Dear Management Consultants,
We need an intervention. It’s about your future, or what’s left of it after the invoice templates stop printing themselves. You’ve had a lovely run at the top of the corporate terrarium: tailored suits, status lounges, and fees that soar on the thermals of “we have the best people, trust us.” Charming. Unfortunately a new colleague just walked in, knows everything, never sleeps, never takes credit, and costs approximately the same as the mints on your hotel pillow. Its name is Artificial Intelligence. It is here for your job, at least the 80 percent that consists of glorified Googling, slide buffing, and spreadsheet whispering.
You protest already. “We provide deep insight. Our clients crave the human touch. We go beyond the data.” Adorable. While you’re saying that, a slide bot is assembling a board pack in minutes, and a large language model is writing a tighter market view than your MBA study group could manage between cold brews and career-coaching sessions. The 2 a.m. fact-finding slog, once your junior team’s rite of passage, now happens in the time it takes to open your laptop. The machine does not require pizza, praise, or “protected weekends.”
Interviews and surveys, those noble time sinks you’ve been billing by the hundreds of hours, are getting transcribed, summarized, sentiment-scored, and cross-referenced before your triple shot stops steaming. Sacred Excel models, the ones you guarded like a dragon sits on gold, are refactored, stress-tested, and cloned by code assistants before lunch, with fewer errors than your “rock star” analyst. Your prized “toolkit,” the one stuffed with templates and last year’s sanitized deliverables, has been outgunned by an engine that ate the internet and didn’t have to attend your Knowledge Management brownbag to do it.
So what exactly remains? Being “indispensable advisors,” allegedly. The politics, the choreography, the eyebrow gymnastics in front of the HIPPO, the gentle art of disagreeing without getting escorted out. True, AI cannot yet roll its eyes at a terrible joke or triangulate six feuding VPs with a single raised chin. But if the analysis is commoditized and the deck builds itself, explain again why the client should pay seven figures for professional handholding. “Two million for therapy, featuring charts.” Try that line in the budget meeting and listen for the silence.
Maybe you plan to hide behind the machines, present their work as your own brilliance, and keep charging champagne prices for tap water. Precious. Clients already know the game. They know you are automating the scutwork. Their next question is simple: why are they paying for ten consultants when one consultant plus a stack of GPUs did the heavy lifting. The pyramid model starts to wobble when the bottom layer is software. When the base disappears, the expensive tip becomes a very lonely triangle.
The jig is up. Two roads remain: automate yourself, or become actually useful.
Automate yourself means drag every repetitive, template-driven task into the bot corral now, before your client’s procurement team does it with a machete and a price benchmark. Be the partner who says, with a straight face, “We delivered in half the hours with twice the insight, because we use AI end to end.” Yes, short-term billing shrinks. Cry later. Better you swing the knife than have someone else do it while quoting an AI SaaS price list.
Become truly useful means stop cosplaying as a human PDF exporter. Show up not to read slides, but to change minds. Use the machines for the grind, then do what they cannot: frame the real question, call nonsense on both the model and the monarch, make the uncomfortable recommendation, and own the outcome. Develop domain mastery so deep that you make judgment calls under uncertainty with context, history, and consequence. Knowledge is cheap now; wisdom is scarce and getting scarcer.
Time for unpleasant arithmetic. A lot of consulting has been process theater. Teams of fifteen where five were sufficient. Three weeks of “analysis” to justify a decision driven by gut and office politics. Complexifying the obvious to imply value. AI is a ruthless simplifier. It collapses the fog. Your client can ask for a second opinion in plain language and get it in plain minutes. “Assistant, is this analysis any good?” If your work cannot withstand that question, you are not selling value, you are selling varnish.
Denial will have its cameo. “We survived ERP, cloud, analytics, we will survive this. Clients need synthesis. Clients have no time.” Some will call from habit. That grace period wanes fast. The next generation of executives grew up with autocomplete and on-demand everything. They do not tremble at frameworks in tidy boxes. They ask, “Did you make that matrix yourself, or did your AI intern do it while you were booking flights?” Then they build an internal strategy pod wrapped around AI and cut your scope to the bone.
