United StatesTechnology and telecom2026 – ongoingCox Communications
Claude-based marketing and sales agents and lead validation for the B2B funnel
Cox Communications reported a 7x return on its first year of AI investment across its B2B funnel
Problem
B2B growth at Cox had slowed and one-size-fits-all campaigns no longer worked. Field sellers researched prospects by hand, and lead lists were expensive to build and inaccurate.
Approach
Cox worked with Anthropic and Accenture and ran Claude through Amazon Bedrock. A marketing agent, CAMI, writes campaign briefs and content for 120 micro-industries and eight customer segments. A sales agent, SAMI, builds call prep, scripts and emails. A lead pipeline uses Claude to research and verify leads, with Haiku, Sonnet and Opus matched to task difficulty. An agentic go-to-market pilot that manages the funnel from first signal to contract close is still running.
Results
Takeaway
Cox matched smaller and larger models to the task, which it tied to the lower cost of validating leads.
Categories
- Size
- Large enterprise
- Industry
- Technology and telecom
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Generative AIAI agents
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
Anthropic describes how Cox Communications, with Accenture, built marketing and sales agents and a lead validation pipeline on Claude. It reports a 7x return in the first year, 55% faster speed to market, 40% higher campaign efficiency, 86% lower lead validation cost and lead accuracy of 97%. The agentic go-to-market work is still a pilot. The page carries no date; the Internet Archive first saved it on September 15, 2026.
Full details are in the original sources.