AI agents for consulting firms

Consulting is sold by people, and a consulting firm's website rarely closes a deal by itself. What it does decide is whether a prospect asks for a conversation, and with whom. Visitors arrive from an article, a conference talk or a referral, with a specific problem in mind. An AI agent helps when it connects that problem to the right practice and partner quickly, and does nothing a partner would not do in a first call.

Who visits a consulting firm website and what they need

• Prospective clients. Executives and functional leaders who want evidence that you have solved their kind of problem in their kind of company. They ask about industries, relevant case studies, who would staff the work, engagement models, typical duration and how to start small.
• Procurement teams. They want to know where to send an RFP, how vendor registration works and who signs the NDA.
• Readers of your insights. People who found an article or report and want more on the same topic, or the author's view on their situation.
• Candidates. Graduates and experienced hires asking about roles, offices and the interview process. They often use the same contact routes as prospects, so routing matters.
• Alumni, journalists and existing clients, each of whom needs a different contact.

Which ASDK agents fit a consulting firm

ASDK is in early access. Sales, Chat and Agent-Ready are early-access agents; Search is coming soon.

Sales: the right partner, not the general inbox

Example flow: a visitor writes, "We're a mid-sized manufacturer and our margins are shrinking." Sales asks the questions you set: role, company size, the problem area, timing. It matches the answers to a practice, offers that partner's calendar through Cal.com, Calendly or Google Calendar, and sends the lead with the conversation attached to your team, and to your CRM through webhooks and Zapier or Make. The partner reads a brief, not a one-line form.

Chat: answers the process questions

Example flow: "Can we start with a short diagnostic?" Chat answers from your "how we work" page and links to it. Candidates asking about interviews get the careers page; RFP questions get the procurement contact. Anything the site does not cover, Chat says it does not know and offers to pass on.

Search (coming soon): your insights library, answered

Many consulting firms publish a lot, and a keyword search box is a weak way to explore it. When Search ships, it will answer a question such as "what have you written about nearshoring?" with a sourced summary and the best articles, and report the topics people searched for and did not find.

Agent-Ready: expertise that assistants can read

Agent-Ready hosts an llms.txt file, as proposed at llmstxt.org, pointing to your practice pages, people and insights; exposes actions such as "contact a partner" through the widget; and on Pro and above reports which AI agents visited.

Guardrails that matter for consulting firms

• No free consulting. The agent explains your approach and points to published work. It does not diagnose a visitor's business or recommend a strategy.
• Ask for no confidential details. Visitors sometimes describe an acquisition, a restructuring or a dispute in the first message. The greeting should ask them not to share confidential information before an NDA, and lead capture should stay minimal.
• Client names stay protected. Only clients you name publicly, with permission, belong in answers. An anonymized case study stays anonymous; the agent must not speculate about who the client was.
• Never confirm or deny client relationships. "Do you work for our competitor?" gets a polite referral to a partner, who can handle conflict questions properly.
• Credentials as published. Bios, titles and credentials come from your people pages only. Google's guidance on helpful content calls fabricated creator profiles and false credentials deceptive, and the same standard applies to anything an agent says about your team.

Every conversation is labeled as AI, and a handoff to a person is always offered.

How AI assistants research consulting firms

Part of a buyer's first research now happens inside AI assistants. In Gartner's May 2026 release, 45% of B2B buyers surveyed had used generative AI during a recent purchase, mostly to gather information on vendors, and 69% preferred to validate AI-generated insights with a sales rep. For a consulting firm, the assistant's summary is the first impression, and the partner call is where it gets checked. An earlier Gartner survey found 69% of buyers saw inconsistencies between a supplier's website and what its sellers said, so the site and the partners should tell the same story.

What to make explicit on your pages:

• Practice areas and industries, described as problems you solve, not as abstract capabilities.
• Named experts with real bios, linked from what they write. Google recommends bylines that lead to information about the author, and Schema.org's knowsAbout property lets a person's profile state their topics in structured data.
• Engagement models: fixed fee, time and materials or retainer, typical duration, and whether you start with a diagnostic.
• Geography and languages you work in.
• Case studies with the industry, the problem, the approach and dated results, anonymized where needed and labeled as such.

The free Agent Readiness Checker shows how your site looks to AI agents today, and Agent-Ready keeps the machine-readable side current.

Where to start

Start with Sales on your practice pages and contact page. Write the routing rules down first: which practice owns which industries and problems, which partners take first calls, and what happens to an inquiry that fits no practice. Add Chat once you have seen which process questions come up most, and Search when it becomes available.