IBM is training consultants, and clients will notice

IBM’s partnership with OpenAI is not just another enterprise AI announcement. The important signal is that IBM is training tens of thousands of consultants to use frontier AI technologies in client work.

That changes the market for consulting and advisory firms.

Clients won’t suddenly expect every adviser to be a machine learning engineer. Most buyers still need help framing problems, setting priorities, getting stakeholders aligned, and making decisions. Those are consulting skills. But they will increasingly expect their advisers to work faster, bring better-informed options to the table, and know where AI belongs in the operating model.

For a consulting firm doing $1 million to $25 million in annual revenue, the risk is not that IBM will take every engagement. It won’t. Your firm likely wins because you understand a sector, have senior relationships, or know how to get work implemented in a specific kind of business.

The risk is a quieter one. A client compares two proposals. Both teams have credible experience. One firm shows a practical AI-enabled delivery approach, can produce a well-researched point of view quickly, and has examples of repeatable tools behind its advice. The other presents a standard deck and says its team is “exploring AI.”

That gap will become harder to explain away.

IBM’s move should prompt a more useful question than, “Which AI tool should our people use?” The question is: where does your firm repeatedly spend senior time on work that could be structured, reviewed, and reused through an AI agent?

For most firms, the answers start in proposals, engagement research, and knowledge management. Those are not glamorous back-office issues. They determine cost of sale, speed to insight, margins, and how much value your firm gets from the IP it creates.

Client expectations are moving from awareness to capability

A few years ago, a consultant could mention AI in a client conversation and appear ahead of the curve. That bar has moved. Many client leadership teams now use tools like ChatGPT, Microsoft Copilot, Claude, or internal enterprise AI environments themselves. They know a first draft can be created in minutes.

What they are trying to work out is where AI can be used safely, where human judgement remains essential, and how to turn isolated experiments into working processes.

That is where an AI-fluent consulting firm has an advantage.

AI fluency does not mean asking every consultant to become a prompt specialist. It means your people understand four practical things:

1. How to define a business task clearly enough for AI to assist.
2. How to provide the right firm knowledge, client context, and constraints.
3. How to review outputs rather than accepting them blindly.
4. How to build a repeatable workflow that improves each time the firm uses it.

This is the difference between a consultant using a public chatbot to tidy up a paragraph and a firm building an operating asset.

If IBM is investing at scale in consultant training on OpenAI technologies, clients will see that investment in the field. They will hear AI-informed questions. They will receive proposals that are more tailored. They will see delivery teams arrive with research briefs, structured knowledge, and a clearer view of possible automation opportunities.

Smaller firms don’t need IBM’s training budget to respond. They do need a deliberate training plan linked to the work their teams do every week.

A good starting point is to study practical examples in the Enterprise DNA insights library, then choose one or two workflows where a well-governed agent can produce a visible result. Training sticks when consultants can apply it to a live proposal or engagement, not when it stays in a generic AI workshop.

The real cost is senior time spent recreating work

Consulting firms often say their biggest constraint is talent. That’s true, but it usually shows up in a specific form. Senior people are doing too much repeatable work because the firm has not turned its knowledge into a usable system.

Consider the major proposal.

A partner receives an inbound opportunity or a referral. The brief looks promising. Someone finds a few old decks. Another person searches folders for relevant case studies. A manager pulls together background research. The pricing approach is debated from scratch because past proposals are difficult to compare.

Then the partner rewrites most of it late at night because the first draft doesn’t sound like the firm.

It is common for a major proposal to absorb 20 to 40 hours of effort before client meetings, revisions, and commercial negotiation. Not all of that time should disappear. Good proposals need judgement, positioning, and an honest view of the client’s situation.

But much of it is retrieval and assembly. The firm has already written relevant capability statements, scope language, workplans, team bios, and pricing logic. It just can’t find or apply them consistently.

Research creates a similar drag. A new engagement starts and a team spends days, sometimes weeks, gathering industry reports, competitor information, company news, financial context, regulatory changes, and internal client materials. Some of that research is necessary. Yet firms repeat large sections of it every time they enter a familiar sector or client type.

