At its Insiders event, UserGems unveiled new Data and Intelligence Agents alongside the UserGems MCP, plus customer results from Frontify and Scrappy ABM showing what an AI go-to-market brain looks like in production.

SAN FRANCISCO, [September 15th 2026], UserGems, the AI command center that turns buying signals into pipeline, today announced two product launches at its UserGems Insiders event: new Writing and Research Agents, and the UserGems MCP, a headless interface that exposes the full command center to any MCP-compatible AI. Together, they let sales and marketing teams build, launch, and audit always-on outbound and ABM campaigns by prompting Claude or ChatGPT directly, rather than working inside a separate platform. The event also featured customer sessions from Frontify and Scrappy ABM on what an AI go-to-market brain looks like once it is actually running.

How to build and scale an AI go-to-market brain

Christian Kletzl, CEO and co-founder of UserGems, opened the event by laying out why go-to-market teams need a "brain" connecting their systems of record to their systems of action in the first place. Kletzl pointed to internal data showing that individual signals, such as a past champion changing jobs or a website visit, matter less on their own than they do in combination: accounts with six or more active signals convert to opportunities at 19 times the rate of accounts with only one or two.

"If you react to only the first signal that fires, you're missing the best opportunities," Kletzl said. He outlined the three components a GTM brain needs to work: business context (a company's own product usage, opportunity, and engagement data), continuously refreshed third-party signals (job changes, intent, hiring patterns, tech stack, custom AI research), and an intelligence layer that scores, prioritizes, personalizes, and orchestrates across sales and marketing. Kletzl compared building that infrastructure from scratch to Waymo's early decision to buy Jaguars rather than build its own cars: teams are better served buying the data and identity-resolution layer and spending their own time on the campaigns that sit on top of it.

New Data and Intelligence Agents, and the UserGems MCP, demonstrated live

Taylor Vo walked through the new Writing Agent and Research Agent by building a revive-closed-lost campaign live: layering signals such as new hires, promotions, and website visits on top of a closed-lost account list so the Scoring Agent could rank which accounts and contacts to prioritize, then using the Writing Agent to draft outreach that automatically pulled in the right signals, value propositions, and customer stories for each recipient's persona and industry.

Vo then rebuilt the same campaign a second time, this time by prompting the UserGems MCP inside Claude. Where most MCP integrations only return data for a rep to act on, the UserGems MCP asks clarifying questions, walks the user through configuring an audience and its workflows and actions, and builds and activates the campaign itself, drawing on the same seven playbooks (including past champions, competitive takedowns, and closed-lost revives) available inside the platform. Vo showed the MCP surfacing an account overview, a prioritized list of contacts to reach out to first, and the Writing Agent's own instructions, all from within Claude, without opening a separate tab.

Frontify: building an internal GTM brain, and buying the piece that can't be built

Alex von Stegmann, GTM Engineer at Frontify, described the four-tier build-versus-buy framework his team uses to decide what to build internally versus buy: personal tools built by individuals, team-recommended tools that are useful but not business-critical, vendor tools required for the whole team to run reliably, and systems of record that need near-perfect uptime.

Frontify has built substantial internal infrastructure on that framework, including Revi, a Snowflake-powered agent that answers leadership's pipeline questions (win rate, sales cycle length, and similar KPIs) with verified, source-checked numbers rather than AI-generated approximations. But von Stegmann said the piece his team can't replicate internally is the ever-changing external context: which accounts resemble Frontify's best customers, and when those accounts are actually in-market. That's the function UserGems' signal-based account and contact scoring serves for Frontify today.

Since adopting it, Frontify's SDR team is more than three times as likely to book an intro with an "A"-scored account than with a "D"-scored account, and roughly 40% of the team's pipeline now comes from UserGems-driven campaigns. Frontify's next step, von Stegmann said, is layering UserGems' external signals directly on top of its internal call and account intelligence through the UserGems MCP.

Turning the brain into an always-on pipeline engine

Trinity Nguyen, CMO of UserGems, presented the framework her team uses internally to turn an AI go-to-market brain into pipeline: Instant campaigns (three to six scorching-hot, company-specific signals such as a past champion changing jobs, that warrant immediate one-to-one outreach the moment they fire), Compounding campaigns (dozens of quieter signals stacked together, such as a competitor's customer hiring for an ABM role, then attending a webinar, then engaging with an ad, as a proxy that an account is moving toward being in-market), and One-off campaigns for events, competitive moments, or new product launches.

Nguyen said Compounding campaigns, while less attention-grabbing than Instant ones, generate the majority of pipeline volume because they draw on far more signals. She cited an internal example in which HubSpot, an existing UserGems customer, launched a new AEO product and stood up a fully targeted, three-audience campaign in a single day, work that would typically take about a month, by creating a new persona and intent signal inside UserGems. Nguyen also described a public-company customer that used the UserGems MCP to identify, de-duplicate, and prioritize its most relevant customer stories against its own scored account list, then automatically apply the results across every active outbound campaign.

Since UserGems fully operationalized this framework internally, Nguyen said, 70 to 75% of UserGems' own pipeline has come from outbound and account-based programs, at the same headcount as before.

31 opportunities in three days from an event playbook

Mason Cosby, founder and CEO of Scrappy ABM, presented results from a client event where a fully automated, signal-driven pre-event program generated 31 opportunities in three days. Cosby's framework maps five core audiences that exist at nearly every event (last year's attendee list, this year's registration list, a geographically targeted list, social posts about attending, and other sponsors at the event), then builds a conversion path connecting each event asset, including free tickets, a speaking session, an executive car service, a happy hour, and a "half a gift" play in which attendees received an empty case for a gifted pair of AI glasses and had to visit the client's meeting suite to collect the actual gift.

That single play brought 80% of car-service attendees to the meeting suite for a demo. UserGems' Writing Agent personalized outreach to each of the five audiences automatically, based on the signals and offers relevant to that segment, and the client's team booked meetings directly out of those conversations rather than following up afterward.

About UserGems

UserGems is the AI command center that turns buying signals into pipeline. We help sales and marketing know exactly who to target, when, why, and what to say for outbound and ABM, at scale.

‍