my case, it was a flat priority list of projects for Q2. From highest to lowest priority, I was meeting with the Engineering Manager and Product Manager on my team to discuss the prior and upcoming sprint. It wasn’t a fancy strategy image or something complicated in Jira. 

Our discussion was effective because we were able to see and discuss accomplishments & learning, where we’re going next, and what is delayed, which is exactly what you need to find ways to discuss throughout a quarter. 

It was a simple flat priority list of ~15 projects in buckets from P1, P2, and P3. Sometimes the least sexiest thing is most useful.