Signal Brief #18: Decagon turns proactive agents into compliant campaigns + 4 more
5 new AI tools for B2B SaaS product managers — each with what it does, why it matters for PMs, and a direct link.
|
SIGNAL BRIEF
Issue #18 · Monday, September 7, 2026 · 5 AI tools for B2B SaaS product managers
|
|
Five AI tools worth your attention today — what each one does, why it matters if you run product at a B2B SaaS company, and where to look. No sponsors, no noise. |
|
Decagon on Tuesday extended its proactive-agents push with Campaign Composer, announced September 2 on the company blog and timed to Decagon Dialogues 2026. Earlier this year, Proactive Agents gave Decagon’s customer-service agents the ability to reach out first instead of waiting for the customer to show up; Campaign Composer turns that ability into full outbound programs. A campaign does not stop at a single touch — it follows up over days or weeks until the outcome is met: remind a customer about an upcoming payment, check in on a prescription refill, sell an upgraded seat on an upcoming flight, picking the case back up where it left off each time. The company’s framing is blunt: building a tool that can place a call or send a text is the simple piece; the hard part is running outbound at scale in a regulated business — configuring the outreach strategy, staying compliant with state and federal regulations, and optimizing based on what works across millions of attempts. Campaign Composer is the control center for all of it, configured by your team in one place. Channels, the hours campaigns are allowed to run, retry logic, attempt caps, and cadence live in the product — yours to change without an SDK build or a ticket to a vendor. Mid-conversation, when a customer says now is not a good time, the agent books a scheduled callback, so the touch becomes a planned next step instead of a dropped thread. The compliance layer is configurable the same way: suppression based on contact-specific fields, opt-out and wrong-party flags, dispositions, and built-in PII and PHI redaction — the controls that let regulated customers meet Reg F and Reg X at scale on a platform that also meets PCI, SOC 2 Type 2, HIPAA, GDPR, and CCPA. The release ships throughput visibility with it: call distribution spreads attempts across the business-hours window instead of firing everyone at once, surfacing which times of day connect best, and Decagon’s Duet reads campaign results to identify the best times to reach people, which messaging performs, and where the funnel loses them. The named deployments mark the territory: Hertz runs outbound car-rental return confirmations, resolving late returns on the spot and recording every outcome; a prescription-refill agent navigates a pharmacy’s phone tree on its own, under HIPAA, and tries again later if it must. Every campaign carries its own definition of a good outcome — and every outbound conversation tells the next one something. Why it matters for PMs: Support agents that only answer are table stakes; the frontier is agents that initiate — and Decagon just published the reference architecture for doing it without becoming spam or a compliance incident. Three lessons. First, compliance as product surface rather than policy document: suppression lists, attempt caps, operating hours, wrong-party flags, and redaction are configuration your operations team changes in the product, not a legal review bolted on afterward. If any agent on your roadmap initiates contact — renewal nudges, churn saves, onboarding outreach — spec it against this bar; expect procurement to ask for the equivalent of your Reg F controls. Second, note where the configurability landed: cadence and retry logic sit with the team that runs the campaign, no engineering sprint behind every change — the same authority-to-the-operator pattern as #15’s control surfaces and #8’s LaunchDarkly configs, applied to outreach. Third, “each with its own rules and its own definition of a good outcome” is the quiet design decision that matters: outcome-defined campaigns are what separate production agents from demos, the same predicted-versus-actual discipline #16’s Product Console brought to features. The named verticals — collections, refills, rentals — are not B2B SaaS, but the pattern transfers directly: follow up until the outcome, inside rules a human set, with every attempt feeding the next. |
You have read tool 1 in full. 4 more are behind the trial.
Each one below has the same three things the tool above did — a one-sentence description, the full why-it-matters-for-PMs line, and the direct link:
- Profound gives the AI marketer an operating memorytryprofound.com
- Userlens measures adoption, not opensuserlens.io
- Onset lets the agent that wrote the code publish the release notesonset.io
- TrackMCP is Google Analytics for your MCP servertrackmcp.com
Subscribe and every issue lands in your inbox in full — each tool's description, why-it-matters line and link, starting with #22.
Start your 7-day free trial$10/month · 7-day free trial · no card required. Or get a free sample issue by email in the form below — tool 1 in full, plus one complete gated write-up: ProdPad from #22. The rest are named.
Already a subscriber? Read the complete issue