From the Local Government edition of September 20, 2026
Denver Water seeks Genesys-integrated internal AI agent assist (RFP 10575)
Denver Water · Municipal/regional water utility customer care and operations · Denver metro, Colorado, United States
- Publisher
- Bidscope
- Original publication
- Bidscope summary of Denver Water solicitation 10575 posted Aug 31, 2026 (due Sept 30, 2026); official opportunities table corroborates listing
- Source retrieved
- 2026-09-21
What happened
Bidscope’s summary of Denver Water solicitation 10575 (Customer Experience AI Chatbot / internal AI Agent Assistant), still open with a Sept. 30, 2026 due date per both Bidscope and Denver Water’s official opportunities table, seeks professional services to deploy an internal assistant for roughly 70 customer-care users. Requirements summarized include Genesys Cloud CX 3 integration; read access to SharePoint, DenverWater.org, and Oracle CC&B; Phase 1 primarily read-only with real-time retrieval, source citations, call summaries, recommendations, analytics, and human review; Phase 2 controlled human-confirmed write-back to Genesys and CC&B; security, privacy, accessibility, records, AI governance, auditability, U.S. data residency; and milestones toward April 2027. Why now: proposal deadline is days away, making this a timely utility procurement control signal for agent-assist patterns.
Why it matters
Near-deadline U.S. utility RFP showing preferred agent-assist pattern: citations, human review before system-of-record writes, and Genesys/CIS integration—highly transferable to city 311/customer-care AI buys.
Evidence and measured results
Bidscope opportunity text: ~70 users; Genesys Cloud CX 3; SharePoint, DenverWater.org, Oracle CC&B; Phase 1 read-only with citations and human review; Phase 2 human-confirmed write-back; due Sept 30, 2026; posted Aug 31, 2026. Denver Water official table: 10575 Customer Experience AI Chatbot, 08/31/2026–09/30/2026.
Limitations and uncertainty
Secondary aggregator summary—not the downloaded official RFP PDF from Denver Water’s procurement portal; verify all requirements in source documents before bidding or advising.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-09-21; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Customer-care agents need faster, cited answers across knowledge and account systems without unsupervised changes to the system of record.
- Stakeholders
- Customer care leadership
- IT/telephony (Genesys)
- CIS/CC&B owners
- Security/privacy/records
- Procurement
- Discovery questions
- Which systems are read-only in Phase 1 vs write in Phase 2?
- What citation and human-confirmation evidence is required?
- Where must data reside, and what certifications substitute for FedRAMP?
- What is the April 2027 milestone definition of done?
- Value hypothesis
- Cited agent-assist with gated write-back can reduce handle time if Phase 1 quality is proven—without guaranteeing FTE reduction stated in the RFP.
- Potential engagement
- Phase 1 read-only agent-assist pilot design against Genesys plus knowledge/CIS sources, with explicit no-write acceptance tests.
- Unsupported claims
- Guaranteed handle-time or FTE savings
- Unsupervised CIS updates are in scope for Phase 1
- Aggregator text equals the signed RFP PDF
- Public customer-facing chatbot is the primary deliverable
Pre-sales engineering
Role takeaway
- Fit
- Internal CSR AI agent assist on Genesys with retrieval from knowledge sites and read/write-gated CIS integration.
- Architecture and integration
- Genesys desktop → assist service → SharePoint/web/CC&B retrieval with citations → human review → optional confirmed write-back of notes/codes/records.
- Prerequisites
- Genesys Cloud CX 3 access
- CC&B/SharePoint connectors
- Citation UX
- Audit logging
- Constraints
- Initial ~70 users
- Phase gates for writes
- U.S. data residency
- Verify against official PDF
- Security
- Enforce RBAC, encrypt data, retain auditable confirmation of writes, and minimize training on customer account content per policy.
- Proof of value
- Citation precision sample
- Agent override rate
- Write-back error rate after Phase 2
- Time-to-knowledge vs baseline
Delivery
Role takeaway
- Work
- Stand up Phase 1 assist, train CSRs on verification, run security questionnaires, then carefully enable Phase 2 write-back.
- Dependencies
- Genesys admin
- Oracle CC&B team
- SharePoint owners
- Security/records review
- Vendor implementation
- Ownership
- Customer Care owns agent adoption; IT owns integrations; security owns go-live for writes.
- Skills and adoption
- Train agents to treat AI drafts as suggestions requiring citation checks before customer commit or CIS write.
- Governance checkpoints
- Phase 1 acceptance
- Write-back enablement CAB
- Accessibility/security packet review
- April 2027 milestone
- Acceptance criteria
- Phase 1 has no CIS mutation
- Every Phase 2 write shows human confirmation
- Citations present on knowledge answers
- U.S. residency attested
- Risks
- Aggregator/PDF mismatch
- Silent account updates
- Poor citation quality
- Scope creep into public chatbot before controls mature
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Prefer agent-assist beside the CSR desktop with citationed retrieval and gated write-back over unsupervised customer-facing bots writing to CIS.
Governance
Who approves, reviews and stays accountable for outcomes?
Encode phase gates (read-only then human-confirmed writes) and AI-governance/audit requirements in the solicitation itself.
Security and privacy
What data, permissions and controls need testing?
Utility account data and U.S. residency constraints dominate; require RBAC, retention, and review before CC&B mutation.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Internal agent-assist can help CSRs serve accessibility needs faster if guidance is verified; RFP accessibility requirements should be tested, not assumed.
Procurement
What should contracts, pricing and exit terms secure?
Utility AI buys should hard-require phase-gated write-back, citations, residency, and auditability in the solicitation—as this package outlines.
Operating model
Which teams own the service once it runs?
Keep humans as the committers to CIS; AI drafts notes and suggestions inside the existing Genesys work pattern.
Publication history
- 2026-09-20Local Government · Issue 154 resources
Stable resource ID: denver-water-cx-ai-agent-rfp-20260930