Faster sameness
The same pitch as last time, sooner. Where most firms are today.
Consulting-grade AI for professional services
Your firm bought AI. What has it won?
Out of the box, AI is good at summarising transcripts and pumping out bland. Neither ability gets your firm considered, shortlisted or chosen by clients with beige blindness. To get you painting in colour, we fix the workflow first, then build AI that knows your firm and knows how to persuade. One firm we did this for went from winning 35% of its proposals to 85%.
Your firm has the AI licences, the training and a prompt library. The drafts still read like every other firm’s, because the AI’s training builds your outputs from every other firm’s words.
Now the partners want to know what their AI investment is bringing them, and more meeting summaries and faster LinkedIn posts aren’t the right answer. With clients pushing fees down because AI makes advice cheaper, partners are looking to BD to find ways to raise revenue using AI.
You won’t grow the firm with a half-day workshop, an enthusiast’s prompt.docx or a monthly licence for a tool that reduces AI to find-and-replace on pitch templates.
Why AI in BD disappoints
Hammer time
In the 1980s, companies poured money into putting a computer on every desk. In 1990, Michael Hammer explained the disappointing results in the Harvard Business Review: companies used the new machines to keep working the old way, only faster. Hammer called it “paving the cow paths”.
BD risks wasting AI on paving the same cow paths. The pitch document still starts from the most similar pitch document anyone can find, when every pitch should start from the client. Credentials still stand in for proof. You don’t get better results from doing the same things faster.
And AI can’t do it better out of the box. Load your old documents into Copilot, Claude, ChatGPT, Harvey or Legora and you get the facts with none of the client-persuading craft you need. The result: your facts, presented the same way as the facts of every firm with a similar AI tool.
Two failures, two fixes
Generic AI
Steeped AI
Old workflow
Old workflow · generic AI
Old workflow · steeped AI
Better drafts, wrong process
Sharper writing, still starting from the last proposal instead of the client.
Redesigned workflow
Redesigned · generic AI
Right process, no craft
The work starts in the right place, then a tool with no persuasion skill writes it.
Redesigned · steeped AI
Consulting-grade AI
The right process, run with your firm’s knowledge and the craft of persuasion built in.
What consulting-grade AI is
Redesign, steep, build
Three pieces of work, in this order.
Redesign the workflow. We find the unwritten rules in how your team pitches, proposes and positions, then rebuild the work to start from the client. AI slots in where it earns its place.
Steep the AI. Steeping means building expertise and knowledge into the AI, not just instructions. Most firms stop at prompt engineering, which gives the AI better instructions. Some go on to context engineering, which loads it with the firm’s documents. Those steps don’t teach AI how to persuade your clients. So we build in two layers.
The expertise
- How persuasion works
- How a buyer reads a pitch
- How an offer is put together
The knowledge
Context engineering done properly
- Your positioning
- Your sectors
- Your wins and losses
- The way your best partners pitch
Build the advisors. Your team gets advisors for pitches, proposals and practice groups, running inside the AI your firm has already approved.
Craft, knowledge, or both
A new marketing hire
Persuasion craftVaries
Your firm’s knowledgeNone yet
Knows marketing, with some persuasion skill. Doesn’t yet know your clients, sectors or wins.
AI loaded with your documents
Persuasion craftNone
Your firm’s knowledgeHigh
Knows what’s in the files. Writes it up the way everyone else does.
A steeped advisor
Persuasion craftHigh
Your firm’s knowledgeHigh
Both layers, built in. Available to everyone in BD.
Add what your clients think, if you want it. An advisor built on your firm’s knowledge knows what your firm believes. Add an ASI study and it also knows what clients in your market weigh when they choose a firm, in their words.
The order of work
1
Redesign
Find the unwritten rules. Rebuild the workflow to start from the client.
2
Steep
Build in the persuasion expertise and your firm’s knowledge.
3
Build
Advisors for pitches, proposals and practice groups, inside your approved AI.
Optional
Add ASI
An Asymmetric Strategic Intelligence study: what clients in your market weigh, in their words.
What you get
About a month, and as little as four hours of your team’s time
We interview the people who do the work now, to learn the result you want and how you get there today. Then we give you versions to try, take your feedback and refine them. Because we know what to ask, your team may give us as little as three or four hours. The whole job takes about a month, depending on how much of the work we’re rebuilding.
At the end, you have:
- the rebuilt workflow, written down
- an advisor or, where your tools can’t host one, a set of steeped prompts built for each step
- training for your team in the new workflow
- a playbook your team keeps
What this did for two clients
Process first, then AI
Searchlight Group · Recruitment
Searchlight pitches for government and not-for-profit work, scored by evaluation committees. Its proposals took six hours each, and it won 35% of them. A law firm’s panel review works the same way: a committee, a scoring sheet and a stack of responses that all sound alike.
We didn’t start with AI. We rebuilt how the firm pitched. Then we built the AI the firm pitches with.
Win rate
Hours per proposal
With it, the firm’s win rate reached 85% in 2025. Proposals now take two hours instead of six, but that time saving is the smaller result. On its own, the time saving just lets you write losing proposals faster.
