Managed AI employees and AI consulting for growing businesses.

Update on the Persistence DEX USDC pools As announced last week, new deposits into the USDC/USDT and DYDX/USDC pools are now paused. Swaps and withdrawals remain open. If you still provide liquidity in either pool, please withdraw and migrate your Noble USDC before Circle's deadline of 12 January 2027. Skip Go (go.skip.build) supports a direct migration to Injective issued USDC. All other pools, and USDT, are unaffected.
Circle is winding down USDC on Noble, the USDC used across Cosmos including Persistence. Minting stops 13 Oct, the Noble contract is paused 12 Jan 2027. Persistence DEX: the USDC/USDT and DYDX/USDC pools are being deprecated. Withdrawals stay open. If you provide liquidity there, please withdraw and migrate your USDC via Skip Go before the deadline. USDT and all other pools are unaffected.
1
3
15
1,649
Reminders to validators and delegators to vote on proposal #155
Persistence validators and delegators: voting is open on proposal #155 for the core-1 v16.1.0 upgrade. Please review the proposal and cast your vote before 23 September, 19:34 UTC. mintscan.io/persistence/prop…
1
1
11
1,563
Persistence validators and delegators: voting is open on proposal #155 for the core-1 v16.1.0 upgrade. Please review the proposal and cast your vote before 23 September, 19:34 UTC. mintscan.io/persistence/prop…
5
18
3,789
Circle is winding down USDC on Noble, the USDC used across Cosmos including Persistence. Minting stops 13 Oct, the Noble contract is paused 12 Jan 2027. Persistence DEX: the USDC/USDT and DYDX/USDC pools are being deprecated. Withdrawals stay open. If you provide liquidity there, please withdraw and migrate your USDC via Skip Go before the deadline. USDT and all other pools are unaffected.
We are discontinuing support for USDC and CCTP V1 on the Noble blockchain as part of our broader transition to CCTP V2, a next-generation cross-chain protocol with faster finality and stronger security. Noble will not receive CCTP V2. New minting of USDC on Noble via Circle Mint stops October 13; the Noble USDC contract pauses January 12, 2027. All USDC on Noble remains redeemable 1:1. Read the details and required actions here: circle.com/blog/circle-is-di…
3
2
15
2,999
Persistence DEX on Babylon Genesis is being deprecated. Liquidity providers can withdraw their positions at any time at app.persistence.one/pools with Babylon selected. Swaps and withdrawals remain available for the time being. Persistence DEX on Persistence Core is not affected.
3
1
15
3,698
We have moved our official Telegram community chat to a new group to (hopefully) reduce spam, bots and inactive accounts 🧹 New link, please join here: t.me/PersistenceOneCommunity The original group is now read-only. Official announcements come only from @PersistenceOne on Telegram and this X account. Reminder: admins will never DM you first, and we will never ask for your seed phrase.
2
1
13
2,646
Admin work rarely fails loudly. It piles up as forms to check, notes to update, bookings to confirm and small exceptions nobody owns. That is useful AI employee territory: visible tasks, clear handoffs, fewer loose ends.
3
17
2,230
Sales follow-up is full of tiny leaks: drafts not sent, pipeline notes missing, stale leads nobody checks. That is good AI employee work because the job is narrow, repeated and close to revenue.
8
1
14
2,406
When an agent touches real work, the run needs to be inspectable. Session counts, errors, traces and tool steps give operators a record of what happened, where it slowed down, and what needs repair.
1
8
2,303
Launching a managed AI employee is the start of the operating work. The role needs approved tools, clear permissions, defined approvals, performance monitoring, and adjustment as the business changes.
2
2
9
2,272
Installing an AI employee should start with the operating setup, not the model. Choose one role, connect only approved tools, set permissions and guardrails, decide which approvals stay with the team, then improve it as the work changes.
1
6
2,333
Customer enquiry is a good first AI employee because the boundaries are visible. It can reply, chase, book and log. Offers, pricing, commitments, exceptions and exact outbound messages stay with the owner.
3
1
10
2,297
An internal knowledge assistant should not just answer faster. It should show what context it used, where the answer came from, and when the question needs a person instead of a guess.
2
8
2,199
Production agents are easier to review when every run leaves a trail: what context it used, which tools it called, where it paused for approval, and what evidence came back. That is the difference between automation and managed work.
4
1
12
2,301
One useful AI employee: lead follow-up. It watches new enquiries, drafts the approved reply, keeps pricing inside the agreed band, and flags exceptions before anything risky goes out.
4
9
2,320
The first useful AI employee usually has a boring job. One recurring workflow. Clear inputs. Limited tools. Known approval points. That is where trust starts.
1
10
2,340
Every agent failure teaches the same lesson: intelligence alone is fragile. A company needs routing, approvals, memory, verification and a safe way to stop when the system is unsure.
2
12
2,433
The biggest AI mistake companies make is jumping to implementation before understanding their own operations. Know where work is stuck before you try to automate it.
2
11
2,508
Small teams directing armies of agents. That is the next company operating model, and it is closer than most businesses think.
1
12
2,541
AI employees will not be optional. The companies that figure out how to hire, brief and manage digital workers alongside human teams will simply outpace the ones that do not.
2
11
2,459