Implementation
A large share of incidents sat in generic categories, leaving the model little reliable signal for similarity or routing.
Now Assist: Recovering a Stalled AI Pilot
How Ramisun approaches Now Assist enablement for organisations holding AI entitlements that are not yet delivering — diagnosing the real constraint, remediating the foundations, and deploying against measured baselines.
Direct Answer
What do you do when a Now Assist pilot fails to deliver value?
When a Now Assist pilot fails to demonstrate value , the constraint is usually data quality rather than model performance. The approach is to run a structured readiness assessment across categorisation, resolution notes, knowledge base health, licensing and release position; remediate the specific gaps found; define AI governance with staged autonomy before any capability goes live; capture measurable baselines; then re-launch against high-volume, low-variance use cases where the outcome can be verified.
Readiness AssessmentCategorisation RemediationKnowledge Base Triage
The Scenario
The scenario: paying for AI capability that cannot be shown to work
Typical Sector💻 Technology & SaaS
MarketUnited States
Organisation SizeEnterprise (5,000+ employees)
Engagement TypeNow Assist Enablement
Indicative Timeline12 weeks to production
Platform ScopeServiceNow ITSM Pro Plus
A pattern seen across the market. An organisation purchases Now Assist entitlements during a platform renewal, runs a pilot focused on incident summarisation, and finds that agents do not use the output and handling time does not move. Sponsorship fades and the entitlements risk being written off at renewal.
The instinct is to conclude the technology underdelivers. In most environments the models are performing correctly and reflecting the quality of the records beneath them — a very different problem with a very different remedy.
- Now Assist entitlements purchased but delivering no measurable value
- Pilot ran for three months with no baseline captured beforehand
- No AI governance framework — approval and escalation paths undefined
The Problem
The Model Works. The Data Underneath It Does Not.
Generative features summarise what exists and retrieve what is indexed. A pilot on weak data faithfully reports that weakness back.
Sampling closed incidents showed a majority with resolution notes too thin to summarise meaningfully.
A significant portion of the knowledge base had no owner and no review in over two years, producing confidently wrong retrieval.
The Solution
How Ramisun Approaches Now Assist Enablement
Each capability is switched on only once the data supporting it can sustain it — foundations first, then staged autonomy.
Readiness Assessment
Ran a seven-point readiness assessment covering categorisation, resolution notes, knowledge health, use cases, governance, licensing and release position.
Categorisation Remediation
Rationalised the category tree and used Predictive Intelligence to back-fill historical records — ML is well suited to this specific task.
Knowledge Base Triage
Retired unowned and duplicate content, assigned owners and review dates to what remained, and restructured for retrieval rather than browsing.
Resolution Note Standards
Introduced a minimum closure standard with template prompts, improving note quality within the first quarter.
AI Governance Framework
Defined approval gates, logging, escalation paths and data boundaries before any capability went live — recommend, then supervised, then autonomous.
Baselined Re-Launch
Captured handling time and deflection baselines, then deployed against high-volume, low-variance use cases with weekly measurement.
Target Outcomes
What This Approach Is Designed to Achieve
No baseline captured
MeasuredValue Demonstrable
Value Demonstrable
Handling time and deflection now tracked weekly against a pre-deployment baseline.
Degraded uncategorised
RemediatedCategorisation Coverage
Categorisation Coverage
Historical records back-filled and the category tree rationalised and owned.
No governance
3-stageAutonomy Ladder
Autonomy Ladder
Recommend → supervised → autonomous, with each stage gated on evidence.
Entitlements at risk
In productionLicence Utilisation
Licence Utilisation
AI capability the business was already paying for is now in daily operational use.
Related Case Studies
More ServiceNow Implementations
Trusted by Global Enterprises
We partner with industry leaders to drive meaningful transformation.
Testimonials
What our clients say
★★★★★
“Ramisun didn't just implement ServiceNow — they redesigned how our teams work. Their AI agents now handle most routine requests.”
VP of IT OperationsEnterprise Healthcare Network
★★★★★
“Their marketplace expertise was the difference. We went from prototype to a certified ServiceNow Store listing our team could sell.”
Co-founder & CTOAI-Funded Startup, Silicon Valley
★★★★★
“Transparent, agile, and genuinely business-first. Every sprint ended with something we could show our executives.”
Director, Digital ProgramsFinancial Services Firm
★★★★★
“The CMDB work alone paid for itself. We finally trust our configuration data enough to automate against it.”
Head of Platform EngineeringGlobal Manufacturer
★★★★★
“Hypercare was real hypercare. Go-live week felt boring in the best possible way.”
IT Service OwnerTelecom Operator
★★★★★
“Ramisun didn't just implement ServiceNow — they redesigned how our teams work. Their AI agents now handle most routine requests.”
VP of IT OperationsEnterprise Healthcare Network
★★★★★
“Their marketplace expertise was the difference. We went from prototype to a certified ServiceNow Store listing our team could sell.”
Co-founder & CTOAI-Funded Startup, Silicon Valley
★★★★★
“Transparent, agile, and genuinely business-first. Every sprint ended with something we could show our executives.”
Director, Digital ProgramsFinancial Services Firm
★★★★★
“The CMDB work alone paid for itself. We finally trust our configuration data enough to automate against it.”
Head of Platform EngineeringGlobal Manufacturer
★★★★★
“Hypercare was real hypercare. Go-live week felt boring in the best possible way.”
IT Service OwnerTelecom Operator
✦ Get Your Free Consultation
Ready to Transform with ServiceNow?
- Free ServiceNow scoping session, no generic pitches
- Tailored roadmap mapped to your exact processes
- Transparent pricing — no hidden fees
- Response within 1 business day
“
An implementation is only as good as the consulting behind it. When the platform mirrors how your business actually works, adoption follows — and so does ROI.Vinnay Nigam, Founder & CEO, Ramisun
100+Implementations
229%3-yr ITSM ROI (Forrester)
8–16 wksTypical Go-Live
Confidential. We never sell data or send spam.
Security & Compliance
Enterprise-grade trust, built in
Responsible AI adoption needs guardrails. Our delivery model is designed around security, auditability, and compliance from the first workshop.
SOC 2 Aligned
Delivery processes mapped to SOC 2 trust principles.
ISO 27001 Practices
Information security management across every engagement.
GDPR Ready
Privacy-by-design data handling and processing controls.
AI Governance
Policy controls, auditability, and human oversight for every AI agent.