Then there is the deeper issue. Every completed project creates intellectual property. Workshop outputs, strategy decks, interview notes, operating models, data analyses, and meeting transcripts all contain useful lessons. At most firms, that material is stored in project folders, personal drives, email, or a document platform with weak search.

The firm pays once to create insight. Then it pays again when another team needs the same insight six months later.

For a firm in this size range, we usually see annual leakage from those patterns in the $80,000 to $300,000 band. That is not a claim that every hour can be automated. It is a practical estimate of avoidable senior effort, slow proposal cycles, duplicated research, and lost reuse across a year.

The commercial issue is simple. If your senior team is consistently busy but margins feel tight, the answer may not be hiring more people. It may be giving the team a better way to use what the firm already knows.

What an AI agent looks like in a consulting firm

An AI agent is not a magic box that replaces a consultant. It is a defined workflow with access to approved information, instructions for how to perform a task, and clear points where a person checks the work.

That distinction matters. A generic chat tool starts with a blank page. A firm agent starts with your methods, your source materials, your standards, and the context of a particular opportunity or engagement.

At Omni ops, we build these workflows around business processes, not around novelty. For consulting firms, three agents often make sense as early priorities.

The Proposal Generation Agent

The Proposal Generation Agent pulls past proposals, case studies, service descriptions, team credentials, and pricing approaches into a tailored first draft for a new opportunity.

The workflow starts with an intake form or a short structured conversation. The person leading the opportunity provides the client name, industry, decision-maker priorities, scope, budget range if known, delivery timeframe, and strategic angle.

The agent then searches only the approved knowledge sources. It identifies relevant prior work, extracts case study evidence, suggests a scope structure, and drafts sections such as the executive summary, client context, approach, deliverables, team, timeline, assumptions, and commercial options.

It can also flag gaps. Perhaps the firm has no approved case study for a relevant issue. Perhaps the proposed timeframe conflicts with delivery capacity. Perhaps similar work has historically required a different price range.

A partner still owns the final proposal. They should challenge the point of view, check claims, decide what the firm will and will not promise, and refine the commercial position. The agent doesn’t replace that responsibility.

What it does is remove the blank-page problem. Instead of starting after 8 pm with scattered files, the partner begins with a structured draft that is grounded in the firm’s actual work.

That can reduce the repeated effort involved in every pitch while improving consistency. It also gives junior team members a better model of how the firm frames and prices work.

The Research Agent

The Research Agent runs structured industry and company research at the start of every engagement. Its output includes sources, summaries, key questions, and a one-page brief that gives the delivery team a common starting point.

A good research workflow does more than collect web links. It separates facts from assumptions. It identifies recent developments, financial signals where relevant, competitor moves, regulation, customer trends, and stated strategic priorities. It records the source and date for each important point.

For a client strategy engagement, the agent might produce:

• A one-page company briefing for the project kickoff.
• A sector landscape with the five to ten issues most likely to shape the engagement.
• A competitor comparison based on public evidence.
• A list of hypotheses the team should test through interviews.
• A source register so consultants can verify key claims.

The human team then decides what matters. That is the consulting value. A research agent can find and organize information at speed. It cannot know, without direction and review, which trade-off a client should make or which political issue will stop a recommendation from being adopted.

Still, replacing the repeated mechanics of research changes how the team spends its first week. Less time collecting. More time forming useful hypotheses.

That is exactly the kind of practical capability clients will expect when larger firms train their consultants in frontier AI systems.

The Knowledge Agent

The Knowledge Agent reads every deck, document, and meeting transcript the firm produces and answers questions across the approved corpus.

This is where many consulting firms have the largest long-term opportunity.

Imagine a partner asking, “What have we learned about pricing transformation programs in mid-market industrial businesses over the last three years?” Or a manager asking, “Show me our strongest evidence for the first 90 days of a post-merger integration program.”

Without a knowledge agent, those questions trigger messages to colleagues, searches across folders, and a fair amount of luck. With the right permissions and source structure, the agent can retrieve relevant material, cite the original documents, and provide a concise answer with links back to the evidence.

