One AI Stack Beats Tool Sprawl
North Shore firms that standardize on one platform cut vendor risk and get their staff to actually use it.
Key Takeaways
- ✓ AI tool sprawl is a symptom of ad hoc adoption, not a strategy, and it usually arrives one subscription at a time.
- ✓ One general-purpose anchor platform, such as Claude or ChatGPT, should handle drafting, research, and client communication by default.
- ✓ Give every overlapping point tool a fixed sunset date the day you sign the anchor platform's business plan, or it never actually goes away.
- ✓ Fewer AI vendors touching client files means fewer data-processing agreements to track, which is a real compliance win, not just a cost saving.
If your Highland Park law firm, Winnetka wealth practice, or Lake Forest insurance agency has added a new AI subscription every quarter this year, you already have the symptom. The fix is not another tool. It is picking one AI stack and retiring the rest.
The fastest way for a North Shore professional services firm to stop AI tool sprawl is to standardize on one general-purpose platform, such as Claude or ChatGPT, as the default for drafting, research, and client communication, then add a narrow point tool only where that core platform genuinely cannot do the job.
What Is AI Tool Sprawl, and What Is It Costing Your Firm?
AI tool sprawl is the accumulation of separate AI subscriptions across a firm, each covering one narrow task, with no single platform serving as the default anyone reaches for first. A partner drafts in ChatGPT. An associate summarizes depositions in a transcription app. The client services team drafts renewal letters in a third tool nobody else at the firm has access to.
None of that is exotic. It is what happens when every team member picks their own tool the week they first need one, and nobody circles back to consolidate. Shadow IT, the long-standing term for software staff adopt without a formal review, describes the same pattern now playing out with AI apps instead of spreadsheets and file-sharing tools.
The cost shows up in three places. First, licensing: five tools at roughly $20 a month each cost more than one shared plan built for a small team. Second, data governance: every point tool is another vendor holding your client files, another data-processing agreement someone should have actually read. Third, and most expensive, adoption: staff who juggle three logins for one job quietly stop using all three and go back to doing the work by hand. If your firm wants a plain read on where it stands today, Bace's AI readiness quiz takes ten minutes and flags the gaps.
"Innovation is saying no to a thousand things."
Steve Jobs, on the discipline of focus over feature accumulationHow Do You Choose a Single AI Stack for Your Firm?
Start with an audit, not a purchase. List every AI tool anyone at the firm currently pays for or runs on a free tier, who uses it, and what task it actually does. Most firms find the list is longer than any partner expected.
Next, pick the anchor platform: the one general-purpose tool that becomes the default for drafting, research, summarizing, and client communication across the firm. Anthropic's Claude and OpenAI's ChatGPT both work as an anchor. The deciding factor for most North Shore firms is which one plugs in cleanly to the software you already run, whether that is Applied Epic, Clio, or Wealthbox, and which one your compliance team is comfortable putting client data through. Judge the two on real differences: context window size for handling a full policy file or case record in one pass, and how clearly the vendor documents data handling for business accounts. Do not judge them on brand familiarity alone.
Then set a retirement rule. Give every overlapping point tool a 90-day sunset date the day you sign the anchor platform's business plan. Without a deadline, the old tools just sit there, still billed, still a second login nobody remembers to cancel. This is also where Bace's AI consulting engagements tend to add the most value: not picking the tool, but running the audit and the migration so consolidation actually happens on a calendar instead of a someday list.
SAMPLE CLAUDE PROMPT
"Here is a list of every software tool our firm currently pays for, with the department that uses each one and what task it handles. Group these into three buckets: tools a general AI assistant like you could fully replace, tools that overlap enough to consolidate into one, and tools that do something specialized enough to keep. For each tool marked replaceable, name the specific task it currently handles."
Where One Stack Beats a Pile of Point Tools
A single anchor platform is not automatically the right call for every task. Specialized tools still earn their place when they do something narrow that a general assistant cannot, such as a dedicated e-discovery platform for litigation or a purpose-built claims-scoring tool for underwriting. The test is whether the specialized tool does something a general platform truly cannot, not just something it does slightly differently.
| Question | Tool Sprawl | Single Anchor Stack |
|---|---|---|
| Who reviews the data-processing agreement | A different person for each vendor, if anyone | One agreement, reviewed once, applied everywhere |
| Where client data actually lives | Scattered across every point tool's servers | One vendor relationship to govern |
| What happens when a staff member leaves | Someone has to remember and cancel five logins | One account to deactivate |
| How staff actually use it | Adoption drops as login friction rises | One habit to build, reinforced daily |
This is not a theoretical concern. McKinsey's recurring survey on the state of AI has tracked adoption climbing at firms for several years running, with generative AI use moving from experiment to daily habit at the firms that stick with it. The firms falling behind are rarely the ones with too few tools. They are the ones where nobody uses any of the many tools consistently, because no single tool is the obvious first choice. Pull a few recent client wins from Bace's case studies to see what a consolidated setup looks like in practice for a North Shore firm.
Pick your anchor platform this quarter, put a 90-day sunset date on everything that overlaps with it, and watch Anthropic's and OpenAI's next enterprise releases before you add a single new subscription. Most of what a fourth AI tool promises is already sitting inside the one you have not fully turned on yet.
If you want a second opinion on which tools in your current stack are worth keeping, a free 30-minute AI audit with Bace Agency covers exactly that. No obligation, and the output is a one-page plan your team can act on this quarter.
Frequently Asked Questions
What is AI tool sprawl? +
AI tool sprawl is when a firm accumulates several overlapping AI subscriptions, each covering one narrow task, with no single platform serving as the default staff reach for first. It shows up as duplicate licensing costs, scattered client data across vendors, and staff quietly reverting to manual work rather than juggling multiple logins.
How many AI tools should a professional services firm run at once? +
Most North Shore firms do best with one general-purpose anchor platform, such as Claude or ChatGPT, handling the bulk of drafting, research, and client communication, plus at most one or two narrow point tools for tasks the anchor platform genuinely cannot handle, such as specialized e-discovery or underwriting software.
Should our firm pick Claude or ChatGPT as the core platform? +
Either can serve as an anchor platform. Judge them on how well they connect to the case, policy, or client management software you already run, how large a file each can handle in a single pass, and how clearly the vendor documents data handling for business accounts, rather than on brand familiarity.
How long does consolidating onto one AI stack take? +
Firms that set a fixed sunset date, commonly 90 days, for every tool that overlaps with the new anchor platform tend to finish the consolidation. Firms that migrate without a deadline often keep paying for the old tools indefinitely because no one owns the cutoff.
Related Articles

What North Shore Firms Should Not Automate
Most AI consultants will automate anything you point at. Here is the short list I tell clients to leave alone.

What Anthropic's Valuation Means for Small Firms
Anthropic just became the most valuable AI startup at a $965 billion valuation. The fad question is closed. The useful question is where AI fits in your firm, and where it does not.

AI Just Got Cheaper. Here Is What to Do
Model prices fell by a large multiple in two years. The instinct is to wait for cheaper. That is backwards. The API bill was never the real cost.
About the author
Written by
Michael Pavlovskyi
Founder, Bace Agency
Michael builds custom Claude and GPT workflows for insurance agencies, law firms, and PE firms on Chicago's North Shore. Speaker at Northwestern and Lake Forest College on practical AI adoption for professional services.
Connect on LinkedInWant to see how AI fits in your firm?
Book a free 30-minute AI audit. No obligation, no pitch deck.
Book a Free AI Audit →