Why Most Companies Don't Need an AI Strategy — They Need an AI Colleague

A top-tier consulting firm will charge you $300,000 or more for an AI strategy. The deliverable: a slide deck, a 6-month roadmap, and a proof of concept that — statistically — won’t survive contact with production.

MIT’s NANDA initiative studied over 300 enterprise AI deployments in 2025 and found that 95% of generative AI pilots deliver zero measurable return on the P&L. Not “modest returns.” Zero.

The problem isn’t that your company lacks a strategy. The problem is that the strategy itself has become the product — and the consulting industry is selling it to companies that just need a tool that works.

The AI Strategy Industrial Complex

The failure numbers are bad, and they keep getting worse.

Gartner predicted that 30% of generative AI projects would be abandoned after proof of concept by the end of 2025 — citing poor data quality, unclear business value, and escalating costs. That turned out to be optimistic. S&P Global reported that 42% of companies scrapped most of their AI initiatives in 2025, up from 17% the year before. The average organization abandoned 46% of its AI proofs-of-concept before they ever reached production.

Bain’s 2024 study found that 88% of business transformations — not just AI, all digital transformation — fail to achieve their original ambitions. IDC pegged it even worse: 88% of AI pilots never make it to production. Only about 1 in 8 prototypes becomes operational.

And yet, AI consulting is booming. McKinsey says 40% of its client work now involves AI. BCG’s AI consulting generated 20% of its 2024 revenue. Accenture booked $900 million in AI consulting in a single year. PwC announced a $1 billion AI investment. EY’s AI revenue jumped 30%.

The disconnect is remarkable: a massive industry selling AI strategy to companies where the overwhelming majority of AI projects fail. Enterprise AI transformation engagements run $100,000 to $500,000+. Even a “small” strategy assessment costs $5,000 to $25,000.

For a 50-person company, that consulting engagement might cost more than the problem it’s trying to solve.

The SMB Dropout

Small and mid-sized businesses got the message that AI matters. They tried it. Then many of them stopped.

A NEXT survey of 1,500 small business owners found that AI tool usage dropped from 42% in 2024 to 28% in 2025. Not because the tech got worse — because the gap between the hype and the actual experience was too wide.

The barriers are straightforward. A Revenued survey found the top challenges: lack of expertise or training (26%), accuracy concerns (18%), and cost (16%). A Service Direct study showed 62% of SMBs cite lack of understanding about AI’s benefits as the main barrier. Only 33% of SMB AI users received proper training, according to Microsoft’s own data.

The tools themselves often exclude the market they claim to serve. Microsoft 365 Copilot originally required a minimum of 300 seats at $30 per user per month — a $108,000 annual commitment. (We break down exactly what Copilot can and can’t do in a separate analysis.) A custom enterprise GPT from OpenAI starts in the millions. These aren’t SMB products. They’re enterprise products with SMB marketing.

The knowledge gap isn’t about intelligence. Small business owners run complex operations every day. The gap exists because an industry built for Fortune 500 companies keeps telling 50-person firms they need the same approach — just smaller.

They don’t. They need something fundamentally different.

What the Research Actually Shows

Here’s the part the strategy consultants gloss over: the studies that show what actually drives productivity gains aren’t about grand transformations. They’re about practical tools people use daily.

A Harvard Business School and BCG study led by Ethan Mollick gave consultants access to AI for real work tasks. The results: consultants using AI completed 12.2% more tasks, finished them 25.1% faster, and produced results rated 40% higher quality. No transformation project. No 6-month roadmap. Just people using a capable tool for their daily work.

Stanford’s Erik Brynjolfsson studied customer support agents using an AI assistant. Productivity rose 14% on average, with the least experienced workers seeing gains up to 35%. His observation: “I’ve done lots of work on the introduction of new information technology over the years, and often companies are happy to get 1% or 2% productivity gains.”

A 2025 OpenAI survey of ChatGPT Enterprise users found they attributed 40 to 60 minutes of saved time per day to AI — just from using it as a daily work tool. Microsoft’s internal data showed the top 5% of Teams power users saving one full work day per month through AI meeting summaries alone.

The MIT NANDA study contained another crucial finding buried in the data: companies that purchased AI tools from specialized vendors succeeded about 67% of the time. Companies that tried to build their own solutions internally? One-third as often.

The pattern is clear. Practical tools embedded in existing workflows beat top-down transformation projects. Every time.

Strategy vs. Colleague

The standard playbook goes like this: hire consultants, spend 6 months on a strategy, build a proof of concept, run a pilot, maybe reach production, maybe see ROI. A process measured in quarters, if not years.

McKinsey’s own 2025 employee survey revealed what actually gets people to use AI: 48% said formal training, and 45% said “seamless integration into existing workflows.” Not a better strategy document. Not a bigger data lake. Integration into the tools they already use, every day.

This is the difference between an AI strategy and an AI colleague.

A strategy is a document. It sits in a SharePoint folder. It has swim lanes and a RACI matrix and deliverables organized by quarter. It requires a steering committee, a change management workstream, and executive sponsorship.

A colleague shows up where you work and starts being useful. You don’t need to reorganize your company around it. You don’t need a data science team. You don’t need to modernize your data infrastructure first. You need a tool that fits into the messaging app your team already has open eight hours a day — and starts helping from the first conversation.

The 67% success rate for purchased solutions vs. ~22% for internal builds isn’t surprising once you frame it this way. Buying a specialized tool means someone else solved the hard technical problems. Your team just has to use it. Building an AI strategy from scratch means your company has to become an AI company first — and for most businesses, that’s not the goal. The goal is to get work done faster.

The Path That Actually Works

If your team already lives in Microsoft Teams — and with 320 million monthly active users, there’s a decent chance they do — the integration point is obvious. We wrote about why Teams-native AI beats standalone tools in detail. You don’t need to convince anyone to adopt a new platform. You don’t need IT to evaluate 40 vendors. You don’t need a digital transformation initiative.

You need an AI that sits in the environment your people already use, remembers their context, and improves over time.

That’s what we built with amaiko. It’s a Teams-native AI assistant with persistent memory and specialized agents — for research, scheduling, email, document work. It works where your team already works. It remembers what you told it yesterday. No consulting engagement required. No 6-month roadmap. No steering committee.

The companies seeing real productivity gains from AI aren’t the ones with the best strategy decks. They’re the ones where people actually use the tool — because it’s right there, in the app they have open all day, doing useful work from day one.

Your team doesn’t need a strategy. They need a colleague.