An industry knowledge base is a structured document — vertical language, compliance realities, buyer archetypes, and pain patterns — that you attach to your AI workspace so every output speaks your market's language instead of generic business English. Most enterprise reps carry years of domain knowledge that their AI tools never see; writing it down once makes every future output more relevant, and the document compounds for as long as you sell in that vertical.
Why generic outputs happen to experienced reps
A rep who has sold into healthcare for six years knows that "compliance" means HIPAA audits and BAA negotiations, that "the buyer" is usually a committee with clinical and IT factions, and that certain words trigger legal review. The model knows none of this unless told — so it writes emails about "streamlining operations" to people who think in terms of denials management and payer mix. The gap is not model quality; it is missing context, and the fix is a one-time documentation effort.
What goes in the knowledge base
Six sections cover most of the value:
Vertical vocabulary. The fifty terms, acronyms, and phrases your buyers actually use, with a one-line gloss for each. This is the highest-payoff section — outputs written in buyer vocabulary read as insider immediately.
Buyer archetypes. The three to five roles you sell to, each with their priorities, their fears, what they get promoted for, and what gets them fired. Not personas from marketing; the real patterns from your calls.
Pain patterns. The recurring business problems in the vertical, ranked by how often they drive deals, each with the verbatim language buyers use to describe it.
Compliance and constraint realities. The regulations, procurement rules, and security requirements that shape every deal — and the claims that trigger legal review or destroy credibility.
Competitive landscape. Who appears in deals, how buyers perceive each player, and the vertical-specific dimensions of comparison.
Seasonal and cyclical rhythms. Budget cycles, industry events, regulatory deadlines — the calendar facts that determine when outreach lands and when it disappears.
How to build it in an afternoon
Do not write from a blank page; extract from what you already have.
Prompt: "You are helping me document my industry expertise for reuse. Interview me about selling into [VERTICAL]: ask me one question at a time covering buyer vocabulary, decision-maker archetypes, recurring pain points, compliance constraints, competitive dynamics, and buying-cycle timing. After each answer, capture it in structured note form. Continue until we have covered all six areas, then compile everything into a single organized knowledge-base document."
The interview format matters — answering questions surfaces knowledge you would never think to volunteer. Budget ninety minutes for the interview and thirty for editing the compiled result. Supplement with real artifacts: paste in three of your best discovery-call extractions and two won-deal proposals, and ask the model to mine them for vocabulary and pain language you missed.
Where to deploy it
Attach the document to a persistent AI workspace so it loads with every conversation — a project-style setup rather than pasting it per prompt. From then on, every cold email, business case, and call-prep brief generated in that workspace draws on the vertical context automatically. Keep a copy as a plain document too: when you run one-off prompts elsewhere, pasting the vocabulary and pain-pattern sections alone recovers most of the benefit.
The team multiplier
A knowledge base built by one rep is an edge; one shared by a pod is an operating system. When a team pools its interview outputs into a shared vertical document, new hires inherit years of context on day one, and every rep's outputs draw on the whole team's pattern library rather than their own slice of it. Assign one owner for the quarterly refresh so the shared version does not drift into neglect.
Maintaining it
The base decays slowly but really — regulations change, competitors reposition, new pain patterns emerge. A quarterly fifteen-minute pass keeps it live: skim your recent call extractions for new vocabulary, update the competitive section from your latest battlecard refresh, and adjust the calendar section for the coming quarter. Date-stamp each revision so you know what vintage of knowledge your outputs are drawing on.
Prompt: "Here is my current industry knowledge base: [PASTE] and here are my call extractions from the last quarter: [PASTE]. Identify new vocabulary, new pain patterns, and anything in the base that the recent calls contradict. Return a marked-up revision."
Frequently Asked Questions
How long does it take to build an industry knowledge base?
About two hours for the first version using the interview prompt — ninety minutes of guided Q&A plus thirty of editing. The quarterly maintenance pass takes fifteen minutes.
I sell into multiple verticals — do I need one per vertical?
Yes, and keep them separate. A merged document blurs the vocabulary and archetypes that make each one effective; build the deepest vertical first and clone the structure for the others.
What's the single highest-value section if I only build one?
Vertical vocabulary with buyer pain language. It changes the register of every output immediately, and it is the section buyers subconsciously screen for when deciding whether you understand their world.
Put It to Work
The interview, mining, and maintenance prompts above are part of the AI setup collection in the Promptifi library — 2,900+ prompts across 16 categories. Browse the library and turn your expertise into infrastructure.