Most AI advice is about what to type. This session is about how to think. Our CSO Schuyler Dragoo is running a 60-minute live working session on using AI with more intention: how to get AI to plan before it works, how to structure questions so each answer makes the next one smarter, and how to tell when AI is giving you a confident answer that isn't actually right for your situation.

She'll also share prompts and techniques you can use right away.

For leaders and operators who already use AI but want more than generic output, and are ready to build more intentional systems over time.

Free Training Lab | Wednesday, March 25 | 1pm ET
Save Your Spot →

• Google added Gemini directly into Sheets, Docs, Slides, and Drive, letting users build spreadsheets, draft documents, and clean data using plain-language instructions. Rolling out in beta to AI Ultra and Pro subscribers. (TechCrunch, Google Blog)

• Microsoft restructured its Copilot organization and moved Mustafa Suleyman to building proprietary AI models, signaling a long-term plan to reduce reliance on OpenAI. Former Snap exec Jacob Andreou now leads the unified Copilot product. (Microsoft Blog, CNBC)

• OpenAI plans to nearly double its workforce to 8,000 by end of 2026, hiring 12 people per day across product, engineering, and a new "technical ambassadorship" role focused on enterprise adoption. Separately, OpenAI is offering private equity firms 17.5% guaranteed returns to form joint ventures that would roll out AI tools across their portfolio companies, as both OpenAI and Anthropic court buyout firms ahead of potential IPOs. (CNBC, Reuters)

• OpenAI acquired Promptfoo, an AI security startup used by 25%+ of Fortune 500 companies, to build testing and compliance tools into its enterprise Frontier platform as businesses push to deploy autonomous AI agents. (CNBC, TechCrunch)

• OpenAI's chief scientist Jakub Pachocki told MIT Technology Review the company is building a fully automated AI researcher, with an "AI research intern" capable of multi-day tasks targeted for September 2026 and a full multi-agent automated researcher by 2028. Pachocki says the system won't be limited to AI research: "any kind of problem that can be formulated in text, code or whiteboard scribbles." (MIT Technology Review)

• OpenAI is in advanced talks to buy fusion energy from Sam Altman-backed Helion Energy, targeting 5 gigawatts by 2030 and 50 gigawatts by 2035. Altman has recused himself from negotiations. Microsoft signed a similar deal with Helion in 2023. (TechCrunch, Axios)

• Anthropic shipped Claude Code Channels, letting users message Claude from Telegram or Discord and have a live local Claude Code session execute the work and reply in-chat. Positioned as a mobile control layer for AI coding agents. (VentureBeat)

• DoorDash launched a standalone "Tasks" app paying its 8 million gig workers to submit videos of household activities like folding laundry and loading dishwashers to train AI and robotics models. Uber and Instacart have made similar moves. (Bloomberg, TechCrunch)

• Anthropic, AWS, Google, Microsoft, OpenAI, and GitHub pledged $12.5 million to the Linux Foundation to help open source maintainers defend against a growing wave of AI-generated security vulnerabilities and malicious code submissions. (OpenSSF, The Register)

• The White House released a National AI Legislative Framework designed to preempt 1,500+ state and local AI bills with a unified federal approach. Separately, the AI Accountability Act would require bias audits for AI used in hiring, lending, healthcare, and criminal justice. (SBE Council, Nextgov)

• Salesforce launched the Agentic Enterprise License Agreement (AELA), offering unlimited AI agent usage for a flat fee over two- or three-year contracts, covering Agentforce, Data Cloud, and MuleSoft with no per-action or per-token charges. (Forrester, The Register)

• Mistral released Small 4, a 119-billion-parameter open-source model under Apache 2.0 with a 256K context window, configurable reasoning, and multimodal input. Self-hostable at no licensing cost with 3x the throughput of its predecessor. (Mistral AI, Testing Catalog)

Harvard Business School researchers studied 78 workers at a global financial services firm and found that AI cut task time by two-thirds across the board. But quality told a different story. Workers close to the domain (web analysts writing about web analytics) scored 3.96 out of 5. Workers in adjacent roles (marketers) scored nearly as well at 3.92. Workers far from the domain (software developers writing marketing content) scored just 3.42, trailing by 13%, even with the same AI tools. The researchers call it "knowledge distance": AI amplifies what you already know, but it can't close a gap in domain understanding.

The bottleneck isn't AI capability. It's making sure the people using AI are close enough to the work to know when the output is right.

