When we're not running Bili, the founding team are market-leading due diligence and strategy consultants to VC investors. We get asked a lot (and I mean, a lot) about how you should be using AI.

This is what we're seeing in the real world:

  • Incremental pockets of AI - first drafts LinkedIn posts, document summaries, etc. Not revolutionary, but helpful to time-poor founders and soloists.
  • High-value niche models - e.g. machine vision projects, medical treatment path screening, and so on.
  • Tightly-bounded cloud LLMs with guardrails and lots of QA - e.g. "Take this data and explain it in user-friendly terms", or "Which of these 10 teams should answer the user's question?". Companies are chaining many micro-prompts, but few (if any!) we've seen are racking-up LLM tokens on big-context conversations.
  • Lots of dev assistance, freeing-up time to build ambitiously, focus on product fit and cut down time spent writing tedious but critical QA tests.

"It's about teams replacing slog-work with LLM-powered digestion to free-up valuable time for creative and strategic planning."

— Rich Kershaw, Co-founder and CEO, Bili

Here's what we're not seeing.

  • One-person SaaS development teams scaling whole platforms with Claude Code, Cursor or Lovable.
  • 10x or 100x developers - we absolutely see effective devs delivering more code, but the idea of a 100x dev ignores the fact that the majority of a dev's time will be spent on ticket refinement, requirement gathering, design and architecture, R&D to test approaches to a problem, etc.
  • Startups laying-off many developers in favour of LLMs. Lots use LLMs to support teams, but in 2026 dev teams are still broadly resembling teams in 2019.
  • Executive teams "supercharging" their work with AI. We struggle to find a team revolutionising *anything* - it's all about incremental support.
  • Big "we need AI everywhere" projects with staggering results - many projects have already stagnated and fallen to the bottom of the priority list.

Here outside the bubble, OpenClaw is a tech toy, beyond the ability (or need!) of the average exec. Agentic AI is expensive, needs hands-on time, sometimes impressive but prone to blind alleys... and often just means a shift from working in Office to working in ChatGPT.

Instead, it's focused features, hidden in HubSpot workflows and improved reports. It's about teams replacing slog-work with LLM-powered digestion to free-up valuable time for creative and strategic planning.

What're we doing about this?

First, we're doing what we've always done at Leckie-Kershaw and Diligency Group - supporting investors and their portfolios with due diligence, strategic thought and guidance on getting good change and product governance setup.

Secondly, we're building a platform - Bili - that uses targeted AI and knowledge from 200+ projects and thousands of team interviews to help founders, investors and their teams focus on performance, risks and human capability building. We're fundraising now and going to market later this year.

About Bili
Bili is an AI-powered business intelligence platform that gives growing businesses and their investors clarity to act and confidence to lead. It integrates with existing tools and data sources to surface insights, automate reporting, and support better decisions — without the overhead of traditional BI solutions. Bili will shortly be launching in private beta ahead of a full launch in 2026.