“When I re-read my old proposals and the ones written with the AI that Taleist trained, I know which proposal I’d go with if I were the client. It’s not about the writing — it’s about the strategic thinking behind every section.”
A global law firm · One firm, five projects
The firm came to us first with a change problem: bringing its partners onto a new CRM. An ASI study found the message that would move them, and we built the AI its team uses every day to tailor its change communications.
The firm came back for a positioning advisor for its central marketing and pitch teams. Then it came back for its pitching engine. Now we’re working with it on image tagging.
Unsurprisingly, the firm made us sign an NDA, so we can’t tell you who it is.
What a steeped advisor does differently
A faster draft, or a better argument
Ask generic AI to answer a tender and it starts from your last response and the list of requirements. It gives you a faster draft.
Ask a steeped advisor and it starts somewhere else. Which requirement will the committee weigh most? What does this buyer need to believe before choosing a firm they haven’t used? Which of your credentials prove it, and which are padding? It gives you a better argument.
One executive summary, before and after
BeforeAfter
Title
Why [the firm] is the right choice for you
Decisions about your companies that stay decided
Opening
“Dear Procurement Team, Thank you for the opportunity to respond to your request for proposal. We are delighted to submit…”
“With [the firm], a decision about one of your companies stays decided…”
Headings
- Market-leading corporate and M&A expertise
- Extensive regulatory experience
- Our long-standing relationship with you
- A truly local capability
- You know each restructuring route’s regulatory path before you choose it
- Your positions stay the same across your portfolio, whichever firm advises
- Red-flag findings on several companies arrive in time to shape your choice
The real tender, scored blind
0 of 4 headings name a benefit to the client
595 words, about half of them cuttable
3 of 3 headings name a benefit to the client
435 words, 23 of them cuttable
Where firms use it
Six places consulting-grade AI pays
Win the pitch
Build the pitch around what the client is weighing, not what you said last time.
Write the proposal the committee scores
Answer the evaluation criteria with proof, in the order the panel reads them.
Write the emails that get answered
The follow-up after a meeting, the client update, the note to a contact who has gone quiet: written to move the reader, in your firm’s voice.
Fill the room
The invitation, the reminders and the follow-up afterwards, written so the right people come and the conversation carries on after the event.
Give a practice group its own advisor
Its sectors, its clients, its language, so partners turn insights into client contact week after week.
Keep the rainmaker’s method
In most firms the best BD instinct lives in one partner’s head. Build it into an advisor and the firm keeps it when the partner retires or moves.
Questions firms ask
Before you start
“Won’t the next AI model fix this?”
It will write more fluently. It still won’t know your clients, and it will still start from the last proposal. A better model runs the old workflow faster.
“We already have prompt training and a prompt library.”
Keep them. They’re a good start. But a prompt can only write to the standard of the person who wrote it, on the day they wrote it. A prompt that produces an answer you like isn’t the same as one that gets everything the model can give, and nobody retests it when the model changes. And a prompt is one instruction pasted into the old workflow. It can’t carry your firm’s knowledge, or the persuasion expertise we build in.
“Isn’t this just prompt engineering?”
Prompt engineering writes the AI better instructions. Context engineering gives it better information. We do both, then add what neither covers: the expertise to persuade, and a workflow that starts from the client.
“Our innovation team is already building AI tools.”
Good. They have the platform and your firm’s knowledge. What they rarely have is persuasion expertise and a redesigned BD workflow to build for. We work with them, inside the tools they’ve built.
“Do we need ASI first?”
No. Most advisors start with what your firm already knows. ASI adds what your clients think, when you want it.
“Will the team use it?”
Advisors get used when what they produce is good enough to put in front of a partner. That’s the standard we build to. The law firm above uses its tool every day.
“Isn’t AI our IT team’s job?”
IT gives the firm safe, approved AI. We don’t choose platforms or run infrastructure. We work inside what IT has approved and build what the BD team does with it.
“What about our AI policy?”
An advisor runs on your firm’s knowledge, so it lives inside the AI your firm already trusts. We build on the platform your policy approves, and we sign your NDA before we start.
“Doesn’t an ASI study already come with an AI tool?”
It comes with a prompt that answers questions about the research. A steeped advisor does the work: it drafts, argues and positions with your firm’s knowledge and the craft built in.
“What does it cost?”
A fixed fee, set once we’ve sized the work. A whole pitch process, from the RFP arriving to proofreading the final submission, is a bigger job than a workflow for event invitations. You choose whether the AI is built on your firm’s knowledge alone, or with an ASI study added.
Book a call
Find out what your AI should be doing
Book a 30-minute call. Tell us the work: the pitches you keep losing, the proposals that eat a week, the practice group that should be growing faster, the rainmaker who’s two years from retiring. We’ll tell you whether to start with your firm’s knowledge or with what your clients think. If AI is the wrong answer, we’ll say so.
Prefer email? asi@taleist.agency
Tell us the work in front of you and we’ll reply within one business day.