AI is changing B2B software. I work in product, pricing and R&D to make sure it strengthens the business — not just the cost-of-goods.

“We know we have to ship AI – but the roadmap is filling up with hype features no real customer actually needs.”

• →Feature discovery and prioritization grounded in real customer value
• →Build / Buy / Wrap – informed decisions, not gut feeling
• →Architecture reviews: RAG, agent patterns, evaluations, guardrails
• →POC-to-production paths that don't get stuck in prototype limbo

From operator practice: roadmap clarity before engineering starts building.

“Our flat fee is eating LLM costs. But usage pricing scares customers off. What now?”

• →Cost-of-goods modeling per feature and per customer segment
• →Pricing models: usage, outcome, hybrid, tiered – matched to buying center
• →Packaging strategy: what's in the base, what's upsell, what's its own SKU
• →Rollout plan including existing-customer migration and A/B tests

From operator practice: pricing models that survive contact with Finance and Sales.

“My team uses Copilot, but we aren't measurably more productive. I hear of 3× output elsewhere – what are they doing differently?”

• →R&D workflow audit: where AI actually helps today, where it only appears to, where it doesn't
• →Tooling stack: Claude Code, MCP servers, subagents, eval pipelines, CI integration
• →Org design for agentic teams: roles, skill profiles, accountabilities
• →Change management and skilling – not just another 'AI pilot' inside the enterprise

From operator practice: measurable productivity gains in pilot squads, then rolled across the org.