AI Agents / Customer Signals Agent

AI agent · Live on the ServiceNow Store

Transcript in. Evidence-backed needs out.

Customer Signals Agent reads your customer meeting transcripts and turns them into structured, evidence-backed customer needs, pain points, and personas, with the customer's own comments kept as evidence for the people building the solution.

Book a demo →Read the docs

Customer Signals AgentCustomer Signals

READS

Meeting transcript (plain text)

Conversation details: customer, date, owner, and primary contact

The customer needs you already have

Customer Signals AgentSuggestions staged for human review

PRODUCES

Candidate customer needs

Pain points and feature opportunities

Personas

Evidence: quotes and sentiment, attached to the right need

The problem

The constraint is no longer building software. It's deciding what to build.

AI-assisted delivery has pushed the effective product-manager-to-engineer ratio from roughly 1:8 to 1:20, and the evidence needed to decide is scattered and unstructured. When a release ships, nobody can reliably say which customer asked for it, in which conversation, and in what words.

“I don't have the time after a customer meeting to synthesize the call into a crisply defined customer need. I'm left with notes, recordings, and a hope that I'll remember what the customer said.”

Needs are buried

Customer needs live in call recordings, meeting notes, chat threads, enhancement requests, and spreadsheets.

Re-keying feedback

Product managers spend as much time re-keying feedback into systems as they do talking to customers.

Context gets stripped

Builders receive requirements without the customer statement that motivated them, so they ask the product manager or go without.

The approach

Conversations become structured product inputs, automatically.

Create a conversation record, add the transcript, and select Analyze. The agent reads the entire conversation, not just the parts someone wrote down, and proposes candidate needs. Every need carries its verbatim evidence, and evidence that matches an existing need strengthens it instead of creating a duplicate. Product managers then review, prioritize, and relate needs to work items.

PRIMARY OUTCOMES

→Capture more customer needs

→Capture the details of each need automatically

→Connect customer evidence to work items so builders understand the need

Create a conversation

Track the customer, date, owner, and primary contact for the meeting.

Add the transcript

Paste or attach the plain-text transcript from your meeting recording.

Select Analyze

The agent extracts personas, needs, and evidence with sentiment and timestamps.

Review and confirm

Confirm or reject each suggestion, then relate needs to work items.

Capabilities

What the agent does for you.

Transcript to structured signal

Proposes candidate customer needs, pain points, feature opportunities, and personas from the whole conversation.

Evidence in the customer's words

Actual customer quotes or close paraphrases are captured as evidence, with sentiment, and attached to the need they support.

Matches existing needs

When a need already exists, new evidence is added to it, strengthening the case instead of creating a duplicate.

Traceability through build and release

Evidence keeps its extraction run, prompt version, model provider, and source. When someone asks why a feature was built, the answer is one click away.

Human review by design

Every suggestion is staged for a person to confirm or reject before it becomes a need.

Built on ServiceNow AI

Runs on ServiceNow AI Agent Studio and Now Assist Skill Kit in your instance, using the model provider your administrator approved.

Get started

Install it from the ServiceNow Store.

Check the requirements, install on a sub-production instance, and analyze your first transcript. The installation guide walks through every step.

Installation guide →Book a demo

SYSTEM REQUIREMENTS

ServiceNow PlatformAustralia Patch 5+ or Zurich Patch 12+

ServiceNow Otto for Platform12.2.0+

LicenseA licensed ServiceNow Otto product, such as Platform, ITSM, or App Creator

LLM providerAn approved provider configured by your instance administrator: Azure OpenAI, AWS Claude, or Google Gemini

InterfaceServiceNow Otto (Now Assist) Panel enabled

Rolesnow_assist_panel_user · x_ic5l_csa.user

Customer Signals

From the blog.

PRODUCT MANAGEMENT

The New Product Bottleneck: What Happens When One PM Can No Longer Keep Up With What Their Team Can Build

AI has pushed the effective PM-to-engineer ratio from roughly 1:8 to 1:20. Here is what breaks in roadmapping and requirements when the constraint shifts from building to deciding, and how to fix the input side of the problem.

Oct 15, 2026

AI Makes Builders Faster. Customer Understanding Makes Them Better.

As AI helps engineering teams produce software faster, the bottleneck moves. The limiting factor is less about writing code and more about deciding what should be built in the first place.

Jul 15, 2026

PRODUCT STRATEGY

Rise of the ServiceNow Builder Persona

Powered by Now Assist for Creator, a single person can now build applications and workflows that once required a full team. IC5 Labs' PBE answers the question that comes first: what should you build?

May 17, 2026

See Customer Signals Agent on your own instance.

A 30-minute session with the team that built it. Bring a real customer conversation, or use our demo data.

Book a demo →View on the ServiceNow Store