Hi all. Is anyone going beyond generative AI and using agentic AI in their customer marketing & advocacy workflows? I'd love to hear any examples of workflows being powered with agentic AI and what tools you're using. Thanks!
Currently building a few AI flows for clients:
Slack-integrated review requests and surfacing of reviews into Slack
Collecting data from Closed-Won and Closed-Lost reports to inform better selling AND surface any potential reference information
I built a customer evidence agent in Glean where all i need to do is pop in the customer name and it will pull together various sources of data (ie: slack channels, SFDC, gong calls) to fill out a summary of their use case, which is then used to create an outline for a customer story. I then use that info in Chatgpt to create a full story based off the prompt we have for writing stories.
I think these are gen AI prompts, correct Alexie Glover Vera? This is great to share for the collective group to learn from. I would be particularly interested in hearing about any agentic AI workflows (autonomous, no human in the loop) anyone has set up for customer marketing / advocacy.
Maybe Amanda Peacock has ideas. I feel like there's a gap here for what agentic is. Connecting tools and leaving them on a scheduled trigger, to me is agentic. Can you further explain? In Slack, such as Alexie mentioned, a new member can join and after 30 days be asked for a review, all in an automated way. Even pulling from all of the tools that Glean is connect with feels automated. Do you have an example workflow in mind?
Ann Davis mine are agentic
Awesome. Thanks Alexie Glover and Vera.
And there is a difference between triggered AI workflows and agentic ai workflows – agentic ai acts as an independent agent that determines its own actions to achieve a goal while triggered AI workflows with AI follow a predefined path to perform repetitive tasks. Definitely not an expert in this realm by any stretch, but we're already using Gemini, NotebookLM, Glean and other AI tools for generative AI workflows – but he is expecting agentic bots set up to execute all human-powered work. Just trying to learn from what others are doing and see what is possible (although it does not feel realistic to use generative or agentic AI for everything).
Hi! I run customer marketing at Glean and here's stuff I've brainstormed for my team to build out eventually:
Win → Community & References Follow‑Through Agent – When an AE posts a win, the agent waits ~14 days, checks if customer admins/power users have joined Gleaniverse, and drafts a short, personalized invite for their admins.
Win → Customer Proof Harvesting Agent – When someone posts an upsell win, the agent mines Gong, Slack, and win notes to assemble a concise customer reference snapshot (problem, solution, impact, quotable lines, key contacts) and posts it to a customer references channel for someone to review as a potential case study.
Win → Expansion Play Suggestions Agent – For single‑department wins, the agent looks at similar accounts and internal patterns, then proposes 2–3 expansion plays (target departments, offers, and messaging) along with ready‑to‑send email cadences for AEs/CSMs.
Kudos → Reference Request Drafts Agent – When someone posts customer praise or screenshots in a customer kudos channel, the agent identifies the persona/company and returns a tight Slack/LinkedIn reply the owner can use to request a reference or story.
If they are senior enough, the agent also suggests specific upcoming webinars/events where we want speakers.
Community Thread → Content Opportunity Agent – When a Gleaniverse thread gets long or high‑engagement, the agent clusters the key questions and themes, then proposes turning it into a blog, webinar, or Glean Academy lesson with a short synopsis and title ideas.
Release → Adoption Follow-Up Agent – 30 days after a tier 1 feature launches, the agent segments small/med customers into “adopted” vs. “not adopted” and drafts celebrate‑and‑ask‑for‑story emails for the first group and concise “here’s how to get value fast” nudges for the second.
Customer Signal Digest Agent – On a weekly cadence, the agent scans Gong, community, NPS, kudos, and tickets to produce a short cross‑functional digest of new advocates, risks, hot topics, and “people to invite” or “people to ask for quotes.”
That is an amazing list Sandhya Simhan TYSM, I hope you can join us sometime to share on your successes with AI (if that would ever work for you).
Yes! I'm hoping to join one of the AI study groups at some point and share some of them 🙂
Thank you Sandhya Simhan. Love your ideas. Can you share the AI tools you use (outside of Glean, of course)?
My personal AI stack is claude code, vercel, supabase, github pages (as needed) + apis for gemini (nano banana pro) and deegram and elevenlabs as needed.