Here’s a free upgrade while the kindness buffer is still funded: stop selling brains by the hour, start selling outcomes with accountability. The manpower rental era is sputtering. No one wants to buy 100 hours of thought when they can buy a month of results with a warranty. Move from “we think” to “we deliver and we stake our fee.” If the internal AI team can do it in two days, you can do it in two days and add the hard part, which is getting people to change behavior and systems to behave. Less ponder, more practice.
And spare us the misty-eyed speeches about empathy that technology will never replicate. If you genuinely care about clients, you will deploy any tool that serves them better, even if it trims your chargeability. The smart consultants are retooling already. The rest will enjoy being focus-grouped out of relevance, perhaps after a final lap selling “heritage expertise” to an AI startup that will thank them politely and then replace them.
Let us read the handwriting on the flipchart together. Evolve into AI-native advisors who trade on judgment, credibility, and execution, or fossilize into a cost line item labeled “legacy.” The upside is real. Less drudgery. More decisions that matter. The downside is also real. Middlemen get squeezed. Ornamental intelligence goes first.
Choose quickly. The clock does not care how many frequent-flyer tiers you have.
Sincerely,
A former consultant and current buyer of consulting services
PS: As a potential purchaser of consulting services, here is a little “do-not-buy” checklist – ten red flags that an engagement is overpriced or better served by AI/productized means:
- No knowledge transfer or capability build for you. The most overpriced projects are those that leave you no smarter or more capable, forcing you to re-hire them. If they aren’t building an internal system, process, or team as part of the engagement, you’re just renting insight that will disappear. In the AI era, demand that they at least set up a dashboard or train your staff on key tools so the value persists.
- Deliverable is mostly a PowerPoint pack of known information. If the scope is “provide an overview of industry trends” and you suspect they’ll just Google and summarize, consider doing that internally with AI. Don’t pay $500K for a book report.
- Large junior team proposed for basic analysis. If a firm wants to put 5+ analysts on data gathering or modeling tasks, ask why your own team (with tools) or a smaller team with AI can’t do it. Paying for a “pyramid” full of juniors makes little sense if their tasks can be automated.
- No clear outcome or success metric. If the proposal is vague (“assist with strategy definition”) and doesn’t tie fees to any result or decision, you risk paying for endless process. Insist on outcome definition or don’t proceed.
- They won’t transfer data or tools at project end. Overpriced engagements often keep the IP – if the consultant’s analysis or dashboards won’t be handed over in usable form, you’ll be stuck re-hiring them. That’s a red flag; prefer projects where deliverables include raw data, code, or tool setups your team can use.
- Over-reliance on “their proprietary model” without transparency. Some firms charge a premium for black-box benchmarks or frameworks. If they can’t explain the magic (especially if an AI is behind the scenes), be cautious. You might replicate results with a bit of research or a smaller expert engagement.
- Long duration proposals for what could be a quick answer. For example, a 6-month study to answer a question that seems narrow. AI and agile methods can often deliver insight faster. Challenge the timeline – are they milking billable weeks? Often, a pilot or shorter study could achieve 80% of the value.
- Project staffed by generalists when specialists are available. If the firm’s team has a steep learning curve (you find yourself teaching them about your industry), you’re funding their education. Instead, use an AI to get them up to speed or hire a specialist boutique with actual expertise.
- Heavy focus on data collection from public sources. That’s a sign you might do it yourself. If the value proposition isn’t in advanced analysis or change leadership, but just gathering data anyone can find – don’t buy it. Use internal resources or data services for that portion.
- “Man-month” estimates that don’t factor technology. If a proposal says “will require 4 people for 3 months” but makes no mention of using tools or automation, they’re likely planning to do things old-school (i.e., inefficiently). Push back – ask how they can do it faster or leaner with technology. If they don’t have a good answer, skip it.
