Google unveils Gemini 3.5 Flash with 4x speed boost
Google's new model runs 4x faster than current versions at half the cost. It rivals GPT-5.5 across benchmarks while being cheaper to operate.
|
For your business
If you spend $500+ monthly on AI tools for document processing, test this model and cut costs by 40%.
|
Uber questions AI spending returns
Uber president Andrew Macdonald said the company can't draw a line between AI investments and delivered features. They hired fewer people to offset AI costs but see no productivity gains.
|
For your business
Audit current AI tool spend and identify which saved actual hours this quarter, not usage metrics.
|
Google Search gets biggest redesign in years
Google added cross-modal inputs and generative UI for custom layouts. The update integrates agents across Workspace and Chrome for automated task completion.
|
For your business
If your team searches supplier data daily, this cuts lookup time from 15 minutes to 2 minutes per query.
|
|
Most companies measure AI usage instead of AI results. Token consumption means nothing if your dispatch team still spends 6 hours daily on route changes. The question isn't whether your team uses AI tools. It's whether those tools save money you can count.
|
|
A real estate investor in the Northeast running a multifamily portfolio. He was losing 6 hours a week on two ends of his business. The small end: receipts, statements, and transaction records from four bank accounts were scattered across email attachments, paper piles, and a half-organized Dropbox. He could not tell you which property had which expense without an hour of digging. The big end: he was raising capital for the next deal with no system for tracking which LPs he had spoken to, who was warm, or who had asked for what kind of update. Investor outreach was happening from memory, and follow-ups were slipping.
We built two automations. First, a document ingestion pipeline that pulls bank statements, receipts, and property docs from his inboxes and Dropbox, tags them by property and category, and stores them in a searchable Notion database with monthly summaries. Second, an investor CRM that pulls signals from his calendar, email, and call notes, scores each LP relationship, surfaces who is due for a touchpoint, and drafts personalized update emails ready for him to review.
Monthly bookkeeping prep dropped from 5 hours to 30 minutes. Investor follow-up cadence went from inconsistent to weekly across 40+ LP relationships. He closed his last raise 6 weeks faster than the previous one and named two new LPs that came directly from the CRM surfacing dormant relationships. The compounding effect: he is spending recovered time on deal underwriting instead of administrative work, which is the actual lever for portfolio growth.
|
|
We build systems like this for logistics, manufacturing, and service businesses.
If that sounds like you, here's how we work.
|
Audit your AI spending this week. Most operations leaders can't name monthly AI costs or connect usage to results.
| 1. |
List every AI tool your company pays for. Include ChatGPT, Copilot, custom integrations, anything with AI in the invoice. |
| |
| 2. |
Pull 90 days of usage data from each tool. Most have dashboards showing active users and consumption. |
| |
| 3. |
Interview three team members per tool. Ask what specific task it saves time on and how many hours weekly. |
| |
| 4. |
Calculate cost per hour saved. Divide monthly tool cost by total hours saved across users. |
| |
| 5. |
Cancel tools costing more than $25 per hour saved. Your team's hourly rate is probably lower. |
|
| |
Process Street
Creates automated checklists and workflows with conditional logic and team assignments.
|
Best for: operations teams managing safety inspections, quality checks, or vendor onboarding that need consistent execution every time
Pricing: $100/month for 10 users with unlimited workflows
|
|
|
—
The companies winning with AI aren't spending the most on it.
|
|