It should not have unrestricted access to every sensitive client file. Permission design matters. Client confidentiality, retention rules, project access, and approved source locations need to be built into the workflow from the beginning.

The goal is not to create a giant pile of documents that an AI can vaguely search. The goal is to preserve the firm’s working memory in a way that helps people deliver better work.

Training should follow live workflows

The wrong response to IBM’s announcement is to buy licenses for a general AI tool, run one lunchtime session, and assume the firm is now AI-enabled.

That approach creates uneven adoption. A few enthusiasts use the tools. Others avoid them because they are unsure about quality, confidentiality, or what good use looks like. No shared capability is built.

A better approach is to formalize training around live workflows.

Start with a small group. This could include one partner, a senior manager who owns delivery quality, someone responsible for business development, and a practical operator who understands the firm’s systems. Give them a clear mandate to improve one process in 30 to 60 days.

Train the group on how to break work into steps, define inputs and outputs, identify trusted sources, write review criteria, and record failure cases. Use a real proposal or a real engagement research task. The work will expose issues that a generic training exercise never will.

Then turn the successful workflow into a standard. Create a short playbook. Clarify what data can be used. Define the human review step. Show examples of good and poor outputs. Build it into onboarding for new consultants.

You can use Omni advisory to frame the operating and governance questions, while Omni apps can support the interfaces your team actually uses. The point is not to deploy technology everywhere. It is to make selected work more repeatable and easier to supervise.

If you need a practical structure before involving your team, download Deploy Your First Business Agent. The accompanying worksheet and checklist helps you map a process, identify source material, assign a human owner, and decide what success should look like before building anything.

Find the work where your firm is leaking value

A sensible AI plan begins with a process inventory, not with a platform comparison.

Look at the last 10 proposals your firm produced. How long did each take? How much of the work was finding prior material, formatting, writing standard sections, and revising inconsistent content? Which proposals involved the most senior people? Which pieces of prior work were useful but hard to locate?

Then look at three recent engagements. How much research was repeated? Where did teams have to recreate an approach, framework, or deliverable because they could not find a past example? What did the client ask for that your firm had answered elsewhere?

These questions uncover the best agent opportunities because they point to work with three characteristics:

1. It happens frequently.
2. It follows a recognisable pattern.
3. It draws on information the firm already owns or can source reliably.

You do not need perfect data to start. You need enough clarity to identify the bottleneck that is costing the firm the most time or limiting its ability to respond quickly.

This is also why an outside view helps. Owners and partners usually know there is duplication. They are often too close to the delivery model to quantify where it begins and what a practical first build should be.

See Omni for consulting firms if you want a view of the specific workflows we assess in advisory businesses. An Omni Audit takes 60 minutes and produces three useful outputs: a map of the highest-value manual work, a prioritised agent opportunity list, and a practical next-step plan. There is no deck built to impress you. The aim is to identify work worth changing.

If you want to discuss your firm’s proposal, research, or knowledge workflow directly, Book a 60-min Omni Audit.

The competitive response is specific, not broad

IBM training thousands of consultants on OpenAI tools is a market signal, not a reason to panic.

Your response doesn’t need to be a sweeping AI transformation program. It needs to be credible, visible, and connected to real work. Build internal fluency around the processes that consume senior hours. Give your people approved tools and clear guardrails. Preserve the knowledge created by each engagement. Review the quality carefully.

Then take that capability into client conversations in an honest way. You can explain how your firm uses AI to accelerate research, improve reuse, and focus consultant time on judgement. You can also show clients where similar operating improvements may apply inside their own business.

That is a stronger position than simply claiming your team uses AI.

The firms that move first will not win because they have the longest tool list. They will win because clients experience a better process. Faster turnaround. Better preparation. More relevant evidence. Less reinvention. A team that can talk about AI without turning every conversation into a technology pitch.

The opportunity is sitting inside the work you already do. The question is which workflow you will formalize first.

For a focused assessment of that decision, review the AI audit for consulting firms, then Book my Omni Audit.