Sources: Harvard Business School Working Knowledge (2025)

You're paying for 50 software licenses and 30 people actually log in. You renewed three contracts last quarter on autopilot because nobody had time to dig into the usage reports. Across a typical 50-person company, unused SaaS licenses alone cost $10,000-$40,000 a year, and that's before you count the tools with overlapping features.

AI spots the waste your finance team doesn't have time to find. Try this with: vendor invoices, license counts, usage reports, contract terms, or even just a list of every tool your company pays for and what each team actually uses.

Try both as a single comprehensive prompt and by running each part separately. You'll often get better results if you start with the spend audit first, then build to the negotiation strategy.

Role: "Act as a skeptical procurement analyst auditing software and vendor spend for a [company size] [industry] company where every dollar of overhead directly affects margins."

Task: "Analyze these vendor contracts and usage patterns to identify waste, overlap, and negotiation leverage, then build a playbook for upcoming renewals."

Context: "We spend approximately [$X/month] across [X] vendors and SaaS tools. We have [X] employees. I suspect we're overpaying on several contracts but haven't had time to audit. Our next major renewals are [list any you know]. Our biggest line items are [top 3 tools/vendors]."

Format: Deliver:
SPEND SNAPSHOT | Total monthly/annual spend across all vendors; number of tools; cost per employee; percentage of budget going to top 5 vendors
WASTE AUDIT | Table: Tool | Licensed Seats | Active Users | Monthly Cost | Cost of Unused Seats | Overlap With Other Tools
NEGOTIATION LEVERAGE MAP | For each vendor approaching renewal: current contract terms; market alternatives; your usage trend (growing, flat, declining); leverage points (competitor pricing, multi-year discount, seat reduction)
RENEWAL SCRIPTS | For top 3 overspend areas: opening line; key data point to cite; specific ask (discount %, seat reduction, feature bundle); walk-away position
COST RECOVERY PLAN | Immediate actions (this week): cancel or downgrade unused licenses; 30-day actions: renegotiate top 3 contracts; 90-day plan: consolidate overlapping tools

Constraints: Distinguish between tools that are underused vs. tools that are genuinely unnecessary; Flag contracts with auto-renewal clauses or cancellation windows approaching; Identify where switching costs make renegotiation smarter than replacement; Be honest about tools your team actually needs even if usage looks low; Estimate total annual savings opportunity.

Tip: Run this before every renewal cycle. Companies that audit vendor spend quarterly recover an average of 15-25% on SaaS costs compared to companies that auto-renew without review.

Instant SOP
Describe something you do regularly that only you know how to do. Approving invoices, onboarding a client, closing the books, whatever lives in your head. Walk through it in whatever order it comes to you, messy is fine. Turn this into a step-by-step procedure clear enough that someone covering for me could follow it without asking questions. Return as: Numbered procedure + decision points where judgment is needed + common mistakes to avoid.

The Five-Minute Handoff
Describe everything you're currently working on. Projects, deadlines, conversations in progress, anything someone would need to know if you disappeared for two weeks. Don't worry about organizing it. Build a handoff document covering what's in progress, who's waiting on what, what's due soonest, and what to watch out for. Return as: Priority-ranked project list + pending decisions + key contacts + landmines to avoid.

Where Was I?
Describe a project or conversation you set down weeks ago and need to pick back up. Paste any notes, emails, or docs you still have from it. Reconstruct where I left off, what decisions were still open, what I was probably about to do next, and what's likely changed since. Return as: Status snapshot + open decisions + your next three moves + one thing to verify before restarting.

Four words changed how I use AI: "and tell me why."

I'd been using AI models the way most people do. Give it a task, get a result, move on. Then I started tacking "and tell me why" onto the end of requests. Ask it to recommend a course of action, then add "and tell me why." Ask it to explain why a system you inherited is designed a certain way, then add "and tell me why." The output goes from answer to explanation. You're not just getting the fish, you're watching how it was caught.

Here's where it gets interesting: once the model explains its reasoning, you can tell it to apply what it just found. "Now take what you learned from that analysis and use it to improve X." The model builds on its own insights in real time, and you're learning right alongside it. It turns a one-shot answer into an evolving conversation where both you and the AI get sharper.

AI isn't just a productivity tool, it's a tutor that works at the speed of your curiosity. Most people stop at the output. The ones getting the most out of it are the ones asking "why" and then pointing the model back at its own findings.

The most valuable knowledge in your company is the kind that only lives in one person's head. The most expensive contracts are the ones nobody has time to audit. AI doesn't fix either of those problems on its own, but the right systems do.

Kiingo helps teams build AI workflows that capture expertise, cut waste, and scale what's already working. No new tools required. Just more from what you have.

Want to see where to start? Book a short discovery call.