Description
Summary AI & Automation Consulting Researcher We're looking for an experienced consultant or researcher with strong knowledge of business transformation, AI and automation, to build a structured knowledge base of how experienced consultants identify problems, choose solutions and implement AI and automation. This is a research and methodology role. You'll do the research and write the knowledge base yourself. You won't be interviewing other consultants to get it. The thread running through the whole thing is the path from business problem to diagnosis to decision to technology to implementation to measurable outcome. We want that reasoning captured in a way someone else can pick up and apply. 1. Consulting reasoning and frameworks Research and build out the practical frameworks an experienced AI consultant uses when assessing a business. That includes: Structuring the audit * How to structure an AI readiness audit from start to finish, from the first conversation with a business through to the final recommendation * What to look at, in what order, and what to ask at each stage * What an AI readiness framework should actually assess, covering areas like data, systems, processes, people and appetite for change * How to score or grade readiness, so two businesses can be compared on the same basis Diagnosis * How to assess a company's processes and operations * How to identify bottlenecks and root causes Opportunities and prioritisation * How to spot worthwhile AI and automation opportunities * How to decide which ones should be prioritised * How to quantify time, cost and revenue impact Choosing a solution * How to work out whether AI, traditional automation, process redesign, buying software or building software is the right answer * How to assess feasibility, risk and implementation complexity * When AI or automation shouldn't be used at all Value * How to calculate expected ROI * How to measure the actual value after implementation Consulting principles Alongside the frameworks, we want the rules of thumb an experienced consultant applies almost without thinking. A principle is a single line of judgement you use to sanity-check a decision you've already reached. Short and usable, not essays. Something like: * Don't automate a process that isn't stable yet * If nobody owns a process, automating it won't stick * At low volumes, doing it manually is usually cheaper than building anything Decision rules Decision rules are more specific than principles. Each one covers a particular situation and says what to do in it, so most of them read as an if-then. For example: * If a process still changes every few weeks, don't build a custom automation for it yet * If an error would reach a customer directly, put a human approval step in front of it Where established consulting frameworks, methodologies or models already exist, find them and incorporate the most useful ones. We'd rather you build on proven thinking than invent something from scratch. 2. Tool library A profile for each major tool and platform, all answering the same set of questions: 1. What is it? 2. What are its main capabilities, and what can it actually do? 3. What is it best for, and what types of workflows or business problems does it suit? 4. When should a consultant choose it, and when should they reach for something else? 5. What systems can it connect to, and what APIs and integrations does it support? 6. What are its limitations? 7. How hard is it to implement? 8. What does it typically cost? 9. What are the main alternatives? Tools we'd expect to see covered, though the final list should come out of your research rather than this one: * Workflow automation: n8n, Make, Zapier, Microsoft Power Automate * Models and AI platforms: ChatGPT and OpenAI, Anthropic and Claude, Gemini * Business systems: Salesforce, HubSpot, Xero * Categories rather than single products: RPA platforms, AI document processing, AI search and knowledge tools, voice AI platforms, agent frameworks Alongside the profiles we want a capability matrix. Where the profiles are organised by tool, this is organised by capability, so someone can start from a need like "extract data from supplier invoices" and see their options. For each capability, set out which tools can do it, what each one actually does in practice, the use cases it fits best, where it falls short, and what the alternatives are. Between the two, a reader should be able to understand what each tool does and when it's appropriate to use it. 3. Top 20 business AI agents Research the 20 most common and commercially useful AI agents currently being built for businesses. We want the agents companies actually pay consultants to build, not a list of speculative ideas. For each one, document: * What it is, and what the agent actually does * Why it gets built, and the business problem behind it * When it's appropriate, and what makes a company a good candidate for it * How it's built, including the typical components, tools and systems involved * What it uses: models, APIs, databases, business systems and anything else *