AI for consultants: where it helps an engagement
A consulting partner doesn't run out of skill. They run out of working memory. Seven active engagements, each with its own thread of decisions, and the material lives in a dozen places. AI can take real work off that load, but not all of it in the same way. Here is where it helps, what to check before you use it on client material, and the part general tools leave to you.
Where AI helps an engagement
Most talk about AI in consulting is about the tool. It's more useful to start from the work. An engagement has three kinds of work that AI can take on, and each one needs something different from you.
1. Research synthesis
The first two weeks of an engagement are reading: the client's board packs, a data room, twenty interview notes, last year's strategy deck. A general assistant can summarise a long PDF, pull themes across interview notes, or compare two versions of a proposal in minutes. The discipline that makes it safe to rely on: ask for the exact passage behind every point, and read that passage before the finding goes on a slide. A synthesis you can't trace back to the source is a hypothesis, not a finding.
2. Drafting
First drafts are where the time savings are most obvious: the recap email after a workshop, the steerco summary, a storyline outline, a scoping note. Give it your rough notes and the audience, take the draft, then rewrite it in your own voice. What it can't supply is the judgment of what this client needs to hear this week, so treat the draft as a starting point, never the send.
3. Client memory
This is the job that breaks at partner level. What did the CFO say about margin in the October review? Did procurement ever confirm the revised scope? Which of the three workstreams did the COO say could slip? The answers are spread across email threads, PDFs, your notes and a voice memo you'll never replay. General assistants help here if you keep each client's material together: ChatGPT's projects, for example, keep "related chats, files, and instructions together," and OpenAI's help centre says a project can use "either default memory or project-only memory" (read 23 September 2026). That works for what you remember to upload. The gap is the correspondence itself: the threads keep moving after you last pasted them in, and letting a general assistant reach your mail directly means connecting the whole account, not the four client threads that matter.
| Engagement job | What AI does well | What you still check |
|---|---|---|
| Research synthesis | Summarising documents, pulling themes across interview notes, comparing versions | The passage behind every point before it reaches a slide |
| Drafting | First versions of recaps, summaries, outlines and client emails | Tone, emphasis, and what this client needs to hear now |
| Client memory | Answering from the material you've given it, kept per client | That the latest thread and the latest decision are actually in there |
Before you put client material into any AI tool
Three checks, in this order. First, your engagement letter and any NDA: some clients restrict where their material can go, and that overrides any vendor promise. Second, your firm's AI policy: many firms name approved tools and workspaces. Third, the plan you're on. OpenAI's enterprise privacy page (updated 8 January 2026, read 23 September 2026) says that for ChatGPT Business, Enterprise and its API, "By default, we do not use your business data for training our models." On a personal plan, OpenAI's data controls article says that when the "Improve the model for everyone" setting is off, "your new conversations won't be used to train OpenAI models." Training is one question. Retention, who can see a shared workspace, and what the client contract allows are the others.
Sources: OpenAI, Enterprise privacy; OpenAI Help Center, Data controls in ChatGPT; OpenAI Help Center, Projects in ChatGPT. All read 23 September 2026.
The consultant's real problem
It isn't a lack of information. It's too much of it, in too many places. The Q3 pricing thread is in email, the engagement notes are in one app, the deck the client sent is a PDF, and the thing you meant to follow up on is a voice memo. Before each call you reconstruct the state of that one engagement from memory, and hope you didn't miss the reply that came in while you were with another client. Research and drafting tools don't solve that, because the problem isn't writing. It's keeping the thread.
How Equerry fits a consulting workload
Equerry is a personal assistant who holds that thread for you. You CC it on the active engagement threads, and replies that keep its address on the thread keep arriving. You share the PDFs and screenshots a client sends from Mail, Files or any app with a share button, dictate a quick voice note in Equerry after a meeting, or save a note. Then you ask: what did the CFO say about margin in the October review? It answers from what you've shared, and when an answer draws on a note you saved, it carries a source chip: tap it to see the passage Equerry used and open the original. Before your next call, it prepares a briefing from what you've shared, timed to your calendar, so you walk in current, including the reply you hadn't opened yet.
Built for client-confidential work
| What consultants need | How Equerry handles it |
|---|---|
| Keep many clients straight | Ask about one engagement and it answers from what you've shared about it |
| Walk into calls current | Briefs you before each call from what you've shared, timed to your calendar |
| Trust the answer | Answers drawn on a saved note carry a source chip that opens the passage and the original |
| Protect confidential material | Sees only the threads you forward or CC; it never asks for your inbox |
| Stay in control of what it remembers | What Equerry remembered is listed in Settings → Your memory, and you can delete any line |
Why deliberate sharing matters here
When your inbox holds signed NDAs and seven confidential relationships, an assistant that wants to connect to your whole inbox is a hard sell. Equerry works the other way: you decide what it sees by forwarding or CC-ing specific threads, and the original stays in your own mailbox. The AI providers it uses are named on our AI disclosure page, so you can put that in front of your firm's policy before you start. It's the same reason Equerry suits client-facing executives more broadly. For a system that works without any AI at all, see how to stay organized as a consultant; to see how the general assistants compare on this job, read the best AI assistants for consultants; and for the prep itself, AI meeting prep.
Frequently asked questions
How can AI help consultants?
In three places on an engagement. Research synthesis: reading a data room, interview notes or a stack of client PDFs and pulling out the themes, as long as you ask for the passage behind each point and check it. Drafting: first versions of the steerco summary, the follow-up email, the storyline outline, which you then rewrite with your own judgment. Client memory: keeping track of what each client said, decided and asked for across weeks of threads and calls. General assistants are strongest at the first two. The third is the one that breaks when you run several engagements at once, and it's the job Equerry, a personal assistant on iPhone and iPad, is built for.
How is AI used in consulting day to day?
Mostly as a fast first pass, not a replacement for the analysis. Consultants use it to summarise long documents, compare versions of a proposal, turn rough call notes into a clean recap, draft client emails and outline decks. The work that still needs a person is judgment: which finding matters to this client, what the CFO will push back on, what was agreed in March. AI helps most when it has the right material in front of it, which is why keeping each engagement's material together matters as much as the tool you choose.
What AI tools do consultants use?
Usually two kinds: a general assistant to think and draft with, and something that holds the engagement's history. Which is best depends on the job, and we compare eight of them, each claim dated and sourced, in our guide to the best AI assistants for consultants. This page is about how AI fits an engagement, not a ranking.
Is it safe to put client information into ChatGPT?
For ChatGPT Business, Enterprise and its API, OpenAI says: "By default, we do not use your business data for training our models" (enterprise privacy page, updated 8 January 2026, read 23 September 2026). On a personal plan, when "Improve the model for everyone" is off, "your new conversations won't be used to train OpenAI models" (OpenAI data controls article, read 23 September 2026). Training is only one question: run the three checks above (engagement letter or NDA, firm policy, plan) first.
Can Equerry keep up with a client thread after I've shared it?
Yes, if you CC it. Add your own Equerry address to the thread once, and replies that keep that address on the thread keep arriving, so you can ask about the latest turn later. Forwarding sends the thread as it stands. It never asks for your inbox, and the original stays in your own mailbox.
Last updated 2026-09-23.
All product names, logos, and brands are property of their respective owners. Equerry is not affiliated with, endorsed by, or sponsored by any other product mentioned; comparisons are provided for informational purposes.